View Full Version : Easy Debian Fremantle Beta Testing
Pages :
1
2
3
4
5
6
7
8
9
10
[
11]
12
13
14
Things that you need to install before LibreOffice, if apt-get isn't able to resolve dependencies:
ure_3.4.3-3~bpo60+1_armel.deb (http://lorienart.pl/cover/ure_3.4.3-3~bpo60+1_armel.deb)
uno-libs3_3.4.3-3~bpo60+1_armel.deb (http://lorienart.pl/cover/uno-libs3_3.4.3-3~bpo60+1_armel.deb)
ttf-opensymbol_2.4.3+LibO3.4.3-3~bpo60+1_all.deb (http://lorienart.pl/cover/ttf-opensymbol_2.4.3+LibO3.4.3-3~bpo60+1_all.deb)Just in case it hasn't been mentioned before, this will resolve the dependencies automatically:apt-get install -t squeeze-backports libreofficebackground:
By default the backports repository has a lower pinning priority than the regular stable repository so apt will try to resolve any dependencies from the stable repository, even if there is a newer version in backports. Only if there's no package with the required name at all in the stable repository the backports repository will be used. In this case that means that regardless of the insufficient version number ure will be taken from the stable repository because it exists there.
-t (target) changes that behavior by giving the specified repository the highest pinning priority for this one command. Thus -t squeeze-backports means: Resolve any dependencies by using the squeeze-backports repository if possible, fetch the rest from whatever repository among all the others has the highest pinning priority.
In our case this means ure (and any other dependency that exists in the backports) will be installed from squeeze-backports. Therefore -t squeeze-backports SHOULD be used whenever you explicitely want to install anything from backports.
Oh and just for fun, a very NOT easy Debian:
http://farm8.staticflickr.com/7020/6618289381_f8bbe83936_b.jpg (http://www.flickr.com/photos/qole2/6618289381/lightbox/)
How Can I install debian on my n9?I think is very hard to do and i am not familiar with linux
I'm struggling around with the locales in ED.
apt-get upgrade and dpkg-reconfigure localeconf shows this output:
[root@m5sulu: /]dpkg-reconfigure localeconf
perl: warning: Setting locale failed.
perl: warning: Please check that your locale settings:
LANGUAGE = (unset),
LC_ALL = (unset),
LC_PAPER = "de_DE",
LC_ADDRESS = "de_DE",
LC_MONETARY = "de_DE",
LC_NUMERIC = "de_DE",
LC_TELEPHONE = "de_DE",
LC_MESSAGES = "de_DE",
LC_IDENTIFICATION = "de_DE",
LC_MEASUREMENT = "de_DE",
LC_TIME = "de_DE",
LC_NAME = "de_DE",
LANG = "de_DE"
are supported and installed on your system.
perl: warning: Falling back to the standard locale ("C").
locale: Cannot set LC_CTYPE to default locale: No such file or directory
locale: Cannot set LC_MESSAGES to default locale: No such file or directory
locale: Cannot set LC_ALL to default locale: No such file or directory
Package `localeconf' is not installed and no info is available.
Use dpkg --info (= dpkg-deb --info) to examine archive files,
and dpkg --contents (= dpkg-deb --contents) to list their contents.
Setting LANGUAGE and LC_ALL works temporary for apt-get upgrade until next chroot login.
[root@m5sulu: /]export LANGUAGE="de_DE"
[root@m5sulu: /]export LC_ALL="de_DE.UTF8"
Is it advisable to change this manualy, or is there a better way? And if so could anybody give me a tip where to set this variables permanent?
Edit:
These settings don't work very well, some symbols and ä ö ü for example are displayed wrong. Trying to get this sorted .....
sulu, thanks for that! Works perfectly.
@eight
I'm experiencing same issue, but I haven't seen any situation, where symbols or my language special characters (pl_PL) - ą ć ś ó ę ł ż ź ń etc - aren't working properly.
where exactly You face actual problems? Would like to check it myself.
[root@m5sulu: /]dpkg-reconfigure localeconfWhat are you trying to achieve with that? There is no package localeconf in the Debian repositories so that can't work.
Do you mean?:dpkg-reconfigure locales
Sorry I meant:
dpkg-reconfigure locales
Copied from the wrong Browse-tab I think.
Though it shows me the same warning.
And the same warning occurs when doing apt-get upgrade
[root@m5sulu: /]apt-get upgrade
Reading package lists... Done
Building dependency tree
Reading state information... Done
The following packages have been kept back:
libavcodec52 libavformat52 libavutil50 libpostproc51
The following packages will be upgraded:
aptitude base-files ca-certificates chromium-browser
chromium-browser-inspector chromium-browser-l10n dhcp3-client dhcp3-common
iceweasel isc-dhcp-client isc-dhcp-common libavutil49 libcups2 libcupsimage2
libfreetype6 libgl1-mesa-dri libgl1-mesa-swx11 libglu1-mesa libgssapi-krb5-2
libjasper1 libk5crypto3 libkrb5-3 libkrb53 libkrb5support0 libmozjs2d
libnss3-1d libosmesa6 libpam-modules libpam-runtime libpam0g
libpolkit-agent-1-0 libpolkit-backend-1-0 libpolkit-gobject-1-0 libsmbclient
libsoup-gnome2.4-1 libsoup2.4-1 libssl0.9.8 libvpx0 libvte-common libvte9
libwbclient0 mesa-utils mplayer openssh-client openssh-server openssl
policykit-1 pulseaudio tzdata tzdata-java update-inetd usbutils x11-common
xbase-clients xulrunner-1.9.1
55 upgraded, 0 newly installed, 0 to remove and 4 not upgraded.
Need to get 0 B/64.3 MB of archives.
After this operation, 762 kB disk space will be freed.
Do you want to continue [Y/n]?
perl: warning: Setting locale failed.
perl: warning: Please check that your locale settings:
LANGUAGE = (unset),
LC_ALL = (unset),
LC_PAPER = "de_DE",
LC_ADDRESS = "de_DE",
LC_MONETARY = "de_DE",
LC_NUMERIC = "de_DE",
LC_TELEPHONE = "de_DE",
LC_MESSAGES = "de_DE",
LC_IDENTIFICATION = "de_DE",
LC_MEASUREMENT = "de_DE",
LC_TIME = "de_DE",
LC_NAME = "de_DE",
LANG = "de_DE"
are supported and installed on your system.
perl: warning: Falling back to the standard locale ("C").
locale: Cannot set LC_CTYPE to default locale: No such file or directory
locale: Cannot set LC_MESSAGES to default locale: No such file or directory
locale: Cannot set LC_ALL to default locale: No such file or directory
Extracting templates from packages: 100%
Preconfiguring packages ...
dpkg: unrecoverable fatal error, aborting:
syntax error: unknown group 'crontab' in statoverride file
E: Sub-process /usr/bin/dpkg returned an error code (2)
Edit:
Overlooked
syntax error: unknown group 'crontab' in statoverride file
After adding
groupadd crontab
"apt-get upgrade" works. :)
Output when running "dpkg-reconfigure locales":
[root@m5sulu: /]dpkg-reconfigure locales
perl: warning: Setting locale failed.
perl: warning: Please check that your locale settings:
LANGUAGE = (unset),
LC_ALL = (unset),
LC_PAPER = "de_DE",
LC_ADDRESS = "de_DE",
LC_MONETARY = "de_DE",
LC_NUMERIC = "de_DE",
LC_TELEPHONE = "de_DE",
LC_MESSAGES = "de_DE",
LC_IDENTIFICATION = "de_DE",
LC_MEASUREMENT = "de_DE",
LC_TIME = "de_DE",
LC_NAME = "de_DE",
LANG = "de_DE"
are supported and installed on your system.
perl: warning: Falling back to the standard locale ("C").
locale: Cannot set LC_CTYPE to default locale: No such file or directory
locale: Cannot set LC_MESSAGES to default locale: No such file or directory
locale: Cannot set LC_ALL to default locale: No such file or directory
/usr/bin/locale: Cannot set LC_CTYPE to default locale: No such file or directory
/usr/bin/locale: Cannot set LC_MESSAGES to default locale: No such file or directory
/usr/bin/locale: Cannot set LC_ALL to default locale: No such file or directory
Generating locales (this might take a while)...
de_DE.UTF-8... done
en_CA.ISO-8859-1... done
en_CA.UTF-8... done
en_GB.ISO-8859-1... done
en_GB.ISO-8859-15... done
en_GB.UTF-8... done
en_US.ISO-8859-1... done
en_US.ISO-8859-15... done
en_US.UTF-8... done
Generation complete.
perl: warning: Setting locale failed.
perl: warning: Please check that your locale settings:
LANGUAGE = (unset),
LC_ALL = (unset),
LC_PAPER = "de_DE",
LC_ADDRESS = "de_DE",
LC_MONETARY = "de_DE",
LC_NUMERIC = "de_DE",
LC_TELEPHONE = "de_DE",
LC_MESSAGES = "de_DE",
LC_IDENTIFICATION = "de_DE",
LC_MEASUREMENT = "de_DE",
LC_TIME = "de_DE",
LC_NAME = "de_DE",
LANG = "C"
are supported and installed on your system.
perl: warning: Falling back to the standard locale ("C").
perl: warning: Setting locale failed.
perl: warning: Please check that your locale settings:
LANGUAGE = (unset),
LC_ALL = (unset),
LC_PAPER = "de_DE",
LC_ADDRESS = "de_DE",
LC_MONETARY = "de_DE",
LC_NUMERIC = "de_DE",
LC_TELEPHONE = "de_DE",
LC_MESSAGES = "de_DE",
LC_IDENTIFICATION = "de_DE",
LC_MEASUREMENT = "de_DE",
LC_TIME = "de_DE",
LC_NAME = "de_DE",
LANG = "C"
are supported and installed on your system.
perl: warning: Falling back to the standard locale ("C").
Ok, now (I think) I know what's going on. I once accidentally added all possible locales to my Easy Debian images, which took a lot of space. So I deleted most of them and only kept de_DE.UTF-8 (UTF-8 is the default encoding in Linux and therefore Easy Debian) and the en and us locales for fallback reasons.
Maemo however uses ISO-8859 locales (e.g. de_DE). Since my Easy Debian image no longer has the ISO-8859 locale any locale-dependent application that gets the Maemo locale throws warnings. I have them too but since it's not critical I simply didn't care for it. I just installed de_DE in Easy Debian via dpkg-reconfigure (but kept the UTF-8 locale as the default) and the warnings are gone.
So to get rid of your warnings look up what your Maemo locale looks like and install it in Easy Debian. For Estel that would mean Maemo most likely has pl_PL. If he uses my Easy Debian image he doesn't have that locale so he should select that in dpkg-reconfigure. While doing that it might also be useful to add pl_PL.UTF-8 as the Easy Debian's default locale.
Thanks, but I'm not using Your image - except for /user/share/X11/xkb directory, which is entirely took from your image.
Anyway, I remember trying to reconfigure locales and selecting pl_PL variants (both ISO and UTF), yet I still have this error. will try again and report back.
OTOH, I tried what qole suggested, and did chown user:users to '/home/user/'libreoffice'. Yet, it still fail to start from user, and does so only from root. i'll try apt-get purge, manually get rid of leftovers, install again and avoid starting it from root - maybe it's going to help.
I'm just worried a little about leftovers from openoffice.org in 'debian' script used by debbie. I would love to get rid of those if they're not required anymore, but unfortunately, i'm not specialist in creating chroots, and qole seems to have other priorities than upgrading everything for LibreOffice.
On the third hand ;) I haven't had any issues using LibreOffice as is, so everyone who don't want to use depreciated and outdated OpenOffice can safely upgrade.
/Estel
// edit
I've also did apt-get upgrade -t squeeze-backports, and everything seems to be working fine (of course pulseaudio is held back like in sulu's first tutorial).
I have them too but since it's not critical I simply didn't care for it. I just installed de_DE in Easy Debian via dpkg-reconfigure (but kept the UTF-8 locale as the default) and the warnings are gone.
Thank you, that did it.
After selecting "de_DE ISO-8859-1" at dpkg-reconfigure and staying with "de_DE UTF-8" as default the warnings are gone. :)
Thought in the first place it could tend to problems with uprades, even better that it's not critical and can be irgnored.
Despite that I've been using my own image and for sure I haven't had such errors before, it seems that after dist-upgrade I was also lacking pl_pl (ISO-8859-1) - 'dpkg-reconfigure locales' - in which i generated pl_PL (ISO one) and set pl_PL.UTF-8 as default - also did the trick for me.
Warnings are gone, but now, polish language special characters aren't properly printed in terminal, but it's minor glitch.
just for the records - 'apt-get purge libreoffice', 'rm -r /home/user/.libreoffice' and 'apt-get install libreoffice -t squeeze-backports' did the trick about running LibreOffice from user, so now shortcuts (so, running through 'debbie libreoffice') are working flawlessly.
/Estel
@sulu
No joy with:
"xvkbd -xsendevent -text "\Alt_L""
...while, for example, F5 mapped as:
"xvkbd -xsendevent -text "\F5""
...works flawlessly. Probably it's something simple, but I'm out of ideas - tried to find correct syntax for xvkbd for a long time, but everywhere alt is listed only as modifier, not resulted keypress...
/Estel
As qole totally ignored my proposition about image with OO replaced with LibreOffice and Iceweasel replaced with Chromium, i don't think he's interested in hosting and adding another image for ED auto installer.
I don't mind hosting any complete images for the auto installer. If you send me a link like sulu did, I will host it. And I won't say insulting things about you.
@qole
Could you explain, what's the "magic" behind bt-hid-script, that allow hassle-free usage of external keyboard inside LXDE?
You have to run setxkbmap for DISPLAY :1 as well as :0. The relevant code from my bt-script:
if [ $EDMOUNT = yes ] ; then
echo "Setting layout for Easy Debian (display :1)."
# Set layout for Easy Debian
setxkbmap -device $EXTKBDID -I -I/usr/share/X11/xkb-chinook \
-rules $KBDRULES -model $KBDMODEL -layout $KBDLAYOUT -print \
| xkbcomp - :1 2>/dev/null
fi
(...)If you send me a link like sulu did, I will host it. And I won't say insulting things about you.(...)
Thanks a lot for answer and the code, but why do you think I meant to insult You? "ignored" was meant to mean exactly what it means in base language, without any layer 2 of irony. I've written about it, You have written few posts after it without picking up this topic = I've written that You ignored it. No ranting, no blaming, no irony intended.
Take my excuses, if my somehow "shortly-informative" style of writing actually left wrong impression.
/Estel
ibrakalifa
2012-01-14, 03:39
i can chroot ubuntu image, but not with fedora one, what should i do??, and anybody know how to chroot meego image easily as easy-debian has??
Well, I'm a little slow to get to things but I finally tried kernel-power v49 and CSSU.
1)When over clocking there was good improvement in starting up LDXE. For example when increasing clockspeed form 600Mhz to 900Mhz the time to start up LXDE went from about 30s to 20s give or take a second or two. This was impressive! Where I was disappointed was that if I start up OO writer from LXDE there was not much improvement. Is this mainly due to RAM limitations or any idea if having the dedicated partition will further improve. Regardless, I will next try to make the dedicated partition for Easy Debian on sdcard. As I am a bit inexperienced at this would someone recommend some posts to follow instructions. Also, any caveats?
2)Just an observation in re CSSU. If I start ED in portrait the LXDE actually appeared in portrait for a second but then went to landscape without me rotating the N900. Doesn't really matter as it is hard to use the HWKB in portrait anyways!
Hi,
I dist-upgrade to squeeze, but without reading the wiki before :-(
So now i have a problem with dbus and hal because i remove
/etc/apt/preferences before upgrade and i get dbus and hal from squeeze.
Note that when i 'sudo /etc/init.d/dbus restart' under the chroot, my n900 reboot !!! :-(
So i put some "exit 0" on top of /etc/init.d/dbus and /etc/init.d/hal to prevent reboot.
Now i can run pulseaudio (the version i get from the wiki) but there is no sound :
$ alsamixer
ALSA lib dlmisc.c:236:(snd1_dlobj_cache_get) Cannot open shared library /usr/lib/alsa-lib/libasound_module_ctl_dsp_ctl.so
cannot open mixer: No such device or address
I notice that the module libasound_module_ctl_dsp_ctl.so was remove between lenny and squeeze from libasound2-plugins.
Thank you.
@guyr:
I'm not sure what happened to your Easy Debian but it seems to me like something went wrong during the dist-upgrade and now you have a mixed alsa installation.
My first attempt to fix it would be to reinstall libasound2-plugins or even all alsa-related packages.
@mscion:
The reason why OO and other heavy applications don't start significantly faster after overclocking is that most of their delay comes from reading all the data from the mass storage where the bottle neck is the mass storage, not the CPU. (btw: this is exactly the type of applications where in a desktop computer replacing an HDD for an SSD would make sense)
As for having a dedicated partition for Easy Debian I can't give you first-hand information since I use image files. But derived from other Debian chroots all you have to do is to copy all the data from the image to an ext2 (less write access) or ext3 (more robust FS) partition on your sd card and adjusting the Easy Debian start script not to mount the image but the partition (or mount it via Maemo's fstab directly).
The reason why I still use image files is because they are easier to handle. I usually have multiple images on my sd card (at least: work image, test image and backup image) and if I mess something up it's easier to restore (well, now that I think about it this isn't quite right but the alternative would be less flexible in terms of storage space).
If you want portrait mode in Easy Debian try the patched x-server packages linked by Estel some time ago! I will add a section for that to the Easy Debian wiki article!
maartenmk
2012-01-23, 20:41
@mscion: A dedicated partition for ED is highly recommended. I did quite some measurements of OO startup times in different configurations some time ago, search the thread.
I also found out that starting OO from within Maemo is faster than starting LXDE and then starting OO.
@Estel: Is your image available yet? I was hoping to see a further improvement in speed and compatibility with Libre-Office
Sorry, was pushed by tight personal schedule. Really, will try to do it ASAP, but this time I won't promise "today or tomorrow", don't want to fail expectations ;)
@Sulu
Because You're kinda specialist here ;) Any highlights, about what should be done before image can be considered "uploadable" for wide audience? I thought for setting locales to EN, instead of my PL. Any other remarks?
/Estel
// Edit
sulu, excuse my noobiness, but where to put files from attachment to:
http://wiki.maemo.org/Documentation/Maemo_5_Developer_Guide/Packaging%2C_Deploying_and_Distributing#Creating_D ebian_Packages
...(lxpanel.tar)? Of course, orientation script goes to /.debian/usr/bin/ (I correctly assumed, that it should be executed from within ED, *not* Maemo itself, right?), but for other files, I have no clue.
---
Also, as ED is far from "testing" now, we have fully working Squeeze release, and this thread grew up to epic size, what do You think about creating proper announcement, after releasing image with all sulu's work + chromium instead of Iceweasel (faster on N900, less RAM-hungry - to the point, that I use it comfortably as "everyday" browser on my device), and LibreOffice instead of OpenOffice?
All this Squeeze thing seems to be big milestone for ED. qole is less active now in Maemo area (no accusing, no demanding, no rant, no offense, no bad words, etc ;) ), and proper announcement by active member (sulu, most likely, or, in worst case, me) would give control over first page (updates), + probably bring fresh, well-deserved attention to Easy Debian. At least, IMHO.
@Sulu
Because You're kinda specialist here ;) Any highlights, about what should be done before image can be considered "uploadable" for wide audience? I thought for setting locales to EN, instead of my PL. Any other remarks?Not really. You're right, I should have set my locales to EN as well. But I just didn't think that it might actually be of public interest. I thought qole would use it as a template for a completely new image.
The only thing you should make sure is that it contains no compromising data like your personal VPN login data or something similar.
Another good idea to minimize the size of the compressed image would be to nullify all the empty space in the image. It just did it by creating a file from /dev/zero inside of the image and deleting it. But there's also the Debian package zerofree which should do that even more comfortably.
Thanks! What about the:
// Edit
sulu, excuse my noobiness, but where to put files from attachment to:
http://wiki.maemo.org/Documentation/Maemo_5_Developer_Guide/Packaging%2C_Deploying_and_Distributing#Creating_D ebian_Packages
...(lxpanel.tar)? Of course, orientation script goes to /.debian/usr/bin/ (I correctly assumed, that it should be executed from within ED, *not* Maemo itself, right?), but for other files, I have no clue.
---
Also, as ED is far from "testing" now, we have fully working Squeeze release, and this thread grew up to epic size, what do You think about creating proper announcement, after releasing image with all sulu's work + chromium instead of Iceweasel (faster on N900, less RAM-hungry - to the point, that I use it comfortably as "everyday" browser on my device), and LibreOffice instead of OpenOffice?
All this Squeeze thing seems to be big milestone for ED. qole is less active now in Maemo area (no accusing, no demanding, no rant, no offense, no bad words, etc ;) ), and proper announcement by active member (sulu, most likely, or, in worst case, me) would give control over first page (updates), + probably bring fresh, well-deserved attention to Easy Debian. At least, IMHO.
(probably, You've started to write, before I edited)
/Estel
ivgalvez
2012-01-24, 07:47
Also, as ED is far from "testing" now, we have fully working Squeeze release, and this thread grew up to epic size, what do You think about creating proper announcement, after releasing image with all sulu's work + chromium instead of Iceweasel (faster on N900, less RAM-hungry - to the point, that I use it comfortably as "everyday" browser on my device), and LibreOffice instead of OpenOffice?
All this Squeeze thing seems to be big milestone for ED. qole is less active now in Maemo area (no accusing, no demanding, no rant, no offense, no bad words, etc ;) ), and proper announcement by active member (sulu, most likely, or, in worst case, me) would give control over first page (updates), + probably bring fresh, well-deserved attention to Easy Debian. At least, IMHO.
Absolutely agree with you.
tanpoaran
2012-01-24, 07:55
@sulu
i was failed install libre office and i'v got this:
root@m5sulu: /]dpkg --configure -a
dpkg: failed to open package info file `/var/lib/dpkg/updates/0017' for reading: Stale NFS file handle
please tell me how to solve this problem..
thank you
sulu, excuse my noobiness, but where to put files from attachment to:
http://wiki.maemo.org/Documentation/Maemo_5_Developer_Guide/Packaging%2C_Deploying_and_Distributing#Creating_D ebian_Packages
...(lxpanel.tar)? Of course, orientation script goes to /.debian/usr/bin/ (I correctly assumed, that it should be executed from within ED, *not* Maemo itself, right?), but for other files, I have no clue.Yes, the orientation script goes to /usr/bin/ and is executed from within ED.
The LDXE, HLXDE and XLXDE directories are lxpanel configurations and go to: /home/user/.config/lxpanel/
If you have a look at the orientation script you'll see that every time the orientation of the desktop is changed another panel profile gets loaded since I found it useful to always have the panel on the short side of the screen. Just like my entire image I considered these profiles to be an example of how the lxpanel might be influenced by the orientation. So if you copy these profiles to your image you'll have my panel layout as well and I'm not sure if that's what you want because it differs quite a lot from the Easy Debian default panel.
The LXDE profile is my panel for the landscape desktop. HLXDE (H from horizontal) is basically the same panel for the portrait desktop, and XLXDE is the HLXDE profile plus an xvkbd keyboard plus a "dummy panel" (width=1px) to prevent other windows from covering xvkbd (or being covered by it).
Unfortunately each of those profiles has to be edited separately, so if you want to add another icon to your panel you have to do it in all three profiles if you want them do be consistent. And the positioning of the xvkbd window depends on your panel layout.
So I suggest to not simply copy my profiles but adjust them to a default panel layout.
Edit: And then there is this orientation.desktop file which creates a (more or less) nice icon (I'm sure one can find a better one, I was to lazy to search) and goes to: /usr/share/applications/
Also, as ED is far from "testing" now, we have fully working Squeeze release, and this thread grew up to epic size, what do You think about creating proper announcement, after releasing image with all sulu's work + chromium instead of Iceweasel (faster on N900, less RAM-hungry - to the point, that I use it comfortably as "everyday" browser on my device), and LibreOffice instead of OpenOffice?I think it's a good idea.
I haven't tested (nor checked for backdoors) them yet but I think these two additions for chromium respectively iceweasel might be useful:
http://www.chromeextensions.org/appearance-functioning/chrometouch/
https://addons.mozilla.org/en-US/firefox/addon/grab-and-drag/
All this Squeeze thing seems to be big milestone for ED. qole is less active now in Maemo area (no accusing, no demanding, no rant, no offense, no bad words, etc ;) ), and proper announcement by active member (sulu, most likely, or, in worst case, me) would give control over first page (updates), + probably bring fresh, well-deserved attention to Easy Debian. At least, IMHO.I appreciate your trust in me but frankly I can't guarantee that I'd be constantly available as a developer or even host of Easy Debian. Everything I've contributed so far is just what I've done for myself. So I'll be happy to continue giving further support but I don't really want to be in charge of a project.
@sulu
i was failed install libre office and i'v got this:
root@m5sulu: /]dpkg --configure -a
dpkg: failed to open package info file `/var/lib/dpkg/updates/0017' for reading: Stale NFS file handle
please tell me how to solve this problem..
thank youI'm sorry, in my experience this is a very nasty error which I've not been able to fix yet. Whenever this occured to me I just restored my (hopefully up to date) backup image and started over again.
Usually this error only happens when the watchdog reboots the N900 during a software installation because it thinks it's frozen due to a runaway process. To prevent this always start processes which are going to occupy the CPU for a long time with nice, e.g.:nice -n 19 sudo apt-get install libreofficeThis way the watchdog will always get the CPU priority when it thinks it's time to check if everything is still alright.
Edit:
just a thought:
With Debian introducing the new armhf (hf= hard float) architecture and the N900's Cortex A8 supporting this architecture might it be possible to re-base Easy Debian on it? I've read about performance increases of up to 40% in special situations (although mostly it won't even be noticeable). If it works this won't happen anytime soon (maybe in a year or two), but I'd like to have some thoughts of people who are more familiar with this topic than I am.
just a thought:
With Debian introducing the new armhf (hf= hard float) architecture and the N900's Cortex A8 supporting this architecture might it be possible to re-base Easy Debian on it? I've read about performance increases of up to 40% in special situations (although mostly it won't even be noticeable). If it works this won't happen anytime soon (maybe in a year or two), but I'd like to have some thoughts of people who are more familiar with this topic than I am.
Not sure about this, but if you compile the whole system (easy debian in this case) with HF you might also need an adapted kernel (unless the interface between libc and kernel is independent of the architecture/ABI, which is probably not the case).
Some day we'll be able to boot an HF-compiled debian with an upstream kernel, but it'll take some time until that happens..
Not sure about this, but if you compile the whole system (easy debian in this case) with HF you might also need an adapted kernel (unless the interface between libc and kernel is independent of the architecture/ABI, which is probably not the case).Since I believe* the existing Maemo kernel makes best use of the hardware, it should already be armhf compatible. No idea if that's actually necessary or if an armhf-compatible libc would suffice, but if it is necessary that shouldn't be a problem.
Some day we'll be able to boot an HF-compiled debian with an upstream kernel, but it'll take some time until that happens..Well I think recompiling the whole Easy Debian software doesn't make much sense. If that's what you want to do I think Arch or Gentoo would be better distributions. Plus, recompiling every binary with armhf support would practically mean to create a new architecture on our own. I have thought about that earlier but I didn't find it practical then and I don't think so now.
That's where the Debian's official armhf port comes in handy now. Theoretically we could start building armhf images right away. The reason why I think there will be a delay of over a year is that it doesn't make much sense to base Easy Debian on Testing or even Unstable since it's supposed to be useable for everybody, not just for Debian experts (heck, I don't even consider myself to be one). Therefore I think the Wheezy freeze is the first date that might make an armhf Easy Debian useful.
*) read: I have absolutely no clue but it would be wise
zlatokosi
2012-01-24, 17:25
Also, as ED is far from "testing" now, we have fully working Squeeze release, and this thread grew up to epic size, what do You think about creating proper announcement, after releasing image with all sulu's work + chromium instead of Iceweasel (faster on N900, less RAM-hungry - to the point, that I use it comfortably as "everyday" browser on my device), and LibreOffice instead of OpenOffice?
All this Squeeze thing seems to be big milestone for ED. qole is less active now in Maemo area (no accusing, no demanding, no rant, no offense, no bad words, etc ;) ), and proper announcement by active member (sulu, most likely, or, in worst case, me) would give control over first page (updates), + probably bring fresh, well-deserved attention to Easy Debian. At least, IMHO.
I second this!
Thanks a lot Sulu, some nice suggestions here.
I haven't tested (nor checked for backdoors) them yet but I think these two additions for chromium respectively iceweasel might be useful:
http://www.chromeextensions.org/appearance-functioning/chrometouch/
https://addons.mozilla.org/en-US/firefox/addon/grab-and-drag/
As for chromium extensions - I've already incorporated "Scrollbars anywhere", which is (AFAIK) most customizable "touch scroll" thing for Chromium. Using it, I was able to enable "drop and drag" (also on text/images), while still retaining capability to select text/drop page elements/whatever. I'm pretty sure everyone will find fancy customization mix for own liking. I also checked it for backdoors, at least as much as my limited skills permit ;)
Unfortunately each of those profiles has to be edited separately, so if you want to add another icon to your panel you have to do it in all three profiles if you want them do be consistent
Of course, such editing can be done from (ED) desktop itself, not only by editing config files by hand? I'm going to check it myself, of course, but I think that writing answer here may be useful for further readers.
---
armhf seems to be great idea, yet I agree, that we should wait a little for at least some* maturity from HF versions. Of course, if someone want to volunteer, and prepare such images ASAP, I'll be glad to test them, nevertheless.
*not that I have a clue about kernel support for it, or whatever ;)
/Estel
Since I believe* the existing Maemo kernel makes best use of the hardware, it should already be armhf compatible. No idea if that's actually necessary or if an armhf-compatible libc would suffice, but if it is necessary that shouldn't be a problem.
I don't think our kernel was compiled with hardfp, as obviously this would have implied that the whole of Maemo would have also been compiled like that.
If we somehow were to replace all packages (incl. glibc) with recompiled versions using hard-float then the kernel would have to be made aware of that (unless the kernel uses a specific, EABI-independent, calling convention for system calls, which I don't think is the case).
Well I think recompiling the whole Easy Debian software doesn't make much sense. If that's what you want to do I think Arch or Gentoo would be better distributions. Plus, recompiling every binary with armhf support would practically mean to create a new architecture on our own. I have thought about that earlier but I didn't find it practical then and I don't think so now.
Well, despite posting in this thread, I did not talk about Easy Debian itself, but debian in general. Heck, I don't even use Easy Debian, but run something similar (used debootstrap for that).
That's where the Debian's official armhf port comes in handy now. Theoretically we could start building armhf images right away. The reason why I think there will be a delay of over a year is that it doesn't make much sense to base Easy Debian on Testing or even Unstable since it's supposed to be useable for everybody, not just for Debian experts (heck, I don't even consider myself to be one). Therefore I think the Wheezy freeze is the first date that might make an armhf Easy Debian useful.
At least my plan is/was to use debootstrap to install the armhf debian port, but then if my guess about the kernel is right, it would not work under chroot, but would need its own kernel (should be no problem thanks to u-boot). But then you can forget about all the maemo things that work while in a chroot, such as making phone calls :)
Re. HardFP
If anyone has time and wants to play around. Make a new partition (or loop-mounted file), and try following this:
http://wiki.debian.org/ArmHardFloatChroot
Report your findings :)
I've finished preparing my ED partition for uploading. It contains all discussed tweaks (LibreOffice instead of Open Office, Chromium instead of Iceweasel, fixed pulseaudio) + few more. I'm now tar.gz'ing it.
I've spent few hours adjusting GUI settings and panel configurations, including portrait ones (which is quite funny, as I don't think I'll ever use that, but, whatever...). sulu's panels were totally depending on his applications (no offense, as he never stated, that it's looking otherwise) - on my image, being qole's v3e, dist-upgraded to Squeeze, with tweaks mentioned before - only 3 icons were pointing to existing things ;) so I decided to re-create them from scratch.
---
In between, I got an idea, that allow us to save precious screen space, and have it all (namely, 800x480 pixels) available for programs, *without* sacrificing existence of panels. Configured to automatically hide, when setup properly, they pop-up on tapping just *outside* the screen. They're perfectly functional, and just disappear when not needed - popping up, they're also on top, so aren't covered by any application, despite it use all available screen.
It require (very) little practice, and, overall, getting used to, but works great. Added benefit, is that we can have larger icons (etc) in panels, to the point of (almost) finger-friendly - popping up and hiding on request, their size doesn't matter.
Honestly, it turned out to work so good (after little tweaking), that it feels just like having 2 new buttons outside N900 screen :) It may be hard to describe by writing, but You'll get what I mean. Just remember, to tap a little bit "outside" screen, on panel's side (tapping need to be a little bit stronger than usual).
Rationale:
Technically, we're using characteristic of resistive touchscreen - it's much less responsible on edges, so tapping "outside", result unavoidably on (registered) short, yet quick "swipe" from inside screen to the edge. Normally, it may be considered con, but in our case, it emulates moving mouse cursor "to the very edge".
On portrait with "docked" xvkbd, I've set top panel to be always visible - it use space from xvkbd head's border (not used for anything), so no reason to hide it.
---
I wasn't able to find sulu's orientation.desktop anywhere, so I created one from scratch, using CSSU's image from orientation-lock applet - it fits perfectly as icon for rotating. Of course it's placed on top panel, being easily reachable.
The only glitch I've found, is that after rotation, desktop background doesn't refresh properly (I've set it up a way, that it *should* look good on both landscape [no change here] and portrait) - after "forcing" refresh (by changing desktop display setting, and changing it back), it looks OK, but fail to do so automatically.
Of course, it's *very* minor thing, yet, if someone know good way to force desktop's image refresh in LXDE (kill? What to kill?), I can still update orientation script inside tar.gz to do so.
---
I'll upload it as tar.gz - so, perfectly ready for unpacking in dedicated ED partition (will post appropriate command + which file tweak to make chroot look @ for partition, not image file). I can also create image file, if someone tell me how to prepare one accepted by ED, step by step (never used images, and I'm not interested in doing so, due to *high* performance loss when using partition images vs real partition)
/Estel
I don't think our kernel was compiled with hardfp, as obviously this would have implied that the whole of Maemo would have also been compiled like that.Would it? As far as I understand it an armhf kernel should be able to execute programs compiled for armel. That would mean that the userland doesn't have to be recompiled necessarily, but it would benefit from it.
At least my plan is/was to use debootstrap to install the armhf debian port,That was my idea too. But unfortunately I won't have time for that until late next week. So if you have any plans of doing it, please do it!
btw: Do you have any application in mind that might be a good benchmark for an armel vs. armhf performance test (ideall without having to install an X server in the chroot)? I guess video encoding might be a good benchmark).
I've spent few hours adjusting GUI settings and panel configurations, including portrait ones (which is quite funny, as I don't think I'll ever use that, but, whatever...).Sounds familiar. ;)
In between, I got an idea, that allow us to save precious screen space, and have it all (namely, 800x480 pixels) available for programs, *without* sacrificing existence of panels. Configured to automatically hide, when setup properly, they pop-up on tapping just *outside* the screen. They're perfectly functional, and just disappear when not needed - popping up, they're also on top, so aren't covered by any application, despite it use all available screen.That tapping outside thing sounds ingenious if it works as you described (not sure if the screens of all the N900s are equally responding). That would actually allow to have four panels, one on each screen edge.
I wasn't able to find sulu's orientation.desktop anywhere, so I created one from scratch, using CSSU's image from orientation-lock applet - it fits perfectly as icon for rotating. Of course it's placed on top panel, being easily reachable.You would have found it in /usr/share/applications of my image. However, whatever .desktop file you created it's surely better than mine since mine was really just made quick&dirty.
The only glitch I've found, is that after rotation, desktop background doesn't refresh properly (I've set it up a way, that it *should* look good on both landscape [no change here] and portrait) - after "forcing" refresh (by changing desktop display setting, and changing it back), it looks OK, but fail to do so automatically.Try calling:pcmanfm -w <image file>I don't know if it works but that's what's written in the manpage.
If it works, add that command to the orientation script. I never bothered with that since I use no wallpaper.
I can also create image file, if someone tell me how to prepare one accepted by ED, step by step (never used images, and I'm not interested in doing so, due to *high* performance loss when using partition images vs real partition)Example for a 4GB image:dd if=/dev/zero of=<filename> bs=4M count=1024
mke2fs <filename>Then just mount the file and copy your data.
guilledoc
2012-01-25, 13:01
I´m waiting for Estel new image to trie it opfully with instruction to make a partition or image that I´ve allways used
Thanks
rafael2k
2012-01-25, 13:33
I'm using a debian sid armhf chroot on my N9, and its running fine.
It does not have all squeeze packages yet, but I already installed all the devel environmnent, qt, pidgin and so on.
Harmattan itself is hard float abi.
I´m waiting for Estel new image to trie it opfully with instruction to make a partition or image that I´ve allways used
Thanks
I'm trying sulu's suggestions now (thanks, sulu!), so in meantime - to use ED on dedicated partition, You need to - obviously - have dedicated partition ;) You can just create one on microSD card, but best approach - performance wise - is to left microSD for swap, and create partition for ED on eMMC.
Rationale:
The worst things for performance, when talking about swap on flash medium, are I/O collisions. Having swap and /home/ + opt in the same (physical) storage is bad. Having /home/ + opt on one medium, and swap + ED on another is better, but You still generate I/O conflicts, while using ED.
On the other hand, having /home/ + opt and ED on one device, while swap on another is *much* better - normally, You would not create many I/O conflicts between /home + opt and ED (most likely, you won't be doing intensive I/O from both Maemo and ED), and even if, it still doesn't affect SWAP, which sit on microSD (another medium).
As for repartitioning eMMC, there is wiki article about it - basically, it's bets to install backupmenu or backupmenu-multiboot (if You don't have it yet...), then connect N900 in backupmenu's mass-storage mode (read&write) to linuxbox - it may be any system run with linux liveCD. Then, You change partition layout using GUI tools.
Recommended size for Squeeze-based Ed is as least 3GB. It should be sufficient, but if You plan to install plentora of huge things (on top of already installed LibreOffice, GIMP etc), You may think about 4GB. I'm using 3GB and I'm still using only 1,8 GB ;)
---
When You have free partition ready, installation would be as simple as untar'ing one (big) file, and modifying small ED script.
/Estel
Try calling:pcmanfm -w <image file>I don't know if it works but that's what's written in the manpage.
If it works, add that command to the orientation script.
Guess what, it works :)
I would say "ready for prime time", but I've noticed strange, worrying issue. Many times (most), after tapping the screen, LXDE register it as "still clicked", despite fact, that stylus isn't touching screen anymore. Thus, it enables "right-click" when not asked for (cause long-press works like that).
It's not very easy to notice it at beginning, but there is easy way of testing it - open xvkbd, terminal, and try to write something, by taping on virtual keyboard. Tap any letter, then, without touching screen any more, wait for a while. In my case, letter auto-repeat, just because cursor is still "over" letter on vkb.
Sulu, could You check, if it also happens in Your case? Any way, ideas what may be causing this, anyone?
Of course, I've tried restarting LXDE, closing ED image etc, to no avail. I don't remember facing this issue before - it *might* have appeared, after installing patched server xephyr. Yet, don't quote me on it, as it's quite hard to notice. Most of the times, you think that long-press menu appeared by accident, and while I was irritated by it few times today, I acknowledged bug only, when, accidentally, deleted top panel and canceled "this dialog is needed to gain keyboard focus in LXDE", so I was forced to use terminal and xvkbd to properly shutdown LXDE. Writing on virtual keyboard clearly indicated bug, by auto-repeating letters (they auto-repeat on moderate speed, when letter is keep pressed), that's why I suggested checking it this way.
That tapping outside thing sounds ingenious if it works as you described (not sure if the screens of all the N900s are equally responding). That would actually allow to have four panels, one on each screen edge.
On 2 devices that I've tested, it works flawlessly - 100% of times - by tapping every edge of the screen. 4 panels are 100% possible - it's just me, who use 2 panels and still have much free space on them - my invention on filling it with content seems small :p
/Estel
It's not very easy to notice it at beginning, but there is easy way of testing it - open xvkbd, terminal, and try to write something, by taping on virtual keyboard. Tap any letter, then, without touching screen any more, wait for a while. In my case, letter auto-repeat, just because cursor is still "over" letter on vkb.
Sulu, could You check, if it also happens in Your case? Any way, ideas what may be causing this, anyone?I can't confirm this. Starting a terminal and clicking a letter in xvkbd doesn't produce any other letters afterwards.
I'm not sure what might cause this bug but it kind of reminds me of the ctrl button issue that I encountered after my dist-upgrade (see regressions list in ED wiki article). I didn't pay attention to it anymore but now that I checked it sems to be gone. No idea if it has to do with the xephyr backport.
However, can you please try if pressing ctrl stops writing further letters? Just to make sure it's not a coincidence if it works, please try Shift too.
It might also be interesting to know what xev says when this happens.
On 2 devices that I've tested, it works flawlessly - 100% of times - by tapping every edge of the screen. 4 panels are 100% possible - it's just me, who use 2 panels and still have much free space on them - my invention on filling it with content seems small :pI can confirm that it works for me too.
But I noticed one minor problem: If there is an icon in the panel where the mouse cursor will be placed, that icon will be launched. So you'll have to take care to click only areas where no icons are in the panel.
Four panels might be overkill, but I think three could be useful:
The two long sides for an application launcher panel with big icons and a task bar panel. One of the short sides for status indicators like a clock or notification area.
Have you been able to test the iceweasel/chromium touchscreen addons yet?
guilledoc
2012-01-25, 15:57
@Estel
So are we seeing your new ED img. soon?
However, can you please try if pressing ctrl stops writing further letters? Just to make sure it's not a coincidence if it works, please try Shift too.
It might also be interesting to know what xev says when this happens.
I've tested ctrl, shift, and blue arrow (also thought about it) - bug still exist.
Yet, thanks to Your suggestion with xev, I've noticed something, which makes this bug even more strange. Xev doesn't have anything to tell about it - just like it doesn't happen (register one click). Then, trying it again with xvkbd, it seems that I was wrong initially - it *isn't* long-press. Long press looks different - xvkbd key becomes "black" (virtually pushed down), and keys repeat quite fast.
In case of my bug, after initial press, xvkbd doesn't look pressed at all, and keys keep repeating at quite slow rate (I called it "moderate" in last post). It looks, like just having cursor over xvkbd key, makes it repeat.
On the other hand, if I *really* long-press some xvkbd key (making it repeat quickly), and then lift stylus (stop it, not touching screen anymore), keys are *not* repeating - of course, until I tap on other key. So, long press before, negates strange "repeating" presses.
I have no fckn clue about what is happening here...
BTW - on other parts that xvkbd, it isn't 100% reproducible - most times, tapping is recognized properly (as one click) - yet, when I try some clicks repeatedly (for example, switching between 4 available LXDE desktops), it start to appear as long-press (submenu appear 0.5 second or so after tap).
Trying the same on xvkbd, I can *never* make it to register "real" long-press wrongly (i.e, key *never* look like pressed - black, pushed down, etc) - of course, unless i long-press it really.
Very strange, but, overall, for other scenarios that using xvkbd, it isn't entirely critical, so I'm going to publish my image - I wonder what other people reports will be. Still, any clues *highly* appreciated.
Have you been able to test the iceweasel/chromium touchscreen addons yet?
As for chromium extensions - I've already incorporated "Scrollbars anywhere", which is (AFAIK) most customizable "touch scroll" thing for Chromium. Using it, I was able to enable "drop and drag" (also on text/images), while still retaining capability to select text/drop page elements/whatever. I'm pretty sure everyone will find fancy customization mix for own liking. I also checked it for backdoors, at least as much as my limited skills permit ;)
So, I'm using Scrollbars Anywhere (Chromium) from the very beginning, and it's working great. As for Iceweasel, I haven't tested it - I'm not using it any longer.
HOLY SH*T!
Quote from Easy Debian wiki:
xvkbd
In LXDE one can also use the xvkbd utility to get a virtual keyboard. If clicking on it with the stylus keeps repeating keys indefinitely, go to the xvkbd menu (bottom left of keyboard), choose "Property", and set "Automatic Click" to "OFF".
*facepalm* Ok, this was "oh sh*t!" moment, after trying to hammer this "bug" for few hours. No idea why it turned on, and I came across wiki mentioning it by pure accident (luck).
So, latest known bug was hammered, I'm creating image and uploading it...
BTW, sulu, do You know any way to enable clipboard sharing between Maemo and ED? It' really PITA, not being able to ctrl +c on ED and paste on Maemo, or the opposite.
/Estel
// Edit
But I noticed one minor problem: If there is an icon in the panel where the mouse cursor will be placed, that icon will be launched. So you'll have to take care to click only areas where no icons are in the panel.
After some practice, it becomes natural to bring panel without clicking icon, even if it's just under the cursor - matter of lifting stylus on the correct moment :p
Hi Estel, I'm not at home right now, but when you get your image ready (please localize for US and include any /home/user files) post it and I'll upload it to qole.org and add it to the installer. If it works for others, I'll put it at the top of the list.
I think it would be a great idea to start a new thread around Estel's image.
Thanks, qole! I localized it to US, packed image with lzma and uploaded, but (as I did it before reading Your post) I'm not sure about the (non-image inclusive) files from '/home/user/'. Only two relevant things that I've found are .chroot and .xbindkeysrc - both are highly modified to my liking (.chroot pointing out to dedicated ext4 partition and with temp size changed from 6MB to 30MB; .xbindkeysrc without TAB and ESC, as I've them binded via Maemo's RX-51). Should I add "vanilla" copies of these to image and repack it? Or have I missed something, and You're talking about other files?
---
Anyway, for adventurous ones, here it is:
http://lorienart.pl/cover/ed-squeeze-final.ext3.lzma
BIG FAT WARNING! Despite filename, it may not be "final" yet - minor things like (mentioned in this post) files from Maemo's /home/user/ may be missing. No big deal for advanced ED users, but everyone else - please wait for possible final "polishing" in line with qole's advices
---
qole, or sulu - if any of You recall out-of-image files, that I possibly missed, and/or test this image and find lacking things, I'll gladly fix it, re-pack and reupload image. Of course, if anyone is willing to fix it her/himself and repack/reupload instead of me, it's perfectly fine.
/Estel
I certainly can't keep up with all the wonderful developements on this thread. Special thanks to Estel, Maartenmk, and Sulu. Couple of questions:
1)In regards to dedicated partition: I just partitioned my mircoSD to have ext3,ext4 and the rest fat32 using a live gparted I happened to have discovered on a linux demo disc I had. First time I ever did anything like this so I was quite pleased until I read Estel's recommendation that it is better to make a dedicated partition on eMMC and create swap space on microSD. So could anyone comment on the improved performance, in a quantitative way, if one goes this route. For example, maartenmk got nearly a factor of two improvement when using dedicated partition on SDcard (see post 2324). Can I expect much better improvement having dedicated partition on eMMC. It would be nice to have similar benchmarks for eMMc case. Anyways, if I followed Estels suggestion, I'd have to do some major house cleaning so I'm reluctant at this point. But significant performance improvement will persuade me to try sooner rather than later...
2)In regards to how Estel's image is packaged with extra files and so forth, I would suggest that all this be done such that it as consistent with current form of Easy Debian. It might help to merge some of Estels changes in chroot and .xbindkeysrc but perhaps, if possible, keep them as comments that ppl can incorporate as desired. For example not everyone will have ext4 (either on eMMC or microSD). Anyways, just a suggestion, Perhaps doing things this way is taking too many steps backwards... Maybe what I need is Easy Estel!
3)Getting back to partitions, When using the Synaptic Pkg Mg, that came with Easy Debian, I happen to notice, in the Debian repositories, the presence of gparted! Has anyone actutally tried this on the N900? I downloaded it and noticed that it needs a superuser password to use but that is circumvented by using the sudo command. It would be interesting if, using ED on microSD that you could treat it as a live gparted to repartition the N900 or if ED is on eMMC that you could repartition a microSD. I will certainly give the latter a try tonight.
[DarkGUNMAN]
2012-01-26, 16:57
I have successfully installed Gparted, working fine on a standard ED image after enabling squeeze repo and forcing update using apt-get, also installed dosfstools to enable fat/fat32. Added a link to the Maemo menu to call it directly.
maartenmk
2012-01-26, 21:19
Mscion, while my measurements were accurate, I believe I did not have swap on SD yet. If you do, your ED running from an image on eMMC could well be already faster than what I started with, and therefore you may not see as big a difference.
After switching my swap to SD I have not done any new benchmarks, but I am almost sure ED became slower, for the reason Estel explained.
Moving ED to eMMC should solve that, so I am guessing you will see similar or slightly better performance to what I had when doing the measurements.
I just reflashed, and I don't have ED re-installed yet (I was waiting for news on Estel's image), so I can't check atm.
But I do intend to install it again, and since these measurements are pretty much the only useful thing I have to contribute, I will try to do some new ones.
;1156124']I have successfully installed Gparted, working fine on a standard ED image after enabling squeeze repo and forcing update using apt-get, also installed dosfstools to enable fat/fat32. Added a link to the Maemo menu to call it directly.
I can confirm it works! Repartioning my microSD was real easy using the debian Gparted on N900. I remember being so impressed that I could compile and run c and fortran codes using Easy Debian on N900. Now this! Awesome!
Question:
If you use easy Debian from the microSD can you then run Gparted to repartion eMMC? My guess is that it is much more complicated than I think but could someone chime in and explain why... Thanks!
If you use easy Debian from the microSD can you then run Gparted to repartion eMMC? My guess is that it is much more complicated than I think but could someone chime in and explain why... Thanks!
Why not?
# gparted /dev/mmcblk0
(assuming that your /dev/* is copied/bind-mounted on ED).
michaelxy
2012-01-27, 17:42
Thanks very much to you all :-)
1)In regards to dedicated partition: I just partitioned my mircoSD to have ext3,ext4 and the rest fat32 using a live gparted I happened to have discovered on a linux demo disc I had. First time I ever did anything like this so I was quite pleased until I read Estel's recommendation that it is better to make a dedicated partition on eMMC and create swap space on microSD. So could anyone comment on the improved performance, in a quantitative way, if one goes this route. For example, maartenmk got nearly a factor of two improvement when using dedicated partition on SDcard (see post 2324). Can I expect much better improvement having dedicated partition on eMMC. It would be nice to have similar benchmarks for eMMc case. Anyways, if I followed Estels suggestion, I'd have to do some major house cleaning so I'm reluctant at this point. But significant performance improvement will persuade me to try sooner rather than later...
While moving from ED-on-SD to swap-on-sd:ED-on-EMMC, as a rule of thumb, I would not expect *as much* improvement in performance, as with moving from image file to dedicated partition. Yet, if You're still using swap on eMMC, by moving it to microSD You're going to feel overall increase in performance for *whole* device, so it's definitely worth the effort. Especially, if You'll also add to this customization of few swap related settings.
In Your situation, as a quick&dirty improvement, I would just "cut" 768 MB from microSd fat32, and create swap there (so, use both Ed and swap on microSD). When time permits, and You'll be able to "clean" some thing on eMMC, You can always "kick" Your dedicated partition into eMMC.
2)In regards to how Estel's image is packaged with extra files and so forth, I would suggest that all this be done such that it as consistent with current form of Easy Debian. It might help to merge some of Estels changes in chroot and .xbindkeysrc but perhaps, if possible, keep them as comments that ppl can incorporate as desired. For example not everyone will have ext4 (either on eMMC or microSD). Anyways, just a suggestion, Perhaps doing things this way is taking too many steps backwards... Maybe what I need is Easy Estel!
:o Thanks, but it wouldn't be possible without sulu messing with dist-upgrade and it complications (I could not fix pulseaudio etc issues in thousand of years...) and of course, qole bringing us ED and easy chroot as whole, on first place.
Anyway, no worries. Image that I've uploaded is (well, at least *should*) be independent of any "personal" tweaks in .chroot and .xbindkeysrc. Also, image was mkfs'ed as ext3, so it can be used by people without kernel-power as well, while still regaining possibility to mount is as ext4 - yet, keep in mind, that some 'special' ext4 benefits aren't available when doing so, see wikipedia article about ext filesystem. I concluded, that people using ext4 will be (most likely) sitting on dedicated partition anyway, so they can copy content of image file to any filesystem they feel fancy.
3)Getting back to partitions, When using the Synaptic Pkg Mg, that came with Easy Debian, I happen to notice, in the Debian repositories, the presence of gparted! Has anyone actutally tried this on the N900? I downloaded it and noticed that it needs a superuser password to use but that is circumvented by using the sudo command. It would be interesting if, using ED on microSD that you could treat it as a live gparted to repartition the N900 or if ED is on eMMC that you could repartition a microSD. I will certainly give the latter a try tonight.
Nice catch. Well, I'm pretty sure that it wont work if You want to repartition Ed partition from it ;) but I'm sure You know it already. I would also not risk repartitioning that include /home + opt (/dev/mmcblk0p2) - remember, we're still in chroot, so AIUI, it would be almost like repartitioning ED from within ED.
On the other hand, I don't see much reason, why we shouldn't be able to cut&slize /dev/mmcblk0p1 (MyDocs), or MicroSD as a whole. Just be sure, that it doesn't involve moving/resizing mentioned /home/ + opt. Also, be sure to disable swap on microSD, before messing with SD.
Repartitioning with GUI independant from other device would be great, but for full solution, we would need some thing like u-boot loading kernel from microSD + ArchLinux from microSD. Doable, yet I think no one tried this up to date.
/Estel
// Edit
As You all have probably noticed, the file I uploaded is image packed as .lzma - just like any other ED image files. You need to unpack it first (can be done via tar -xf --lzma <file>) - warning, on device, it's prone to same things, that haunted images provided by qole, i.e. device can get unresponsive during unpacking. In my case - using swap on microSD and tweaked settings, there is no such problem - it unpacks gracefully, and device is 100% responsive during unpacking (yet, unpacking takes 100% of CPU, so other programs aren't blazing fast), but in other setup's it may be better to first unpack it on other machine. AFAIK, it's safe to do so under any OS - may be even windoze + 7zip (unpacked image file won't suffer due to windows filesystems not handling permissions - for the same reasons, image file can be held on MyDocs).
While unpacking on device - no matter of setup - it may be good to do it with nice command, so (for example) 'nice 19 tar -xf --lzma ed-squeeze-final.ext3.lzma' - of course, if You already cd'ed to directory, where You want to see unpacked image.
---
After un-lzma'ing, image can be used directly with ED (just remember, to change 'ext2' to 'ext3' in /home/user/.chroot). It's content can be also moved to dedicated partition - to do so, just trivially mount image somewhere ('mount -o loop <imagefile> <mounpoint>), and 'cp -a' it's content to desired place (dedicated partition). Then, modify /home/user/.chroot, to make it point to Your dedicated partition, instead of image file.
---
When qole will review my questions about files from /home/user/, and image will be confirmed to working fine for few brave folks testing it, I'll land on ED installation, as "first" choice. I've also prepared .desktop files for Maemo (to easy launch LibreOffice or Chromium), so maybe I'll be also time for small update of package - if not, .desktop files can be also downloaded separately.
Hello Estel, thanks a lot for making your ED image available and all the work you did. Same goes out to qole and the other ED contributors.
I fail at extracting the lzma archive directly on the n900 through lzma. I think it's not a problem related to your image, but I thought I'd post it here.
When trying to extract the archive the following error shows up: lzma: SetDecoderProperties() error.
Nokia-N900:/home/user/MyDocs/1Files/OS+img# lzma -d ed-squeeze-final.ext3.lzma
lzma: SetDecoderProperties() error
Nokia-N900:/home/user/MyDocs/1Files/OS+img# unlzma ed-squeeze-final.ext3.lzma
unlzma: SetDecoderProperties() error
Nokia-N900:/home/user/MyDocs/1Files/OS+img# tar xf ed-squeeze-final.ext3.lzma
lzma: SetDecoderProperties() error
tar: Child returned status 1
tar: Beende mit Fehlerstatus aufgrund vorheriger Fehler
Nokia-N900:/home/user/MyDocs/1Files/OS+img# tar xzf ed-squeeze-final.ext3.lzma
gzip: invalid magic
tar: Child returned status 1
tar: Beende mit Fehlerstatus aufgrund vorheriger Fehler
Nokia-N900:/home/user/MyDocs/1Files/OS+img#
The command "xzf" forces tar to use gzip, but it should work with "xf". Lzma seems to be broken or wrong configured on my n900.
Though it works with p7zip to extract the archive after renaming to .7z.
Nokia-N900:/home/user/MyDocs/1Files/OS+img# p7zip -d ed-squeeze-final.ext3.lzma
/usr/bin/p7zip: ed-squeeze-final.ext3.lzma: unknown suffix -- ignored
Nokia-N900:/home/user/MyDocs/1Files/OS+img# mv ed-squeeze-final.ext3.lzma ed-squeeze-final.ext3.7z
Nokia-N900:/home/user/MyDocs/1Files/OS+img# p7zip -d ed-squeeze-final.ext3.7z
7-Zip (A) 9.04 beta Copyright (c) 1999-2009 Igor Pavlov 2009-05-30
p7zip Version 9.04 (locale=de_DE,Utf16=on,HugeFiles=on,1 CPU)
Processing archive: ed-squeeze-final.ext3.7z
Extracting ed-squeeze-final.ext3
I'm running CSSU kernel-power49 and have testing and devel repositorys activated.
Nokia-N900:/home/user/MyDocs/1Files/OS+img# lzma -V
LZMA command line tool 4.32.0beta3
LZMA SDK 4.43
a tip for people who aren't so technically inclined with linux (i am one of them :)).
download the file manually, copy it to a laptop and extract it there. it is soooooooooo much faster. then copy it back and edit the easy debian config (/home/user/.chroot) file and you're done in a jiffy.
@eight
Thanks for pointing that out, of course my tar noobines was showing - I've corrected command to not include -z. I wonder why You have this error at all - image was created via aptosid's tar, so maybe it's some version mismatch. If that is the case, I'm probably doomed to PITA lzma'ing on device. Which will probably take a two days or so ;)
In any case, I'm glad that You've found a way to decompress it on device.
By the way - I've used --lzma, as it creates ~450 MB image, vs 650 MB image of gzip - yet, the latter is *much*, much faster, for both compressing, and decompressing, especially on device. What do You think people - it's the 1/3 size save worth the effort?
@Everyone who tested it already - does it work for You as it should? Also, remember, that if You do not like "hiding" panels (again, the easiest way to pop-up them, is to tap (a little harder than usual) just *outside* the screen, at corresponding edge - it's just like having more hardware buttons on N900 ;) ), this feature can be disabled via right-click (or long-press) on panel, then, in last tab, unchecking "hide panel". Their sizes can be also easily adjusted - I use relatively big panels, as they're hiding when unneeded anyway, and big ones are easy to click, even with fingers. Yet, everything is customizable.
/Estel
michaelxy
2012-01-28, 06:10
I get the error:
Mount failure!
/media/mmc1/ed-squeeze-final.ext3 failed to mount on loop0
mount: mounting /dev/loop0 on /.debian failed: Invalid argument
my old ED image runs normal. I unzip it with 7zip on w7.
You're trying to mount it via /home/user/.chroot, or via hand? If first option is true, have You modified .chroot to mount as ext3, not ext2? If both true, Could You post here Your .chroot file?
Anyone else having this problem?
/Estel
maartenmk
2012-01-28, 16:01
I have exactly the same problem, mounting by hand also failed. Here's my .chroot
# Sample config for chroot
# Device or image containing Debian filesystem.
# Default: first in /home/user/MyDocs/debian*.img*, /media/mmc1/debian*.img*
# Some examples:
IMGFILE=/home/user/MyDocs/ed-squeeze-final.ext3
#IMGFILE=/media/mmc1/debian-squeeze-m5.img.ext2
#IMGFILE=/dev/mmcblk1p2
#IMGFILE=/dev/mmcblk0p4
# Filesystem used; must always be set when using a partition.
# Default: from extension of IMGFILE, or ext2.
IMGFS=ext3
# Mount point for Debian.
# Default: /debian
CHROOT=/.debian
# New /tmp dir size for printing / PDF creation
# Default: 6M
#TMPSIZE=6M
# Debian user to drop privileges
# Default: user
#DEBUSER=user
Got stuck at the same point.
Nokia-N900:/home/user# mount -o loop -t ext3 /home/user/MyDocs/1Files/OS\+img/ed-squeeze-final.ext3 /mnt/ed-img/
mount: mounting /dev/loop0 on /mnt/ed-img/ failed: Invalid argument
And I did some research on the lzma error. Tested a lzma archive from qole.org which extracts without the error.
Nokia-N900:/home/user/MyDocs/1Files# lzma -d maemo-sdk-v1_2.img.ext2.lzma
...
Edit:
Well, this reboot problem below was caused by setting smartrefex (vdd1 sr) enabled in qcpufreq.
Disabled VDD1 and now works well, even overclocked.
*-----------------------------------------------------------------------------------------------------------------------------------------------------*
But then the N900 reboots after a few seconds till minutes after starting lzma extraction.
If this happens /dev/mmcblk0p1 (MyDocs) gets read-only. Mount looks normal but dmesg shows it and you'll get an error like this:.
Nokia-N900:/home/user/MyDocs/1Files# lzma -d maemo-sdk-v1_2.img.ext2.lzma
lzma: Cannot open output file maemo-sdk-v1_2.img.ext2
Nokia-N900:/home/user# mount
...
/dev/mmcblk0p1 on /home/user/MyDocs type vfat (rw,noauto,nodev,noexec,nosuid,noatime,nodiratime, utf8,uid=29999,shortname=mixed,dmask=000,fmask=013 3,rodir)
Nokia-N900:/home/user# dmesg
...
[31237.416351] FAT: Filesystem error (dev mmcblk0p1)
[31237.416381] fat_free_clusters: deleting FAT entry beyond EOF
[31237.416412] File system has been set read-only
If this happens at your N900 you can fix it this way:
- umount /dev/mmcblk0p1
- fsck and repair /dev/mmcblk0p1
- remount it to MyDocs
Nokia-N900:/home/user/MyDocs/1Files# cd ../..
Nokia-N900:/home/user# umount /dev/mmcblk0p1
Nokia-N900:/home/user# fsck -r /dev/mmcblk0p1
fsck 1.41.3.maemo0 (12-Oct-2008)
dosfsck 2.11, 12 Mar 2005, FAT32, LFN
FATs differ but appear to be intact. Use which FAT ?
1) Use first FAT
2) Use second FAT
? 1
Reclaimed 1024 unused clusters (16777216 bytes).
Perform changes ? (y/n) y
/dev/mmcblk0p1: 5812 files, 903148/1277519 clusters
Nokia-N900:/home/user# mount -o rw,noauto,nodev,noexec,nosuid,noatime,nodiratime,u tf8,uid=29999,shortname=mixed,dmask=000,fmask=0133 ,rodir /dev/mmcblk0p1 /home/user/MyDocs/
This reboot even happens with nice -n 19 [command]. First I thought it could be due to overclocking, but happens also at 600 Mhz.
It seems, that archive get somehow corrupted during upload - my local copy works, but when I downloaded it, I'm also unable to unpack. Doing so like eight described result in false-positive unpacking, but image is unmountable.
Sorry for inconvenience - will upload again, yet, this time, I'm going to download it myself and check, before dropping link here ;)
/Estel
maartenmk
2012-01-29, 13:52
I repeated the tests I did in posts 2324&2329 (opening a file in OOCalc directly from Maemo, editing a minimal amount, and closing OOCalc again), now concentrating on the locations of both swap & ED partitions. The setup is as similar as possible, but can't be fully compared of course.
Each test was run 3 times. The numbers behind each line are the times in seconds:
- Swap on emmc, ED on sd: 95,105,90
- swap on emmc, ED on emmc: 95,75,75
- swap on sd, ED on sd:95,80,75
- swap on sd, ED on emmc: 80,70,65
I then overclocked to 900mhz, and ran the last test twice more:80,60
So as expected 'swap on sd & ED on emmc' is the fastest, and overclocking helps, but not a lot.
I can't really explain the second place for both on emmc though. I am also surprised about the bad performance of the first setup. But the conclusion is clear enough for me.
When Estel's image is back up, I can try the same sequence again, so we will have some idea on the performance of Libre-office.
I got pissed a little by incompatibilities between aptosid tar --lzma and Maemo tar --lzma implementations, so I decided to do everything on device. I'm already during cp -a everything to image file. Then, I'm going to --lzma it, so it will probably take a week or so :P (seriously though, I expect it to be finished tomorrow). After that, I'll upload it again, then download, unpack, and test. If correct, will drop info here.
/Estel
Holy s*** again! Of course, it's my tar noobines again. The file inside .lzma archive isn't corrupted, it's just .tar with wrong name. So, after un-lzma'ing, just untar it (plain tar, without compression), and you should get final, working .ext3 file.
I'm going to fix that in uploaded file anyway, but - people that already downloaded it, and haven't deleted yet - I've confirmed that it's 100% working after un-lzma'ing *and* un-tar'ing after.
/Estel
BTW, sulu, do You know any way to enable clipboard sharing between Maemo and ED? It' really PITA, not being able to ctrl +c on ED and paste on Maemo, or the opposite.It would require to exchange clipboard contents between different xserver displays an unfortunately I have no idea how to do this.
I repeated the tests I did in posts 2324&2329 (opening a file in OOCalc directly from Maemo, editing a minimal amount, and closing OOCalc again), now concentrating on the locations of both swap & ED partitions. The setup is as similar as possible, but can't be fully compared of course.
Each test was run 3 times. The numbers behind each line are the times in seconds:
- Swap on emmc, ED on sd: 95,105,90
- swap on emmc, ED on emmc: 95,75,75
- swap on sd, ED on sd:95,80,75
- swap on sd, ED on emmc: 80,70,65
I then overclocked to 900mhz, and ran the last test twice more:80,60
So as expected 'swap on sd & ED on emmc' is the fastest, and overclocking helps, but not a lot.
I can't really explain the second place for both on emmc though. I am also surprised about the bad performance of the first setup. But the conclusion is clear enough for me.Some time ago I made some benchmarks between my class 10 microSD card and the eMMC. The overall result was that the eMMC is about 1.3 times faster than the microSD card. I guess that might explain the results.
maartenmk
2012-01-29, 19:18
I can confirm the image file works after untar-ing. And it, or at least Libre-Office Calc) is indeed significantly faster, the same test as in post #2560 gave these results:
- swap on sd, ED w. Estel's image 'installed' an emmc partition: 75,55,50
(previously 80,70,65)
The UI also reacts better, before I sometimes had to click the menu bar twice to open a pulldown menu.
@ Estel: it looks like the Chromium extensions you installed did not carry over into the image file.
Edit: test results added
It would require to exchange clipboard contents between different xserver displays an unfortunately I have no idea how to do this.
Is this of any help sulu? Seems like an almost identical situation unless I've missed something:
http://ubuntuforums.org/showthread.php?t=1431913
Is this of any help sulu? Seems like an almost identical situation unless I've missed something:
http://ubuntuforums.org/showthread.php?t=1431913Nice find! xsel seems to be exactly what we need.
I just tried the script from ubuntuforums.org and at least for text it works. I also tried to copy images with it but unfortunately that didn't work. Instead I only got a single dot (the script's default empty content) when trying to insert an image. I'm not sure if the problem was on the copying or on the pasting side.
You'll find an optified version of Debian Lenny's xsel package (I just moved the files around, so there's no change in the source code) attached to this post. Install it under Maemo and follow the instructions at ubuntuforums.org!
Thanks for testing, guys!
Last night I've uploaded fixed version - this time, it was lzma'ed on device. BTW, it's awesome how much difference using swap on microSD and correct swap-related settings does - *without* using nice, I was able to do other things on device during whole compressing, + I got "surprise" call, and was able to pick it up + talk without any single glitches. Whole process took only few hours (compressing 3GB file).
I'm announcing it just today, as I've first downloaded it, unpacked (also on device, so I'm 100% sure it will work using lzma -d, or ED installer), and tested.
The link haven't changed (I just replaced file):
http://lorienart.pl/cover/ed-squeeze-final.ext3.lzma
It can be mounted as ext3 and ext4 flawlessly. I've also successfully mounted it as ext2 and enabled LXDE, yet, haven't performed any further ext2-related testing.
/Estel
It can be mounted as ext3 and ext4 flawlessly. I've also successfully mounted it as ext2 and enabled LXDE, yet, haven't performed any further ext2-related testing.
/Estel
Thanks - i'm currently using it with ext2 :)
As for lack of Chromium extensions in resulted image, I've traced down the problem - for some reasons, using workaround for stupid Chrome Store behavior ("developer mode" for installing extensions), it's not possible to keep them on Debian partition (dialog, where Your provide path to extension folder, can't point to hidden directories, starting with "dot", and our mount is ".debian").
No big deal, though - those are small files, and one can keep them - for example - in /home/user/chromium-extensions. On how to download and load in developer mode (need to be done only once), follow:
http://talk.maemo.org/showpost.php?p=1147612&postcount=2495
for Scrollbars anywhere, You can also download from this instead:
http://lorienart.pl/cover/scrollbars-anywhere.tar.gz
...untar, and jump directly, to part of instructions, that explain loading in dev mode. This is exact copy of my extension setup (although, I'm pretty sure, that You'll need to load and configure them anyway).
For best effects, in Scrollbars Anywhere options, use following (leave rest unticked):
Button: Left
Tick "Use grab and drag style scrolling"
Top speed: 20000
Friction: 1
This will enable scrolling page like in microB, yet, You'll be able to easily select text holding ctrl. Also, ctrl + click open link in new tab.
Overall, by *unticking* "use system border and title bar" (*right click on upper Chromium bar", You got effect like this:
http://lorienart.pl/cover/chromium.png
To easily navigate through opened tabs, use ctrl + number (control+ blue arrow + number). With possibility of binding left alt somewhere, we could use that to achieve back & forward - for now, one need to pres buttons. New tab is ctrl +t.
/Estel
Today is the first time i used chromium - the first time ever really using a debian browser and it works great!
Some "Must Have" features (for me) were missing on N900's browser - ones that i always have on my PC such as saving all tabs on exit, undo for closed tabs (history) and keyboard control for switching/opening new/closing tabs.
Chromium works really well with several tabs and deals great with hebrew characters which other browsers had problems dealing with.
I also used it with portrait mode and full screen which is very cool.
Estel - This image of debian you prepared is really awesome - i already worked with GIMP, LMMS and LibreOffice and all are working great till now :)
I'm using the image mounted to ext2 partition.
When using Ctrl+ to zoom in, the page does not format correctly and the left part of it is not accessible (not even by scroll bar or dragging). is there any solution for this?
How do you install flash support?
mr_pingu
2012-01-31, 22:40
Anyway to start it from LXDE? Icon seems to be present but it doesn't start chromium actually... As the statusbar of maemo when starting from chroot terminal is taking much screenspace
Also I found scrolling better when
Button: Left
Tick "Use grab and drag style scrolling"
Scaling: 125 - 150%
Top speed: 40000
Friction: 1
And maybe releasing the .desktop files of LibreOffice, chromium or did I miss them, saw you talking about it somewhere though... Yet have to find more issues. Not really an easy debian user before as I always failed with updating... Now let's see if I can get all pentest tools working in ED ;)
maartenmk
2012-01-31, 22:49
You can start it from the menu, at least it works for me.
Also, if you switch off 'use system border and title bar' (see Estel's mail above), the Maemo menu is no longer in the way. Note that it only worked after I had restarted and set the statusbar to semi-transparent. (I don't know if both of these are necessary)
mr_pingu
2012-01-31, 23:28
Hmm my /home/user/.config/chromium/ seemed to be corrupted or something. Deleted that folder and now it's working from LXDE don't know what caused this...
I'm glad You all liked it :)
@impeham
I can't reproduce chromium behavior You've reported, when using "ctrl +" and "ctrl -" . In my case, pages are rendered correctly, and it's perfectly possible to drag page in any direction (yet, I'm using scrollbars anywhere, as I've written in last post - shouldn't matter, in that case). Are You sure, that it isn't related to some particular site?
@mr_pingu
Maybe You've opened Chromium via "Ed root" Maemo icon (one opening terminal as Easy Debian root)? This makes Chromium unusable from user, i.e. from LXDE, or from proper Maemo icon launching program directly. Same apply for LibreOffice.
Generally, using ED root terminal from within Maemo should be used *only* to do "root'ish" tasks - like dpkg-reconfigure locales, etc. In any other case, open regular Maemo terminal, and execute 'debbie'. This result in Ed terminal as 'user'.
---
.desktop files for Maemo - allowing to (hassle-free) launch Chromium and LibreOffice - are available as this post attachment. Untar it to (Maemo) /usr/share/applications/hildon/*
/Estel
*Easiest way to do it, is - as root:
mv location/where/you/saved/desktop-files.tar.gz /usr/share/applications/hildon/
cd /usr/share/applications/hildon
tar -xzvf desktop-files.tar.gz
after that, it's safe to remove tar archive:
rm desktop-files.tar.gz
udaychaitanya16
2012-02-01, 18:49
I will be very greatful is somebody tell me what to do after downloading the image file.I have placed it in MyDocs.Read the wiki ,installed chroot package.When I started the image installer it did not recognise the preloaded image.I dont know how to mount the image.Infact I have debian installed in my laptop and gparted is there.Please tell me in a simple manner.It would be joy to have squeeze.
Thank you.
first of all, You need to unpack img. Use lzma (from terminal, or any GUI program You like) to achieve it. For casual users, I strongly recommend unpacking on desktop/notebook, then transferring image file itself to N900. Without proper swap-related configuration of Your N900, You can have unresponsiveness/reboot problem while unpacking.
Unpacked image is 3GB file (ed-squeeze-final.ext3). Place it somewhere - for example, on MyDocs. Then, edit /home/user/.chroot file (You should have it, as You installed easy chroot), find appropriate line (every option is explained in #comments), and edit it, to point into Your file, i.e. /home/user/MyDocs/ed-squeeze-final.ext3.
Thats all. For more detailed informations, please refer to Easy Debian Wiki page.Also, keep in mind, that using image file, performance won't be amazing - 'cp -a' image file to dedicated partition (little more advanced topic) would give much better results.
/Estel
@impeham
I can't reproduce chromium behavior You've reported, when using "ctrl +" and "ctrl -" . In my case, pages are rendered correctly, and it's perfectly possible to drag page in any direction (yet, I'm using scrollbars anywhere, as I've written in last post - shouldn't matter, in that case). Are You sure, that it isn't related to some particular site?
After using it for a while, i noticed that it sometimes happens and sometimes not - it happened twice for maemo talk, but now it seems to work ok - i'll try to isolate to get what exactly is causing this.
Estel - how do you install the flash plugin? i noticed that it is not included even as disabled in the plugins list...
demolition
2012-02-02, 00:00
Thanks for testing, guys!
I'm announcing it just today, as I've first downloaded it, unpacked (also on device, so I'm 100% sure it will work using lzma -d, or ED installer), and tested.
The link haven't changed (I just replaced file):
http://lorienart.pl/cover/ed-squeeze-final.ext3.lzma
It can be mounted as ext3 and ext4 flawlessly. I've also successfully mounted it as ext2 and enabled LXDE, yet, haven't performed any further ext2-related testing.
/Estel
Please would you point to notes for installation and any "known issues". I can read perfectly well but get a bit lost on new territory - I become unsure if points made are "you need to know this" or merely "isn't interesting that...". The beauty and problem of all fora!
Things like - is it best to install on-device or on-SD? Does it need to be in its own partition or just plonked on the eMMC?
- a pointer to some clear advise that doesn't meander too much, would be greatly appreciated.
As an aside,
- I see it's running chromium as the main browser. Are mozilla ones just too unwealdy?
- Once this edit of ED becomes established, how easy would be to migrate libraries known to work on the N900 to the maemo repos? e.g. ffmpeg v0.7 (or at least 0.55) or RTMPdump 2.3 etc., etc.?
Great work people. Really looking forward to trying this out!
udaychaitanya16
2012-02-02, 08:16
@ESTEL
I am very much pleased with your post.And i admire the fact that you did not tell me "if you dont know this you dont deserve this thing:D.
what I did is, downloaded the image of you and extracted in my linux laptop.Archive manager did all the thing for me.
Then I placed it in /home/user/MyDocs.Installed easy debian chroot package.
Then I had to edit the /home/user/.chroot in leafpad pointing to your image file.
At last I am typing this post in chromium web browser:).
If you could tell me the size of dedicated partitions and locations I would do that to taste the speeds you people are experiencing.
On top of every thing,as qole suggested please start a new thread with a nice tutorial.
Thank you very much for this nice image.
EDIT: where did this thanks button go?
@impeham
Well, i didn't did *anything* to enable flash. It should work out of the box... I use tweakflashversion (from Maemo) to "cheat" sites, that I use flash 10 - no idea, if it is relevant, or not. It was working (for me) with Iceweasel on old image, and it's working in Chromium same way.
I think I would remember doing anything "special" in ancient past, but who knows... Maybe someone else got an idea? Is it working for others, out of the box?
@demolition
As stated many times before in this thread (no problem, i understand, that it all may sounds confusing), best scenario is to use it on dedicated partition. Now, the "best of best" is to use it on eMMC *and* use (Maemo) swap on microSD *only* - no matter what type of card You have, it may be even slowest one. This gives performance benefit for whole Maemo, and to ED too, as added "bonus". If, for any reason, You still want to use swap on eMMC, it may be wise to have ED on dedicated partition - yet, in this case, it depends on SD card performance.
As for "why Chromium" - well, Iceweasel is too heavy/slow behemoth. Chromium is *much* faster - at the same time, providing most useful extensions (lazarus form recovery, scrollbars anywhere, etc), that we can find on Iceweasel.
- Once this edit of ED becomes established, how easy would be to migrate libraries known to work on the N900 to the maemo repos? e.g. ffmpeg v0.7 (or at least 0.55) or RTMPdump 2.3 etc., etc.?
I have no idea. Just keep in mind, that we're chrooting here, so having thing working on ED, doesn't mean that it will work flawlessly in Maemo, and - more importantly - that Maemo components (especially, closed-source) will cooperate with new version.
For both You and udaychaitanya16 - to use ED from dedicated partition, just create one (via any tool, that You feel fancy), then - no matter if from within Maemo, or from desktop PC, with N900 connected and it's partitions reachable:
1. Mount (empty) dedicated partition that You've just created
2. Mount Your ED image (mount -o loop /path/to/ed-squeeze-final.ext3 /path/to/mountpoint)
3. cp -a /path/to/image/file/mountpoint/* /path/to/partition/mountpoint/
4. Edit /home/user/.chroot, to make it point into Your partition, instead of image file (for example, /dev/mmcblk0p4, if Your dedicated ED partition is 4th on eMMC).
That's all. Just remember, that to gain best speed benefits, You should have Your Maemo swap on microSD (with eMMC swap disabled *completely*). Further speed improvements are possible to gain via correct swap-related settings, as - most of the times - ED programs are quite swap-intensive - yet, swap settings are out of scope of this discussion.
/Estel
// Edit
udaychaitanya16 to use "Thanks" button, you need to verify Your email address from profile settings.
Hi,
please help me with the following case. I have repartitioned my internal memory of N900. I made 8G for apps, 18G for MyDocs and remained approx 3G as ext3 for EasyDebian installation. It is /dev/mmcblk0p4 . The question is now how to mount it that Debian Image Installer to see it and to give me chance to install the image there. I have tried to make dir called "EasyDeb" in /home/user/ and to mount /dev/mmcblk0p4 to this dir, but installator did not recognize it as suitable. It gave me again to chose between MyDocs and mmc1. What I am doing wrong?:confused:
Installer doesn't allow You to use dedicated partition - You need to do it manually.
Background story - when qole created it for the first time, repartitioning was considered hi-tech topic (which was a little exaggerated, to say at least - repartitioning using GUI linuxboxes busted this myth), so he focused totally on using Ed via image file.
To be honest, preparing correct partition layout is one of easiest things, that "power user" can do - I wonder, if adding support for dedicated partitions (i.e., automatically extract image content there, as bonus step for installer) would be much hassle to add into installer?
qole, what do You think? Of course, it would be 100% optional.
/Estel
udaychaitanya16
2012-02-02, 18:54
@Estel
worst thing happened to me.I wanted to re partition my emmc and 16 GB sd card.I started with making a live usb of Gparted since I could not un mount FAT partition of N900 from my installed Gparted in my computer.
Before that I read a wiki about partitioning.It suggested to have backupmenu.so I installed that and backed up root and opt.Booted gparted and it could see the partitions on emmc.
Then I did silly,funny partitioning.I gave 3GB for ED and called it debian,3GB with a lable home,768Mb as linux swap.I gave the lables on Gparted only.
In case of sd card I have allocated 1GB space for linux swap.
It was a successful operation until I realized that I could not back up any thing from backupmenu.Then it went into reboot loop.I had to flash it.That was not a problem since I have downloaded image with me always.
After that the first thing I did was to install rootsh.It installed flawless but I could not upgrade it because I dont have enough memory to upgrade it. I could not understand what df -h gave me nor can paste it since I dont have ssh.In the settings menu it told me I have literally 1 KB available.I wanted to have 3 GB available for installable applications and ended up with 1 KB.
I am ready to flash a hundred times.Do I need a backupmenu again?If so how can I install that with my great partitioning scheme?
Please save me from this partitioning trap.
1. Start device in backupmenu mode
2. connect it to gparted live [whatever] and select (on device) "mass-storage mode: read&write".
3. You should see your partition scheme in gparted. Check what5 went wrong, and re-repartition it, when appropriate.
Remember, that You can only restore backupmenu backups for same kernel version, as one that was on device, when You created backup. So, if You backuped with kernel-power from -devel, be sure to install it, before trying to restore backup.
In worst case, You can flash everything (including vanilla eMMC layout), and try repartitioning again.
Also, keep in mind, that You should not use ext4 for home partition, if You're not using kernel-power *and/or* You don't know how to modify backupmenu to work with it. In such case, stay with ext3, You aren't losing much.
---
This is quite off-topic here, so in case of further problems, please find similar thread, or, if such one is lacking, create Your own, them drop me a link to it via PM. I'll be glad to help, but not in ED thread.
/Estel
maartenmk
2012-02-02, 23:03
@impeham
Well, i didn't did *anything* to enable flash. It should work out of the box... I use tweakflashversion (from Maemo) to "cheat" sites, that I use flash 10 - no idea, if it is relevant, or not. It was working (for me) with Iceweasel on old image, and it's working in Chromium same way.
[...]
For me it doesn't work from Chromium, it consitently gives me the error:
"the following plug-in has crashed: /usr/lib/browser/plugins/libdebianflashplayer.so"
For me it doesn't work from Chromium, it consitently gives me the error:
"the following plug-in has crashed: /usr/lib/browser/plugins/libdebianflashplayer.so"
I can confirm this, i can't reproduce videos with chromium. "about: plugins" reports that i have the 9.0 r31 version (usr/lib/browser/plugins/libdebianflashplayer.so).
in maemo, i use tweakflashversion on the default pr1.3 libflashplayer.so, using the 10.1.64 (froyo version) i can reproduce videos perfectly with microb.
is there any way to replace the libdebianflashplayer.so with the libflashplayer.so i have in maemo? i tried with a symlink. it doesn't work, but i have not skills and i've just started with easydebian (i don't even know where all my confs are saved, and i really made a lot of mods, like installing icons, themes and a few chrome plugins, no idea where they are saved tho)
can we just update the debian flashplayer? or replace the default .so? i guess that libdebianflashplayer.so comes inside the img so replace it would be useless or at least not permanent, just asking, i don't want to start another flash 10.1 epic thread
Hello, friends.
Seems that the installation on dedicated partition has to be done manually. Could someone help me with some additional info what I need to do/read. If there is such a information just give me the link and I will read/study it. Then if I have question I will ask, otherwise I will just do it and will post the result/problems I met.
Regards
http://talk.maemo.org/showpost.php?p=1159430&postcount=2579
One page ago...
Estel,
Can you estimate how much was the performance gained by using a dedicated swap partition for debian and installing debian image to a dedicated partition?
i assume you made both on the microSD card (class 10?) - is that right?
i currently mount the image file from the internal MyDocs and not using another swap partition for debian - chromium is not working so smooth but reasonable - if the change in performance is significant i would do it your way.
Thanks for your hard and contributing work!
[EDIT]
i just made a simple test and moved the debian image file to the SD card (i am using class 10) and mounted it from there.
just by doing so, performance is about doubled with all debian applications - libreoffice, gimp and chromium - much much better!
udaychaitanya16
2012-02-03, 20:37
1. Mount (empty) dedicated partition that You've just created
2. Mount Your ED image (mount -o loop /path/to/ed-squeeze-final.ext3 /path/to/mountpoint)
3. cp -a /path/to/image/file/mountpoint/* /path/to/partition/mountpoint/
4. Edit /home/user/.chroot,
1.mounted dev/mmcblk0p4
2.mount -o loop /home/user/MyDocs/ed-squeeze-final.ext3 /media/ed
"ed"is my mount point
3.This got my nerves out.what idid is cp -a /dev/mmcblk0p4/ /media/ed
4.edited .chroot directing to /dev/mmcblk0p4
When I started the debian lxde its a failure.What wrong am I doing there.
I guess you want to copy the estel-image to mmcblk0p4?
First mount the image and the partition mmcblk0p4 to a directory:
mount -o loop /home/user/MyDocs/ed-squeeze-final.ext3 /media/ed
mkdir /mnt/debian-tmp
mount /dev/mmcblk0p4 /mnt/debian-tmp
Then your copy command should look like this:
cp -a /media/ed/* /mnt/debian-tmp/
With your command above, you copy from (empty?) mmcblk0p4 to the mounted ed-image.
----> the general principle is alwas: cp <-opt> <source> <destination>
You can also copy/backup directly from partition (mmcblkXpY) to image file and verse visa with dd.
Example:
dd if=/dev/mmcblk0p4 of=/home/user/MyDocs/part0p4.ext3
dd if=/home/user/MyDocs/part0p4.ext3 of=/dev/mmvblk0p4
There's been a surge in new user questions lately. This surprises me, since I didn't expect there to be new users for a discontinued product (the N900).
My second N900, the one I got after my first one's USB port broke, is looking old and battered.
Is there somewhere you can buy new N900s, or are you trying new things on your old N900s?
i'm from the philippines and found two brand N900's in two cities. it's still possible that there are people here who just bought a brand new device. plus, have you seen reggie's post: talk.maemo.org/showthread.php?t=82019?
there are still lots of regular TMO visitors and we still get spikes from time to time. as somebody mentioned somewhere, "The N900 refuses to die." :)
demolition
2012-02-04, 01:33
@qole
Viva en-nine-hundred... mmm ... not sure that has much of a ring but you get what I mean. Because of its uniqueness, lots of us won't let the N900 die!
As far as buying new one is concerned: yup, it's still possible in the UK, unless all the vendors are spoofs! Getting one from from an actual shop is unlikely. A lot of devices for sale are second hand though. The N900 got a bad repuatation for unreliability and being over-complicated, especially amonst sales staff. I also think Nokia had a bad reputation well before Nov 2009. Therefore, the N900 didn't really sell, especially in zones dominated by carrier subsidies. It's possible that the devices that didn't sell might have been bought up en-masse for resale elsewhere?
As you and many others commented: speed aside, the N900 is unmatched by any other device currently available.
Back to ED: have you tested the most recent image? Any thoughts/comments/etc.?
I'm back at home, and I'm just downloading Estel's image tonight. I'll let you know when things are up and running via the installer.
But again, I strongly suggest Estel takes discussion of his image to a new thread here in Alternatives, leaving this thread for discussion of the scripts and stuff.
I've added Estel's image to the installer. But it is really clear when running LXDE that we need to figure out a way to set up different environments for different images.
I suggest we look into using the embedded startup scripts (http://wiki.maemo.org/Easy_Debian#Startup_scripts) to configure the user's environment when using these contributed images.
Thanks, qole! I waited with new thread, until image appear "officially" in the installer. Do you feel like adding LibreOffice and Chromium .desktop files into the package as well, or do You think, that they should be downloadable from somewhere else (dedicated thread, or via embedded startup script)?
Also, could You elaborate more on:
But it is really clear when running LXDE that we need to figure out a way to set up different environments for different images.
...? I don't exactly get what You mean - what things, that require environmental variables tweaking have You observed? Would love to get it straight before "announcing", as I'm quite a noon when it comes to such things (enviroment tweaking).
/Estel
// Edit
I'm not surprised by high interest in N900-related things at all. First of all, Maemo and N900 have no very high momentum - IMO, higher than in *any* moment of it's history, including release (talking about actual contributions, not chit-chat - see CSSU and kernel-power).
Second, You may agree with it or not, but for many people, N9(50) is a total or almost-total failure. There is (still) no competing device for N900, even on far horizon - latest Nokia products, seems to be targeted on quite different target group - Personally, I would say, that it's close to what WebOS used to be. But, of course, it's off-topic here.
tanpoaran
2012-02-04, 10:13
@qole
how to install printer drivers on easy deb??
i've succeed installing libreoffice yesterday.. it's great, but now i want to install cups and i see this line..
dpkg-reconfigure cups
Stopping Common Unix Printing System: cupsd.
Starting Common Unix Printing System: cupsdFATAL: Could not load /lib/modules/2.6.28.10-power49+pre50/modules.dep: No such file or directory
FATAL: Could not load /lib/modules/2.6.28.10-power49+pre50/modules.dep: No such file or directory
FATAL: Could not load /lib/modules/2.6.28.10-power49+pre50/modules.dep: No such file or directory
tell me what should i do??
thank you
Estel: I mean we could put image-specific configuration files inside the image, and then use the embedded startup scripts (http://wiki.maemo.org/Easy_Debian#Startup_scripts) (probably onfirstchroot-ext.rc to copy the config files into the /home/user directory. Alternatively, the image's onfirstchroot-ext.rc script could unmount the Maemo /home/user directory and use the image's version instead. That's probably the best, it would maintain per-image config files and let the user personalize them.
As for the N9, it is still very early days. There's a bluetooth keyboard app in the Nokia store already, and open mode works well. Interestingly, there's no keyboard focus problems on Harmattan, even though they said the keyboard focus problem wasn't a bug on the N900...
Concept of using config file per-image is great, but when it goes to my image, it doesn't contain any specific thing in /home/user ;) Personally I use modified .xbindkeysrc and different .chroot, but it's not dependent on any particular image, it's just about my preferences (running from dedicated partition as opposed top image file, etc).
I'll try to prepare announcement thread today.
/Estel
// Edit
As per N9(50), time won't change hardware fails - capacitive screen, lack of SD slots, lack of easily-removable battery, etc. Using external keyboard is also hardly a solution for developing/overall productivity on-the-go.
Using Estel's prepared image, when launching Chromium -> about:plugins, this is the list i see - there is no flash plugin - Estel is this what you have and flash is working when you use youtube?
Estel: When I ran it, LXDE from your image wasn't configured very well; the theme, panels, fonts, and background were not set correctly. That's what I meant.
ivgalvez
2012-02-05, 17:52
Estel: When I ran it, LXDE from your image wasn't configured very well; the theme, panels, fonts, and background were not set correctly. That's what I meant.
That's the same thing that has happened to me. I thought it was due to my lack of knowledge around ED as it's the first time I have set it up, but the appearance of LXDE seems unmodified, even the panels are not autohiding or configured for finger use as you explained before.
Maybe it's a wrong image?
@impeham
Nope, in my case Shockwave Flash is also listed - 9.0 r31, provided via /usr/lib/browser/plugins/libdebianflashplayer.so
If you wait a little, we can move this discussion to new thread.
@qole, ivgalvez
Damn, I have no idea why it was looking like that. Thinking about that, it's probably missing configuration from /home/user/.config/lxpanel and similar things.
qole, do you remember which files from /home/user/(.config?) exactly are used by ED? I would upload them, as a "hotfix", for separate extraction.
As a long-term solution, I would use - as You suggested - "onfirstchroot" scripts, to unmount Maemo /home/user, and it would use own ED content. uploaded in new version of image.
Does anyone have idea, how such script should like? Or, maybe better ideas to solve that?
/Estel
The Chromium plugins problem is another example of the missing home directory config files. I've got a few pots on the stove, but I'll try to make a script for you to unmount the /home/user folder in the image...
@impeham
Nope, in my case Shockwave Flash is also listed - 9.0 r31, provided via /usr/lib/browser/plugins/libdebianflashplayer.so
If you wait a little, we can move this discussion to new thread.
Of course i will - this was my next question - seems like chromium related issues :)
udaychaitanya16
2012-02-07, 10:12
I guess you want to copy the estel-image to mmcblk0p4?
First mount the image and the partition mmcblk0p4 to a directory:
mount -o loop /home/user/MyDocs/ed-squeeze-final.ext3 /media/ed
mkdir /mnt/debian-tmp
mount /dev/mmcblk0p4 /mnt/debian-tmp
Then your copy command should look like this:
With your command above, you copy from (empty?) mmcblk0p4 to the mounted ed-image.
----> the general principle is alwas: cp <-opt> <source> <destination>
You can also copy/backup directly from partition (mmcblkXpY) to image file and verse visa with dd.
Example:
dd if=/dev/mmcblk0p4 of=/home/user/MyDocs/part0p4.ext3
dd if=/home/user/MyDocs/part0p4.ext3 of=/dev/mmvblk0p4
Everything worked except copy command.As soon as i issued that command my device froze for sometime and gave some error messages about epiphany.After that It got switched of itself.When I edited chroot and started ED it complained of non existing image in said partition.Anyways Thank you.I will patiently wait for a new thread and dedicated partition related tutorial.
guilledoc
2012-02-08, 15:12
sorry for the silly question but how to run from estel img chromium or libre office from terminal with debbie?
thanks
sorry for the silly question but how to run from estel img chromium or libre office from terminal with debbie?
thankschromium:debbie chromium-browser
libre office writer:debbie lowriter
spreadsheet:debbie localc
michaaa62
2012-02-08, 17:34
Everything worked except copy command.As soon as i issued that command my device froze for sometime and gave some error messages about epiphany.After that It got switched of itself.When I edited chroot and started ED it complained of non existing image in said partition.Anyways Thank you.I will patiently wait for a new thread and dedicated partition related tutorial.
You might want to set 'nice -n 19' before the copy command to ease the burden of the I/O-process like thatnice -n 19 cp -a /media/ed/* /mnt/debian-tmp/
If you are having copy / decompression problems when using the MyDocs directory, you probably need to run the following two commands before trying the copy command:
sudo nr_requests_setup
sudo swapoffon
If that doesn't work, reboot your phone and try again.
Had also problems with reboots while unlzma'ing and found out that on my n900 enabled SmartReflex VDD1 causes this.
Tested it with 900 to 600 Mhz and always reboots with vdd1.
After disabling vdd1 these high-load-tasks work on my device flawlessly, even at 900 Mhz.
So make sure you turn off SmartReflex VDD1 on your N900, although it should be off by default:
sudo "echo 0 > /sys/power/sr_vdd1_autocomp"
udaychaitanya16
2012-02-11, 12:19
First I have turned off smartreflex from /etc/pmconfig,though cpu is at 805.
Then entered
sudo nr_requests_setup
sudo swapoffon
There after copy command.
Though the device is not switched off itself,since 15 minutes x terminal froze .Even the backlight is not turned off.Still waiting to observe what happens.:cool:
magic_doc
2012-02-11, 15:13
Had also problems with reboots while unlzma'ing and found out that on my n900 enabled SmartReflex VDD1 causes this.
Tested it with 900 to 600 Mhz and always reboots with vdd1.
After disabling vdd1 these high-load-tasks work on my device flawlessly, even at 900 Mhz.
eight, what kernel are you using?
The VDD0/1-Settings is to activate SmartReflex, but this runs only reasonable with KernelPower49.
With the one in extras-testing you can safely overclock to 805MHz at the most devices and the battery-savings are awesome.
Only a hint and sorry for the off-toppic. :D
More informations can be found here (http://talk.maemo.org/showthread.php?t=71879).
HTH and HAND
CU Michael
I'm using power49 withd vdd1 and vdd2 enabled at 900 Mhz for some days now.
Works well expect of unlzma'ing big archives. To be more precise, dd .. or tar -cfz (-xfz) of my 3 Gig debian partition works without issues.
Only lzma gives reboots even with nice -n 19.
I have tested it with this archive: http://qole.org/files/maemo-sdk-v1_2.img.ext2.lzma
Switching SR VDD off for lzma tasks like this is quickly done.
And the battery-savings of VDD2 are really good. A big difference, compared to only-vdd1-mode and directly noticeable. :)
Ok, here's how to disconnect the image's /home/user directory from the Maemo version, allowing for per-image user settings.
Create a script in the image's /var/run directory called "onfirstchroot-ext.rc", or add the following lines to the existing script:
umount -f "$CHROOT/home/user/MyDocs"
umount -f "$CHROOT/home"
mount -o bind /home/user/MyDocs "$CHROOT/home/user/MyDocs"
If there are some settings directories you wish to share between the two environments, you can use the syntax of the last line (the "mount -o bind" line) to bind other home/user directories from Maemo to their corresponding directories in Debian.
Now you can copy all of your Debian application settings into the image's /home/user dir, and things will be the same on everybody's N900.
ESTEL: Please either do this with your image and re-upload it, or give me the zip file of your home directory settings files, and I'll do it on my qole.org version.
ivgalvez
2012-02-13, 19:13
I have been testing Estel's image installed on my SD card (ext3, non dedicated partition) with no problems.
Now I've tried to gain some performance by enabling swap on SD:
sudo swapon -p 0 /dev/mmcblk1p3
sudo swapoff /dev/mmcblk0p3
As a result:
~ $ cat /proc/swaps
Filename Type SizeUsed Priority
/dev/mmcblk1p3 partition 795208 64376 0
Everything seemed to work OK. tried Fennec and it was slow as usual. However when I run ED LXDE I receive this error:
Mount failure!
/mnt/data/easydebian/ed-squeeze-final.ext3 failed to mount on loop0
mount: mounting /dev/loop0 on /.debian failed: Invalid argument
Mount failure!
/mnt/data/easydebian/ed-squeeze-final.ext3 failed to mount on loop0
mount: mounting /dev/loop0 on /.debian failed: Invalid argument
When did you download the image? Maybe your problem is related to this (http://talk.maemo.org/showpost.php?p=1157384&postcount=2561) and you try to mount a tar archive? Estel updated the image some days ago, if you downloaded before just untar it before mount.
tar -xfv ed-squeeze-final.ext3
ivgalvez
2012-02-14, 08:01
When did you download the image? Maybe your problem is related to this (http://talk.maemo.org/showpost.php?p=1157384&postcount=2561) and you try to mount a tar archive? Estel updated the image some days ago, if you downloaded before just untar it before mount.
tar -xfv ed-squeeze-final.ext3
Thanks, but that's not the case as the image was working before changing the swap options. Now I have rebooted (swap again in eMMC) and yet the same error. It seems that image is corrupted.
is it possible to extract all the files from the .ext3 image to a partition (without mounting and copying the files from the mounted folder because this will also copy the /home/user and /home/user/mydocs folders)?
i would like to extract it to an ext4 partition i made for it on the sd.
ivgalvez
2012-02-17, 17:23
Estel: Did you upload your configuration files for your image to anywhere else?
Temporal
2012-02-25, 01:20
1. Mount (empty) dedicated partition that You've just created
2. Mount Your ED image (mount -o loop /path/to/ed-squeeze-final.ext3 /path/to/mountpoint)
3. cp -a /path/to/image/file/mountpoint/* /path/to/partition/mountpoint/
4. Edit /home/user/.chroot, to make it point into Your partition, instead of image file (for example, /dev/mmcblk0p4, if Your dedicated ED partition is 4th on eMMC).
Sorry for bothering, but, I'm with a problem:
~$ debian-lxde
Starting Debian echo "chroot is now open!"
Chroot dir specified: /.debian
/dev/mmcblk0p5 specified in ~/.chroot
Mounting...
using device: /dev/mmcblk0p5
Using ext3 file system
mounting device: /dev/mmcblk0p5
.
..
...
....
Updating the /etc/mtab file...
Everything set up, running chroot...
chroot is now open!
Starting Debian /usr/bin/Xephyr :1 -screen 800x480 -br -ac
Chroot dir specified: /.debian
/dev/mmcblk0p5 specified in ~/.chroot
/.debian has a qmount already!
/dev/mmcblk0p5 already mounted on /.debian...
Everything set up, running chroot...
/usr/bin/Xephyr: error while loading shared libraries: libX11.so.6: cannot open shared object file: No such file or directory
What am I doing wrong? The img file works just ok...
Temporal
2012-02-26, 20:36
Ok, I believe I have a bug report:
1)Every time I reboot, the only way to be able to mount by "loop" is first running sudo debian. I don't know why;
2)The only way to run debian-lxde is by first mounting the partition (in my case, mmcblk0p5) to /.debian;
Now I just need to mount using event.d files to be able to run lxde (so I don't need to manually mount every time I run lxde), so, at least for now this works for me.
Sorry for late reply, I've just gone through through 2,5 weeks of posting. I'll filter out and publish /home/user config files ASAP, then, I'll create new image containing required config files and replace old one with it.
/Estel
I just tested a Debian Wheezy chroot in armhf. In principle it works but there are some odd things that I don't understand.
To check whether there is any performance gain I ran nbench. For reference I first did this in in an armel chroot. The result is as follows:\h:\w$ ./nbench
BYTEmark* Native Mode Benchmark ver. 2 (10/95)
Index-split by Andrew D. Balsa (11/97)
Linux/Unix* port by Uwe F. Mayer (12/96,11/97)
TEST : Iterations/sec. : Old Index : New Index
: : Pentium 90* : AMD K6/233*
--------------------:------------------:-------------:------------
NUMERIC SORT : 311.04 : 7.98 : 2.62
STRING SORT : 48.123 : 21.50 : 3.33
BITFIELD : 1.4078e+08 : 24.15 : 5.04
FP EMULATION : 94.282 : 45.24 : 10.44
FOURIER : 355.5 : 0.40 : 0.23
ASSIGNMENT : 5.8009 : 22.07 : 5.73
IDEA : 771.67 : 11.80 : 3.50
HUFFMAN : 517.72 : 14.36 : 4.58
NEURAL NET : 0.52156 : 0.84 : 0.35
LU DECOMPOSITION : 16.355 : 0.85 : 0.61
==========================ORIGINAL BYTEMARK RESULTS==========================
INTEGER INDEX : 18.351
FLOATING-POINT INDEX: 0.660
Baseline (MSDOS*) : Pentium* 90, 256 KB L2-cache, Watcom* compiler 10.0
==============================LINUX DATA BELOW===============================
CPU :
L2 Cache :
OS : Linux 2.6.28.10-power49
C compiler : gcc version 4.4.5 (Debian 4.4.5-8)
libc : libc-2.11.3.so$W5 (Debian 4.4.5-8)
MEMORY INDEX : 4.581
INTEGER INDEX : 4.578
FLOATING-POINT INDEX: 0.366
Baseline (LINUX) : AMD K6/233*, 512 KB L2-cache, gcc 2.7.2.3, libc-5.4.38
* Trademarks are property of their respective holder
Then I did the same in the armhf chroot:\h:\w$ ./nbench
BYTEmark* Native Mode Benchmark ver. 2 (10/95)
Index-split by Andrew D. Balsa (11/97)
Linux/Unix* port by Uwe F. Mayer (12/96,11/97)
TEST : Iterations/sec. : Old Index : New Index
: : Pentium 90* : AMD K6/233*
--------------------:------------------:-------------:------------
NUMERIC SORT : 365.68 : 9.38 : 3.08
STRING SORT : 47.982 : 21.44 : 3.32
BITFIELD : 1.6907e+08 : 29.00 : 6.06
FP EMULATION : 51.779 : 24.85 : 5.73
FOURIER : 1512.4 : 1.72 : 0.97
ASSIGNMENT : 6.0761 : 23.12 : 6.00
Illegal instructionThe first thing you'll notice is the "illegal instruction" error. I repeated the test maybe a dozen times and the error did not always occur at the same point. Mostly it happened after the STRING SORT (never before) or BITFIELD test but there was no pattern that seemed to make sense. The above run that even completed the ASSIGNMENT test was the very first try. Each of the following runs failed one test earlier until the 5th one failed after the STRING SORT test. After that the pattern was completely random. Because there is no pattern I believe the reason is not Maemo or the N900 being armhf incompatible but some strange effect of other Maemo kernel settings or programs similar to the watchdog.
Does this make sense to anybody? I'm running KP49 without DSP profile and without CSSU btw.
The second thing you'll notice is that the result of the armhf FOURIER test is more than 4 times as high as the one of the armel test which in theory should come in handy in multimedia related tasks. On the other hand the FP EMULATION test dropped by 45%. Does anybody have an idea why the FP EMULATION result changed or does someone know the practical implications of that change?
To get more practical results I made a video decoding benchmark with mplayer and youtube's 240p version of the big buck bunny video:
armel:\h:\w$ mplayer -benchmark -nosound -vo null /root/bbb.flv
MPlayer 1.0rc3-4.4.4 (C) 2000-2009 MPlayer Team
mplayer: could not connect to socket
mplayer: No such file or directory
Failed to open LIRC support. You will not be able to use your remote control.
Playing /root/bbb.flv.
libavformat file format detected.
[lavf] Video stream found, -vid 0
[lavf] Audio stream found, -aid 1
VIDEO: [FLV1] 400x226 0bpp 24.000 fps 256.8 kbps (31.3 kbyte/s)
================================================== ========================
Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family
Selected video codec: [ffflv] vfm: ffmpeg (FFmpeg Flash video)
================================================== ========================
Audio: no sound
Starting playback...
VDec: vo config request - 400 x 226 (preferred colorspace: Planar YV12)
VDec: using Planar YV12 as output csp (no 0)
Movie-Aspect is undefined - no prescaling applied.
VO: [null] 400x226 => 400x226 Planar YV12
V: 596.4 0/ 0 5% 0% 0.0% 0 0
BENCHMARKs: VC: 31.237s VO: 0.117s A: 0.000s Sys: 16.393s = 47.748s
BENCHMARK%: VC: 65.4215% VO: 0.2452% A: 0.0000% Sys: 34.3334% = 100.0000%
Exiting... (End of file)
armhf:\h:\w$ mplayer -benchmark -nosound -vo null /root/bbb.flv
MPlayer svn r34540 (Debian), built with gcc-4.6 (C) 2000-2012 MPlayer Team
mplayer: could not connect to socket
mplayer: No such file or directory
Failed to open LIRC support. You will not be able to use your remote control.
Playing /root/bbb.flv.
libavformat version 53.20.0 (external)
Mismatching header version 53.19.0
libavformat file format detected.
[flv @ 0x40a06b10]Estimating duration from bitrate, this may be inaccurate
[lavf] stream 0: video (flv), -vid 0
[lavf] stream 1: audio (mp3), -aid 0
VIDEO: [FLV1] 400x226 0bpp 24.000 fps 256.8 kbps (31.3 kbyte/s)
Clip info:
starttime: 0
totalduration: 596
totaldatarate: 317
bytelength: 23668451
canseekontime: true
sourcedata: BADC20661MH1320360564963154
purl:
pmsg:
Load subtitles in /root/
================================================== ========================
Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family
libavcodec version 53.34.0 (external)
Mismatching header version 53.32.2
Selected video codec: [ffflv] vfm: ffmpeg (FFmpeg Flash video)
================================================== ========================
Audio: no sound
Starting playback...
Movie-Aspect is undefined - no prescaling applied.
VO: [null] 400x226 => 400x226 Planar YV12
V: 596.4 0/ 0 5% 0% 0.0% 0 0
BENCHMARKs: VC: 32.116s VO: 0.116s A: 0.000s Sys: 15.791s = 48.023s
BENCHMARK%: VC: 66.8774% VO: 0.2411% A: 0.0000% Sys: 32.8815% = 100.0000%
Exiting... (End of file)As you can se there is virtually no difference, which I don't understand either. Obviously it works in armhf at all, so the nbench "illegal instruction" error doesn't affect mplayer. On the other hand I'd expect either a noticeable increase due to the better FOURIER test results or a decrease due to the worse FP EMULATION results, depending on how mplayer decodes the video. But the result suggests that there aren't any FP operations involved at all. Is that normal for this codec?
@sulu,
I'm no expert, but the only difference between armhf vs armel (or rather, hardfp vs softfp) is in the use of FP registers as a calling convention (as opposed to using the stack).
Both hardfp and softfp will use the floating point unit in exactly the same way.
So, if most of the work done by mplayer is done within a single function (perhaps with other functions inlined), you should see no difference.
Frankly I'm no expert either.
the only difference between armhf vs armel (or rather, hardfp vs softfp) is in the use of FP registers as a calling convention (as opposed to using the stack).Are you sure about that?
If I read [1] correctly then there are ARM devices out there that have no FPU at all and armel works on them. So this either means that armel doesn't use the FPU at all (and instead emulates FP operations using integer functions) or can detect if there is a FPU and make use of it. As I understand it the latter would require quite some overhead during runtime which makes me believe the former is true.
Under this assumption my understanding was that armhf requires an FPU which will then run FP operations much faster than armel using its integer emulation. This might also explain armhf's poor FP EMULATION test result since it simply wouldn't need to be optimized for that.
Anybody, please correct me if I'm wrong!
[1] http://wiki.debian.org/ArmHardFloatPort#Rationale
btw: As you can see I ran the armel test under Squeeze and the armhf test under Wheezy. I will repeat the armel test under Wheezy to see if the "illegal instruction" error is architecture or release related.
Edit:
I'd welcome any suggestion for a real world FP benchmark! I guess some of the (de)compression algorithms of different archive types make heavy use if it. But I don't have the knowledge which one would be suitable.
I didn't have any need for sound in Easy Debian so far - Maemo satisfies all my needs there - but just out of curiosity: should the Music Player in LXDE with the one title "Lament ..." work out of the box? It doesn't for me. Also youtube videos don't play sound in iceweasel during the short intervals the videos are played at all. I'm just wondering whether I broke something by installing a lot of additional (and actually useful) stuff.
LXDE with the one title "Lament ..."
Does anybody know which that song is ?...
Please Help me...
I really need that Song..:(
I repeated nbench under Wheezy armel which worked fine.
Then I tried it under armhf again but changed the compiler to build nbench. First I tried with gcc-4.4 which resulted in a segmentation fault after the FP EMULATION test instead of gcc-4.6's Illegal instruction.
Then I tried nbench compiled with gcc-4.5 which even completed the IDEA test but then rebooted the phone. So it seems something is terribly unstable here. Either it's the compiler under armhf or some component in the underlying FP hardware/Maemo software.
Can somebody who has any other armhf-compatible device please try if nbench completes reliably if compiled with any of Debian Wheezy's armhf gccs?
Just for completeness, this is the gcc-4.5 result as far as it worked:\h:\w$ ./nbench
BYTEmark* Native Mode Benchmark ver. 2 (10/95)
Index-split by Andrew D. Balsa (11/97)
Linux/Unix* port by Uwe F. Mayer (12/96,11/97)
TEST : Iterations/sec. : Old Index : New Index
: : Pentium 90* : AMD K6/233*
--------------------:------------------:-------------:------------
NUMERIC SORT : 415.04 : 10.64 : 3.50
STRING SORT : 45.57 : 20.36 : 3.15
BITFIELD : 1.1854e+08 : 20.33 : 4.25
FP EMULATION : 51.798 : 24.86 : 5.74
FOURIER : 1461.6 : 1.66 : 0.93
ASSIGNMENT : 5.8777 : 22.37 : 5.80
IDEA : 1206.9 : 18.46 : 5.48
mr_pingu
2012-03-06, 15:52
Estel, did you already upload your new image with the files from /home/user/ ?
My ED image got corrupted so I am forced to download a new image ;) Then I started wondering if you had fixed your image
Unfortunately, not yet. Sorry, have been quite occupied lately. BTW, after doing so, I'll "announce" it here for sure.
/Estel
rhadesan
2012-03-19, 08:18
Hello!
I am trying Easy Debian right now, and I am not able to perform apt-get update. It look good at the begining, but then:
Err htp://packages.tspre.org lenny/main Packages
504 Gateway Timeout
Err htp://packages.tspre.org lenny/non-free Packages
504 Gateway Timeout
Err htp://packages.tspre.org lenny/contrib Packages
504 Gateway Timeout
Fetched 808kB in 1min 10s (11.5kB/s)
W: Failed to fetch htp://packages.tspre.org/dists/lenny/main/binary-armel/Packages.gz 504 Gateway Timeout
W: Failed to fetch htp://packages.tspre.org/dists/lenny/non-free/binary-armel/Packages.gz 504 Gateway Timeout
W: Failed to fetch htp://packages.tspre.org/dists/lenny/contrib/binary-armel/Packages.gz 504 Gateway Timeout
E: Some index files failed to download, they have been ignored, or old ones used instead.
[root@m5v3e: /]
Is there something wrong?
PS: I have to do s/http/htp/ because of rules of this forum.
@rhadesan:
You're right, the server is down.
Just deactivate that entry in your sources.list by prepending: #
It's quite a long time since I checked but afair there are no critical packages coming from this repository. It's safe not to use it.
Estel and sulu:
Could either of you please send me your LXDE config files from your home directory?
I really don't think I should leave your images as part of my installer with the current config problems.
mr_pingu
2012-04-05, 21:31
I just started to make my own image based on estels but is there really something wrong? I mean, what's missing besides LXDE config? I think it's quite stable and not much to edit?
I just wanted these files for completeness sake, without knowing what exactly was missing...
Estel and sulu:
Could either of you please send me your LXDE config files from your home directory?
I really don't think I should leave your images as part of my installer with the current config problems.I'm not quite sure which of the files in my .config directory are from Easy Debian and which are from Maemo applications. Therefore I included everything that I think might be from ED. If something is missing nevertheless please tell me!
Thanks sulu, it's shame on me, but accidentally, I've unsubscripted from this thread :o Thus, totally forget about leaving image incomplete, without config files.
/Estel
Hi, is there any way to install ff11 (iceweasel 11) on ED? I tried adding experimental repos but it gave me broken dependencies, something about gtk. I also tried to extract the deb to a folder and run the binary but that didnt work either
1. You can try to download missing dependencies one by one, and see, if - at the end - You get something usable, thus, creating another releases mix, that we've just avoided ;)
2. You can wait for new stable release, and try to dist-upgrade it, then - after that - fix all conflicts, just like Sulu di. (and, in very small addition, I).
Is Iceweasel 11 with it's great slowness worth that? Answer Yourself ;)
/Estel
I'ts just that I am a web developer and I want to be able to show off my work on my phone, and as firefox is my dev browser, most possibly I can show off work in progress. 3.5 is slowly becoming tooo old. On the other hand I think that after 3.6 they've improved a bit in terms of speed.
I found that I had also somehow unsubscribed from this thread, which strikes me as VERY odd.
Anyway, hopefully I will have time this long weekend to rebuild sulu's image with the extra files and re-publish on qole.org.
I found that I had also somehow unsubscribed from this thread, which strikes me as VERY odd.
So, there are already two of us unsubscripted for reasons unknown. Bug?
Anyway, I tried to create archive with my config files related to ED, but it resulted in blazing ~16MB, mainly due to chromium temporary files, databases erc., also located in /home/user/.config/chromium. Same apply for LibreOffice (or OpenOffice) files, including sketches of my personal documents, etc.
Does anyone know an easy way to post it, without - at the same time - providing users unnecessary cache of web sites visited by me, my cookies, encrypted password cache (sic), or other things? Of course, i mean other way than manually examining every file?...
BTW, wouldn't chromium, LibreOffice etc. rebuild it's config files to default, in case of lacking them? Maybe it's important to only post lxde, pcmanfm and things like that?
I would really like to allow people hassle-free usage of my image (which may be felt - by many users - as preferred, due to already included LibreOffice instead of old OpenOffice, and fast Chromium instead of slow Iceweasel, + LXDE panels tweaked to fit N900 as much effectively as I was able to), yet, this whole config things start to become more complicated than work on image itself.
Of course, it's probably just me and my lack of knowledge, so any hints really appreciated.
/Estel
maartenmk
2012-04-07, 11:26
So, there are already two of us unsubscripted for reasons unknown. Bug?
My subscription was still ok.
[....]
BTW, wouldn't chromium, LibreOffice etc. rebuild it's config files to default, in case of lacking them? Maybe it's important to only post lxde, pcmanfm and things like that?
LibreOffice works fine from your image.
Chromium also works, but it didn't have your customisations, and since these improve usability a lot, I would appreciate them being included. (I did try to recreate them, but I'm not sure I got it right, it's kinda slow at times)
I would really like to allow people hassle-free usage of my image (which may be felt - by many users - as preferred, due to already included LibreOffice instead of old OpenOffice, and fast Chromium instead of slow Iceweasel, + LXDE panels tweaked to fit N900 as much effectively as I was able to),
[...]
/Estel
Agreed. While I was amazed how well Easy Debian worked on the N900, it still is a matter of compromise, so any tweaking helps.
My subscription was still ok.
As is mine. It seems just the more important people were unwillingly unsubscribed. Sabotage, I say. :D
Ok, I have uploaded the new sulu image in place of the old one. The panels now work correctly, including the very cool orientation turn icon. But the portrait mode is 'upside down' for me. No big deal, though.
I'm proud to say I did everything on-device, including lzma and upload to qole.org.
Chromium also works, but it didn't have your customisations, and since these improve usability a lot, I would appreciate them being included. (I did try to recreate them, but I'm not sure I got it right, it's kinda slow at times)
Things for chromium were posted here:
http://talk.maemo.org/showpost.php?p=1158202&postcount=2568
...And, just for convenience, .desktop files for Maemo:
http://talk.maemo.org/showpost.php?p=1158803&postcount=2573
...allowing hassle-free launching of Chromium and LibreOffice via regular shortcuts.
As for the latter, it can't be included in image (except for some post-first-chroot magic?), qole would need to add it to chroot package, like he did with OpenOffice and Iceweasel in old versions of ED.
For Chromium extensions, I don't have idea for good way of including them, without making it mandatory - IMO they're a must-have due to drag'n'drop scrolling, but I imagine that few people may don't like it.
/Estel
// Edit
I'm currently packing config files without ones from LibreOffice and Chromium, will post it here in minutes.
// Edit 2
Here it is (see attachment).
qole, Your version of sulu's image also contain /home/user totally separated from Maemo one, as You suggested (via script) some posts ago? Do You feel, that You could also re-build and "approve" my image the same way?
By the way - it is perfectly safe now, to perform apt-get upgrade (and apt-get update, obviously) - repositories are set - by default - to contain Squeeze and squeeze-backports. It's a month already from latest image, so You may find some updates available, and it's perfectly safe to perform.
mr_pingu
2012-04-07, 23:12
Thank you for LXDE config,10 times more convenient than my fast setup. Already had chromium running with scrollbars anywhere, but scrolling is sometimes slow and laggy... Estel, hows the performance on your device?
It's just not how I expected the performance to be. I expected a little bit more "smooth" scrolling. Would using a partition, increase the performance much?
In my case scrolling is generally smooth as silk - and for sure, isn't worst after installing (and configuring) scrollbars anywhere. BTW, it's also worth to install smallest possible modification, that disables scrollbars - we aren't using then anymore, so it's a gain of few additional pixels for web content (just like in microB - unfortunately, I wasn't able to find extension, that hides scrollbars but show them during scrolling only...)
I'm also sure, that using dedicated partition helps performance *a much* - and if I remember correctly, someone in this thread already proved this by measurements.
---
As it was some time since last discussion, for people that may be confused a little - best way to open hidden bars in my LXDE config, is to tap just a little *outside* screen. Due to our resistive screen's characteristic, for LXDE it perfectly simulates a motion to the edge of screen. This way, we can use such big, finger/stylus friendly icons (or even bigger, if someone feel it fancy).
/Estel
maartenmk
2012-04-08, 16:50
Looks like I have about the same performance as mr_pingu with Chromium, scrolling is sometimes smooth, but often laggy, I certainly do not think it is a full replacement for Opera or MicroB. And I do use a partition.
I also had higher expectations after reading Estel's experiences with it.
Also, consider using swap *only* on microSD, and using *internal* eMMC partition for ED (eliminates plentora of I/O conflicts on eMMC). If all of those is valid, maybe You have some memory-hungry widgets or whatsnot? I'm quite minimalistic in that matter, so my "standby" memory footprint is relatively small. On the other hand, full-fledged browsers like Chromium are memory hogs, so every megabyte saving You from using swaps count here.
BTW, of course, preferences wary - for me, opera isn't any alternative to microB or Chromium via ED (any single of them).
/Estel
I will get Estel's config files put into the image and re-uploaded.
Estel: what is the mysterious broken icon on the top panel?
http://qole.org/images/Estel-LXDE.png
it is supposed to be (upgraded) orientation-switch. It *should* have proper icon, and result as switch between landscape, portrait, and portrait with virtual keyboard on-screen.
I suppose, that it's non-functioning at all, eh? It must've been something that also diden't live through image transfer. I'll try to search for script/icon and post it here in a moment.
Thanks for putting work into it and spotting errors!
/Estel
//Edit
Here it is (see attach.)! Sorry for troubles.
Hi, is there any way to install ff11 (iceweasel 11) on ED? I tried adding experimental repos but it gave me broken dependencies, something about gtk. I also tried to extract the deb to a folder and run the binary but that didnt work eitherI just started trying to dist-upgrade ED to Wheezy. I did not plan to release it before Wheezy is frozen but you have a point here that might make me reconsider that.
My intention was to create a very minimal image, mainly only consisting of LXDE and audio.
The idea was that somebody else with basic knowledge in Debian administration could easily remaster that image for end-users without having to throw out too much "sulu-specific" bloat.
Give me another week or two and I may be able to come up with an image where you can install a recent version of Iceweasel, Chromium or whatever! But since it will be a Testing image you'd have to use it with caution - usually the same caution that always comes with running Testing.
Does anyone know an easy way to post it, without - at the same time - providing users unnecessary cache of web sites visited by me, my cookies, encrypted password cache (sic), or other things? Of course, i mean other way than manually examining every file?...I know no easy way to do that. But I always thought that it might be prudent to separate ED config files from Maemo config files. This could be reached by not bind-mounting Maemo's .config directory into ED. This on the other hand might bring other side effects that I haven't considered yet. Any thoughts on that?
BTW, wouldn't chromium, LibreOffice etc. rebuild it's config files to default, in case of lacking them? Maybe it's important to only post lxde, pcmanfm and things like that?I think so. At least for chromium I know that's the case. Essentially my config archive only contains LXDE-related stuff.
I would really like to allow people hassle-free usage of my image (which may be felt - by many users - as preferred, due to already included LibreOffice instead of old OpenOffice, and fast Chromium instead of slow Iceweasel, + LXDE panels tweaked to fit N900 as much effectively as I was able to), yet, this whole config things start to become more complicated than work on image itself.I totally agree. Do you remember which "bugs" your image had (and probably originated in mine) that should be addressed in a new release? I remember the missing /etc/groups.
But the portrait mode is 'upside down' for me. No big deal, though.I know. I did that intentionally because I always felt Maemo's portrait mode to be upside down. ;)
But for the sake of consistency it's most likely better to have ED portrait mode in the same direction. I will keep that in mind for my next image.
If you want to change that on your own have a look at the orientation script (/usr/bin/orientation afair)! There you'll find xrandr commands where you'll just have to replace "left" with "right" (or vice versa, I'm not sure anymore).
As for the cool icon, could somebody please post the icon alone including it's full path? Personally I don't care for the icon but it would make sense to have it in my next image if it's really used the way I have in mind.
qole, Your version of sulu's image also contain /home/user totally separated from Maemo one, as You suggested (via script) some posts ago?If that means what I think it means I might have missed an important development. Can I have a link to that discussion please?
zlatokosi
2012-04-10, 12:06
debbie vs. chroot
A quick question:
When I start Chromium with debbie (or from within LXDE), it never remembers my preferences. But if I start it in chroot, everything works fine.
Any ideas what is going on?
Here is the output when I start chroot.
/sbin/debian: /home/user/.chroot: line 1: : not found
Chroot dir specified: /.debian
/media/mmc1/ed-squeeze-final.ext3 specified in ~/.chroot
/.debian has a qmount already!
/media/mmc1/ed-squeeze-final.ext3 already mounted on /.debian...
Everything set up, running chroot...
Thanks in advance.
If that means what I think it means I might have missed an important development. Can I have a link to that discussion please?
sulu, if you are building a new image, please start with the one at qole.org. I have added the script that unmounts the Maemo /home/user directory (but keeps the MyDocs directory mounted), so that it uses the configuration files from within the image instead of the Maemo ones. This means that all of your settings will be maintained on all installations.
This is the method I used:
Ok, here's how to disconnect the image's /home/user directory from the Maemo version, allowing for per-image user settings.
Create a script in the image's /var/run directory called "onfirstchroot-ext.rc", or add the following lines to the existing script:
umount -f "$CHROOT/home/user/MyDocs"
umount -f "$CHROOT/home"
mount -o bind /home/user/MyDocs "$CHROOT/home/user/MyDocs"
If there are some settings directories you wish to share between the two environments, you can use the syntax of the last line (the "mount -o bind" line) to bind other home/user directories from Maemo to their corresponding directories in Debian.
Now you can copy all of your Debian application settings into the image's /home/user dir, and things will be the same on everybody's N900.
As for the cool icon, could somebody please post the icon alone including it's full path? Personally I don't care for the icon but it would make sense to have it in my next image if it's really used the way I have in mind.
Don't get it wrong way, but it is attached to the post before Your post: ;)
http://talk.maemo.org/showpost.php?p=1190038&postcount=2653
Good luck with Wheezy, it would be awesome to have such up-to-date version fully working, as with Squeeze!
debbie vs. chroot
A quick question:
When I start Chromium with debbie (or from within LXDE), it never remembers my preferences.
I'm quite convinced, that You've started Chromium as root for the first time, so it's created it's config files with write permissions for root only. You need to manually delete chromium's config files and start it again as user ('debbie' method) - it will re-create config files - or, if You're sure that You know which file require what permissions exactly, You can manually chmod them (not recommended).
---
qole, dumb question mode again - will using You script (to separate Maemo's and ED's config files) on already populated image, cause need to move every ED's config file from /home/user/.config to ED equivalent, manually?
BTW, do You plans to do same thing to sulu's and mine image, before posting fixed versions of both?
Thanks for Your expertize and work on fixing Squeeze images, again.
/Estel
Ok, here's how to disconnect the image's /home/user directory from the Maemo version, allowing for per-image user settings.
Create a script in the image's /var/run directory called "onfirstchroot-ext.rc", or add the following lines to the existing script:
umount -f "$CHROOT/home/user/MyDocs"
umount -f "$CHROOT/home"
mount -o bind /home/user/MyDocs "$CHROOT/home/user/MyDocs"
If there are some settings directories you wish to share between the two environments, you can use the syntax of the last line (the "mount -o bind" line) to bind other home/user directories from Maemo to their corresponding directories in Debian.
Now you can copy all of your Debian application settings into the image's /home/user dir, and things will be the same on everybody's N900.That seems strange to me! First you mount Maemo's $HOME from the Easy Debian scripts in Maemo and then you unmount it again from within Easy Debian.
I see the advantage of not having to update the Easy Debian package, but it seems like a bad idea as a permanent solution to me.
I see three possible permanent solutions:
1. We stick with the old mixed $HOME (i.e. Maemo's $HOME mounted).
2. We go for two separated $HOME's by removing the mount command from the package's script.
3. We make this mount line optional, either via a comand line switch, a config file or some other method.
I'll vote for 3. but since I can't fully foresee the consequences (different settings leading to different side effects making support harder) I'd prefer 2. over 1. as a fallback.
mr_pingu
2012-04-12, 16:01
I prefer option 2, while configs remain separated the mass storage remains accesible in both.
Everytime I need a file from maemo in easy debian, I always stored it in MyDocs, besides I think it's cleaner to have them separated.
Imagine: What if you have installed program x in maemo and in ED, do they share the same config then?
Estel, I couldn't get your orientation switch working. chmod'ed +rwx
the binary but LXDE wouldn't want to update it's icon nor change orientation when clicked on...
Some offtopic: I have another maemo chroot, I always use qchroot and quserchroot to chroot into it, but I would like to have acces to the real MyDocs. What do I need to do to make that happen, looks like I need a small script for that?
Damn, by placing proper .desktop icon in correct place inside ED image, You - basically - create program shortcut (available from LXDE panel). .desktop file is pointing to correct file in /usr/bin (orientation script), and to correct icon. You should be even able to remove this icon from panel, then, add it again manually from scratch, as it's listed in Ed programs shortcuts.
So, what could go wrong in Your case? Are You sure, that location of all files are correct, i.e. no 'duplicated' paths, etc?
Anyone else having this problem?
/Estel
// Edit
I also think, that sulu's proposed option 2 is best - if someone want shared $HOME, she/he can always set it in script executed on every chroot. Default settings should be different home's.
I prefer option 2, while configs remain separated the mass storage remains accesible in both.For most users that's true. But some people (including me) have switched the home and MyDocs partitions. Except for the very few cases of taking photos with my N900 I practically don't use MyDocs.
But these are special cases and Maemo's home can still be mounted elsewhere.
Imagine: What if you have installed program x in maemo and in ED, do they share the same config then?Theoretically yes. For example I have Chromium installed in ED and Maemo and since the ED version is newer the Maemo version always complains about some incompatibilities (but otherwise works fine).
Estel, I couldn't get your orientation switch working. chmod'ed +rwx
the binary but LXDE wouldn't want to update it's icon nor change orientation when clicked on...What happens if you call /usr/bin/orientation manually from within roxterm?
Some offtopic: I have another maemo chroot, I always use qchroot and quserchroot to chroot into it, but I would like to have acces to the real MyDocs. What do I need to do to make that happen, looks like I need a small script for that?Basically all you have to do is mount it somewhere just like it is done in ED. Suppose you want to mount it to /mnt in your chroot and your chroot directory is /home/user/chroot just run this in Maemo:mount -o bind /home/user/MyDocs /home/user/chroot
With the older images I was able to use a modified keboard which works under maemo and have óćą etc. in ED. Now with Estels image the keys don't work. I did the setxkbmap pl and did the locale but still there are no Polish keys. How can I fix it?
Edit:
The solution is there in this topis :)
I even asked for the same thing
sudo synchroot is the trick :)
Hi all, I'll have Estel's image up soon. I ran out of time on the long weekend and life overtook me.
I like the method I've posted for separating the home directories, because it is very image-specific. You can set it up per-image. The default script mounts the home directory, but this behavior can be modified, depending on which image is mounted. It is also 100% backward compatible with all previous images, which I like, because I don't want to upload a new version of the package.
Estel's orientation icon in the .config files that he posted is broken, it points to some strange help .desktop file in .config. I have fixed it to point to the proper .desktop file in my version.
Ok, new Estel image on qole.org. Please download and troubleshoot!
I have had a few problems with the new Estel image:
I first reconfigured the locales to pl (maybe this was my mistake)
Then I've tried to put @setxkbmap pl but the file didn't want to save. (I've opned it from within ED)
Also when I was trying to write @ it gave me the TAB
So I opened the file to put the setxkbmap pl line from the chroot and then it worked. After that I did sudo synchroot from maemo xterm and now it seems that it works.
Is it normal that I can't see the panels inside ED? I have to touc the edge o the scren to see them. Is it a space savieur?
Also there is no image on the upper panel like here http://talk.maemo.org/showpost.php?p=1189927&postcount=2652
I have had a few problems with the new Estel image:
I first reconfigured the locales to pl (maybe this was my mistake)
Then I've tried to put @setxkbmap pl but the file didn't want to save. (I've opned it from within ED)
Also when I was trying to write @ it gave me the TAB
To have Your Maemo's keybinding inside ED, You need to copy /usr/share/X11/xkb/symbols/nokia_vndr/rx-51 to same place inside Your image - most likely (when You've image mounted, i.e. ED started) /.debian/usr/share/X11/xkb/symbols/nokia_vndr/rx-51
Before doing so, You may consider examining rx-51m already bundled inside image - it contains 99,99% of keys from big computer keyboard, including F1-F12 mapped to upper row. As for @, it is blue arrow + shift + space, due to bug not allowing mapping TAB for above combination.
I decided to bundle sophisticated remap inside image, to give people full-button acces out of the box - for ones that wan't to use own keymap, copying rx-51 is only method anyway, so it doesn't hurt anyone.
Is it normal that I can't see the panels inside ED? I have to touc the edge o the scren to see them. Is it a space savieur?
Yep. This way, You have full screen (well, without 2 pixels) available for programs, at the same time, having big, stylus/finger friendly icons on panels. Of course, You may disable hiding them, and/or make icons smaller/bigger - everything is customizable. As it was mentioned few pages ago, best way to bring panels is to tap just *outside* screen, near the edge, where panel is to appear - it's like having 2 new touchpad buttons :)
Also there is no image on the upper panel like here http://talk.maemo.org/showpost.php?p=1189927&postcount=2652
No idea, maybe, while fixing it, qole forget to include icon (provided by me with "broken" .desktop file and script itself) inside image?
/Estel
anthonie
2012-04-16, 14:31
For those trying to install E17 (Enlightenment) as per instructions from here (http://talk.maemo.org/showpost.php?p=879368&postcount=22), the sources.list is a bit outdated and needs some fresh juice.
Package repo's
--------------
Get some key, add the corresponding repo to your list, save it
wget http://packages.enlightenment.org/repo.key.old
apt-key add repo.key.oldwget http://packages.linuxniche.net/deblet/repo.key
apt-key add repo.keyNow add the repo's to your sources.list
deb http://packages.enlightenment.org/debian sid main extrasas well as
deb http://packages.linuxniche.net/deblet deblet mainand
deb ftp://de.debian.org/debian sid mainSave your sources.list file, update and follow the instructions as posted by maddog.
EDIT: As sulu mentions, don't forget to turn them off again once you're done installing enlightenment. They will break things at some point if you forget.
mr_pingu
2012-04-16, 14:47
What's that "rotation switch" supposed to do? On the new uploaded image I still got that red cross and no visible function...
Yeah what's with the repo's? I used to find aircrack-ng there(old images), now it seems gone
deb ftp://de.debian.org/debian sid mainDon't forget to disable that entry again after installing e17!
Otherwise things may break. Audio and Keyboard layout in Easy Debian WILL break if the corresponding packages (pulseaudio & xkb-data) are upgraded for some reason.
What's that "rotation switch" supposed to do? On the new uploaded image I still got that red cross and no visible function...Does calling /usr/bin/orientation from roxterm show any result?
That's what the icon is supposed to do.
mr_pingu
2012-04-16, 15:02
Does calling /usr/bin/orientation from roxterm show any result?
That's what the icon is supposed to do.
Yes that works, it rotates the screen portrait ;)
Yes that works, it rotates the screen portrait ;)Ok, then it's not my fault and you can blame Estel and/or qole. :D
Seriously, please have a look at the file: /usr/share/applications/orientation.desktop
There you'll find this line: Exec=/usr/bin/orientation
If that's true the panel starter should work, even if the icon graphic is missing. If it doesn't add a new starter pointing to: /usr/share/applications/orientation.desktop
Edit:
Do you have an "Orientation" starter in LXDE's menu? If so, does it work?
T
I decided to bundle sophisticated remap inside image, to give people full-button acces out of the box - for ones that wan't to use own keymap, copying rx-51 is only method anyway, so it doesn't hurt anyone.
/Estel
So how to have Polish letters on hw keyboard with your image and without tinkering anything with the rx-51 file?
You can probably blame me, I must have uploaded the wrong image. Arggh! Sorry, will fix tonight.
Ok, sorry for the false alarm!
The correct Estel image is now posted on qole.org... It includes the orientation icon and the chromium grab+drag plugin that Estel suggested. I also compacted it as much as possible, to improve the download time.
Here it is (http://qole.org/files/debian-m5-estel.img.ext3.lzma).
http://talk.maemo.org/showpost.php?p=1193416&postcount=598
This seems to be case for thing that was bugging me for ages - MyDocs not mounted automaticaly, after using mass-storage mode. It's worth to add, that after such event, next try of mass-storage doesn't work at all.
I have Easy Debian mounted all the time (automaticaly during boot), as I don't want to have longer starting time for ED programs, and I use them very often. So, I can 100% confirm that ED is the case here.
Any ideas why it's like that, and how to fix it? AFAIK, it's happening no matter of image used.
So how to have Polish letters on hw keyboard with your image and without tinkering anything with the rx-51 file?
IIRC, just use shift+blue arrow + a for dead_ogonek (so, You have to press letter You want to merge with ogonek *after* that, or press space, in case You want to have ogonek itself. It works the same way for dead_accute (shift + blue arrow + c)and dead_abovedot (blue arrow + z). Lstroke is invoked directly by shift + blue arrow + l, as there is no working dead_stroke :( (unfortunately).
Yet, I strongly encourage everyone to use own, preffered keyboard mapping, by copying RX-51.
/Estel
It is quite important that you issue a "sudo closechroot" (or use the icon) before connecting the USB cable for "Mass Storage Mode".
The Easy Debian mount definitely interferes with the USB mode.
ArchiMark
2012-04-19, 15:12
Been trying to download 'Estel' image file since yesterday...usually what happens is that file downloads partially and then process stops....best I've done is last night where it downloaded almost all the file except the last 1 MB.....
Just tried again and download is stopped at about 158MBs......
Is there
A. a recommended way to do the install using the installer
or
B. is there an alternate method to do the install by downloading file onto my computer (assuming that I have better luck that way...) and then copying the file onto my microSD card and putting card into N900 and then completing install from there?
Thanks!
B. is there an alternate method to do the install by downloading file onto my computer (assuming that I have better luck that way...) and then copying the file onto my microSD card and putting card into N900 and then completing install from there?The fastest option should be to download the file to your computer, extract it there, copy it to your N900 (either via USB or network) and in Maemo change the path to your image file in /home/user/.chroot
It is quite important that you issue a "sudo closechroot" (or use the icon) before connecting the USB cable for "Mass Storage Mode".
The Easy Debian mount definitely interferes with the USB mode.
Acknowledged, but why it's that, technically? BTW, Pali updated ke-recv recently, so maybe a workaround for that could be implemented too? Could You elaborate more?
/Estel
i'm using Estel's image (thanks man !!! :D) but i couldn't map the "ç". Tried key <BKSP> { type[Group1] = "PC_FN_LEVEL2", symbols[Group1] = [ BackSpace,ccedilla ] }; but no sucess... Does anybody knows how to do it ?
Thanks.
The problem is that when you plug in the USB cable, the N900 tries to dismount and make all of its SD partitions and the MyDocs partition available to the PC as external drives. If an image file on one of those partitions is mounted as the chroot on the N900, it can't dismount the partitions.
@qole:
Negative, I don't use image file for ED at all - I have dedicated, independent partition for ED (on eMMC). Yet, as You can see from OP and my reports, MyDocs and SD card is unmounted *properly*, yet, MyDocs fail to mount again. It happens only, when ED chroot is active (i.e. mounted, not necessary in direct use).
i'm using Estel's image (thanks man !!! :D) but i couldn't map the "ç". Tried key <BKSP> { type[Group1] = "PC_FN_LEVEL2", symbols[Group1] = [ BackSpace,ccedilla ] }; but no sucess... Does anybody knows how to do it ?
Thanks.
Does it works for You on Maemo, via hardware keyboard remapping? If yes, just copy rx-51 file (as per my instructions from few posts back) into Easy Debian, and, from within it, execute setxkbmap <Your language code>.
/Estel
@Estel
When you do NOT have your chroot partition mounted on the N900, does your PC mount that partition when you connect the phone?
=Does it works for You on Maemo, via hardware keyboard remapping?/Estel
Nope. But on Maemo i use the virtual kb. Right now im using on ED "dead_cedilla + c" but since i saw "ccedilla" on rx-51 im trying set it up. Never mind !!! Is just one more tap. Thanks for replay. Your .img rocks !!!
@Estel
When you do NOT have your chroot partition mounted on the N900, does your PC mount that partition when you connect the phone?
Nope, it only mount MyDocs and MMC as usual. With chroot mounted PC also mounts usual pair - the difference is that after closing mass storage mode, MyDocs isn't mounted again by N900, if chroot mount is present. MMC is mounted fine, ho ever.
Another weird thing is that, with chroot mounted, MMC in mass storage mode is accessible from both N900 and PC - that is, it doesn't get unmounted from N900, during mass storage mode enabling (no error about it being in use appears, and it's accessible for read&write for Maemo and PC simultaneously - with possible risk of filesystem corruption, if both would try to write same file at same time.
/Estel
As for chromium extensions - I've already incorporated "Scrollbars anywhere", which is (AFAIK) most customizable "touch scroll" thing for Chromium. Using it, I was able to enable "drop and drag" (also on text/images), while still retaining capability to select text/drop page elements/whatever. I'm pretty sure everyone will find fancy customization mix for own liking. I also checked it for backdoors, at least as much as my limited skills permit ;)
/Estel
How does this work? Is it really included in your latest image? If so, how do I active it?
Thanks for your image!
Edit:
Found this on the command line...
CJ PLUGIN: thread 0x2b81bc28: NP_GetValue return
[12594:12594:21132130036:ERROR:chrome/browser/extensions/extension_error_reporter.cc(53)] Extension error: Could not load extension from '/home/user/Downloads/scrollbars-anywhere'. Manifest file is missing or unreadable.
IIRC, qole said that he fixed it in latest version of image (included already). Do You use version from His server?
In case You don't want to replace image just for that, You can (re)install extension manually, as per my instruction (bundled in last 10-15 pages of this thread, include "cheating" google store to actually download extension).
/Estel
IIRC, qole said that he fixed it in latest version of image (included already). Do You use version from His server?
In case You don't want to replace image just for that, You can (re)install extension manually, as per my instruction (bundled in last 10-15 pages of this thread, include "cheating" google store to actually download extension).
/Estel
Image downloaded today from his server..
I will look into that and see if I make it.
//Wille
Hello again!
I have fixed the "Scrollbars anywhere" extension but now I can't get java applets to work.
GCJ PLUGIN: thread 0x2b821c68: NP_Initialize
GCJ PLUGIN: thread 0x2b821c68: plugin_test_appletviewer
GCJ PLUGIN: thread 0x2b821c68: plugin_test_appletviewer return
GCJ PLUGIN: thread 0x2b821c68: NP_Initialize: using /usr/lib/jvm/java-6-openjdk/jre/lib/../../bin/pluginappletviewer
GCJ PLUGIN: thread 0x2b821c68: NP_Initialize return
GCJ PLUGIN: thread 0x2b821c68: GCJ_New
GCJ PLUGIN: thread 0x2b821c68: plugin_data_new
GCJ PLUGIN: thread 0x2b821c68: plugin_data_new return
GCJ PLUGIN: thread 0x2b821c68: plugin_get_documentbase
Invalid pipe names provided. Refusing to proceed.
Any ideas? Does it work for you?
I have been fighting with the keyboard shortcuts and found out that the .xbindkeyrc file dosen't exist in the chroot /home/user and because of that xbindkeys fails. Is this an error or is it changed from how the wiki says:
Note that the relevant file /home/user/.xbindkeysrc does not reside in the Debian image but in the home directory.
Should I move it in to the chroot or is there something missing in Estels image?
EDIT:
I have tried to move it in to the chroot and now the keybindings works! :) I'm quite sure that this worked in the old Estel image though, without copying the file..
No idea, could You put a link to example site causing this problem?
No idea, could You put a link to example site causing this problem?
Well, for me it never works but you could test on this site:
http://profs.etsmtl.ca/mmcguffin/learn/java/01-drawingLines/
It should be a java applet there with black background and green lines.
It seems to me that the md5sum in imagelist-fremantle-01 are not up to date.
I got for debian-m5-sulu.img.ext2.lzma
21c11b164acd0a730f740ceecd612da4
and for debian-m5-estel.img.ext3.lzma
cea873aa149dba449d3c54e7152fe12c
wheras in imagelist-fremantle-01
3,Estel,http://qole.org/files/,debian-m5-estel.img.ext3,lzma,Large Debian 6 with LibreOffice,75d4d0a24d1b62424919accd0a1be9af
4,sulu,http://qole.org/files/,debian-m5-sulu.img.ext2,lzma,Debian 6 image by sulu (DE),90dea3dad97b336fa82a14be50e8fce4
Estel - just started using your new image - works great - i like this chromium theme :)
This is how i use it:
Created an ext4 partition on SD
Copied the ~3GB file to it
Mount it to /.debian
Is there a benefit to copying the image's files to the partition and use it directly instead of using the mount?
Thanks.
Thanks for Your appreciation :). As for running native partition vs loopback image - there is *much* benefit of doing so. Currently, You're not gaining (almost) anything over keeping image file in vfat. If You actually have files on partition, natively, You're benefiting running whole (Easy) Debian natively (chroot = native). Getting files out/to loopback image is second biggest bootleneck, next only to low amount of RAM.
---
In my personal opinion, "default" loopback-image configuration of Easy Debian is derivative from times, when repartitioning was "black magic", and qole wanted to save end-users need of doing it. In addition to that, even now, end user can try Easy Debian without changing anything in their filesystem, which is quite big benefit.
Honestly, though, repartitioning through GUI solutions is now trivial for everyone, both ones that have Linux background or not. I would not use image on loopback for any "real" work on Debian programs (especially, resource hungry, like browsers)
[unrelated rant]
Have anyone else noticed, that lately, we're placing web browsers amongst most resource-hungry programs? Hell, even LibreOffice runs better than lightest, feature-complete browsers, like Chromium. Iceweasel/Firefox is such Behemot'ish cow now, that it's even not worth mentioning.
At the same time, ancient microB works blazing fast, and do render most pages well (would render everything well, if we would be able to update underlying Gecko to more recent version - without loosing any speed...). Honestly, it would be 100% full-fledged browser, if we would be able to add few small features to closed UI (+ upgrading Gecko behind it would be almost trivial).
We already know who we need to stab endlessly, for closed microB UI? Suckers, they own this code fully, haven't and won't do *anything* with it ever, yet, they can't release it due to "business reasons".[/unrelated rant]
/Estel
Estel - how do you extract the image to the partition (and how to repack it back from a partition to image)?
mount (empty) partition, mount image file as loopback (mount -o loop <imagefile>, if I recall correctly), then cp -a content of the latter to the former.
then, You need to modify .chroot file, pointing it to partition, instead of image file (refer to Easy Debian wiki for instructions about that).
/Estel
// edit
As for backing it up to image file, You just revert whole process - mount Your file (clear files there, if it's not empty), cp -a everything from partition into file.
freemangordon
2012-05-01, 19:45
@estel - I didn't try the latest version, but in the one i have here, there is a bug in qchroot, causing the infinite (almost) mounting when easy debian is on external sd card (as a directory) and the partition in question is mounted in /media .
here is my fix, in /sbin/qchroot, hope it will be helpful
#Any external devices
echo ... >/dev/stderr
if [ `echo $CHROOT | grep '/media/mmc'` = $CHROOT ]; then
# We are using directory on eMMC or on SD card so skip our partition from mounting
MNTPART=`echo $CHROOT | cut -f 3 -d /`
MNTD=`cat /proc/mounts | grep -v /media/$MNTPART | grep ' /media/' | awk '{print $2}'`
else
MNTD=`cat /proc/mounts | grep ' /media/'|awk '{print $2}'`
fi
for MDRV in $MNTD ; do
echo Mounting $MDRV
if [ ! -d "$CHROOT$MDRV" ] ; then
mkdir -p "$CHROOT$MDRV"
fi
mount -o bind "$MDRV" "$CHROOT$MDRV"
done
mount (empty) partition, mount image file as loopback (mount -o loop <imagefile>, if I recall correctly), then cp -a content of the latter to the former.
then, You need to modify .chroot file, pointing it to partition, instead of image file (refer to Easy Debian wiki for instructions about that).
/Estel
// edit
As for backing it up to image file, You just revert whole process - mount Your file (clear files there, if it's not empty), cp -a everything from partition into file.
After trying, i just remembered what was my problem in doing this the last time i tried to. here is what i did:
mounted the image with:
mount -t ext4 /home/user/MyDocs/debian-m5.ext4 /media/debiantemp -o loop,noatime
made sure that /media/debiantemp is empty
used 'mount | grep debi' to make sure MyDocs and mmc are not mounted (if they are then they are being copied too).
mount /dev/mmcblk1p2 /.debian
cp -a /media/debiantemp/* /.debian
the copy starts and i can see that the process goes well until something like 10 - 15 minutes of copying and then my device crashes.
i retried this for 3 times - it seems to crash everytime - probably at the same point.
checking space taken in /media/debiantemp ~1.9GB
checking space taken in /.debian (after restart after crash) ~500mb
Am i doing something wrong?
Am i doing something wrong?Assuming your partition is big enough you're not doing anything wrong.
The problem is the N900's watchdog*, which is supposed to reboot your phone if some process stops responding. Whenever you put the N900's CPU on full load for more than (say) 10 minutes (which can easily be done by file system I/O) you'll likely trigger its reboot routine.
You can alleviate that problem by running every program that's likely to trigger the watchdog with a lower nice priority:nice -n 19 <your command>
Unfortunately that doesn't solve the problem because even if your program has a low priority (and therefore the watchdog will be served first) due to a high load caused by your program it might take too long to free the memory that the watchdog requires and the device will still reboot.
The only safe way to avoid the watchdog being triggered is to make sure the N900 won't be on full load for too long. So in your case it should help to not copy everything at once but to copy one (or some) directory at a time, wait until the load is down to normal again and then start the next one.
*) rant: In my opinion from a technical/hardware point of view the watchdog is the biggest flaw in the N900's design. I've had numerous reboots due to the watchdog, none of them being justified. There was no runaway process, I just wanted to do some task that did not finish within a few minutes (dist-upgrade Easy Debian, compile something, etc.).
So if someone has found a way to disable the watchdog completely (without putting the N900 in R&D mode) I'd be grateful to hear that!
@Estel:
While we're at ranting, I totally agree with your browser rant. Unfortunately text browsers like lynx or w3m don't work that great on the N900, but you might want to try chimera2 [1] in Easy Debian or dillo3 [2] once I've managed to dist-upgrade to Wheezy (which I can't do on the N900 due to the watchdog, therefore I've started to tinker with qemu on my PC).
Afair there was a "dillo3 on Maemo" thread some time ago (it might as well have been a youtube video I stumbled upon). Both browsers are severely limited in their functionality, but for simple browsing they should suffice.
There are also midori [3] and arora [4], which might be alternatives. They are supposed to be lightweight, but according to my experience that highly depends on your system and I haven't found out yet why that varies so much.
[1] http://packages.debian.org/squeeze/chimera2
[2] http://packages.debian.org/wheezy/dillo
[3] http://packages.debian.org/squeeze/midori
[4] http://packages.debian.org/squeeze/arora
freemangordon
2012-05-02, 09:14
by installing swappolube one should be able get rid of wd kicking
by installing swappolube one should be able get rid of wd kicking
I have it installed with recommended settings but i can't see any option to disable this watchdog - what excatly di you mean?
freemangordon
2012-05-02, 09:44
I have it installed with recommended settings but i can't see any option to disable this watchdog - what excatly di you mean?
you can't. could you try the settings bellow
echo 0 > /proc/sys/vm/page-cluster
echo 512 > /sys/block/mmcblk0/queue/nr_requests
echo 512 > /sys/block/mmcblk1/queue/nr_requests
echo 0 > /sys/block/mmcblk0/queue/iosched/slice_idle
echo 0 > /sys/block/mmcblk1/queue/iosched/slice_idle
echo 32 > /sys/block/mmcblk0/queue/iosched/quantum
echo 32 > /sys/block/mmcblk1/queue/iosched/quantum
echo 80 > /sys/block/mmcblk0/queue/iosched/fifo_expire_sync
echo 80 > /sys/block/mmcblk1/queue/iosched/fifo_expire_sync
echo 180 > /sys/block/mmcblk0/queue/iosched/fifo_expire_async
echo 180 > /sys/block/mmcblk1/queue/iosched/fifo_expire_async
you can't. could you try the settings bellow
echo 0 > /proc/sys/vm/page-cluster
echo 512 > /sys/block/mmcblk0/queue/nr_requests
echo 512 > /sys/block/mmcblk1/queue/nr_requests
echo 0 > /sys/block/mmcblk0/queue/iosched/slice_idle
echo 0 > /sys/block/mmcblk1/queue/iosched/slice_idle
echo 32 > /sys/block/mmcblk0/queue/iosched/quantum
echo 32 > /sys/block/mmcblk1/queue/iosched/quantum
echo 80 > /sys/block/mmcblk0/queue/iosched/fifo_expire_sync
echo 80 > /sys/block/mmcblk1/queue/iosched/fifo_expire_sync
echo 180 > /sys/block/mmcblk0/queue/iosched/fifo_expire_async
echo 180 > /sys/block/mmcblk1/queue/iosched/fifo_expire_async
It works :) just finished the copy process successfully.
Man you just solved a long time issue i had with this device - every time i tried to copy a big file it crashed and i assumed the problem was that the N900 can't handle such big files because of some buffering issue - since it happened i got scared everytime i needed to copy big files from one place to another, so i started to do this only using N900 -> PC -> N900 back which was an annoying process. i'm glad to finally realize the cause of this.
What exactly from this settings do i need to keep permanent and how? does this affect anything else?
Thanks for the help
It's worth to add, that having swap on microSD and ED partition on eMMC also helps with performance. Honestly, I haven't had any reboots, no matter if I lzma 3GB image file for ED (on device), copy <whatever> size/files count directory, or any other heavy I/O thing. Of course, I also modified files suggested by freemangordon, but I use slightly different settings (tweaked to my liking):
echo 5 > /proc/sys/vm/page-cluster
echo 256 > /sys/block/mmcblk0/queue/nr_requests
echo 512 > /sys/block/mmcblk1/queue/nr_requests
echo 7 > /sys/block/mmcblk0/queue/iosched/slice_idle
echo 7 > /sys/block/mmcblk1/queue/iosched/slice_idle
echo 4 > /sys/block/mmcblk0/queue/iosched/quantum
echo 4 > /sys/block/mmcblk1/queue/iosched/quantum
echo 125 > /sys/block/mmcblk0/queue/iosched/fifo_expire_sync
echo 125 > /sys/block/mmcblk1/queue/iosched/fifo_expire_sync
echo 250 > /sys/block/mmcblk0/queue/iosched/fifo_expire_async
echo 250 > /sys/block/mmcblk1/queue/iosched/fifo_expire_async
I must add, that combination of swap on (class 2) microSD and above settings, not only device doesn't reboot during heavy I/O, but is also fully responsible (seriously, @900 mhz, even during packaging ED image with lzma and compressions (1 above default), I can do other things on device without much problem!).
/Estel
It's worth to add, that having swap on microSD and ED partition on eMMC also helps with performance. Honestly, I haven't had any reboots, no matter if I lzma 3GB image file for ED (on device), copy <whatever> size/files count directory, or any other heavy I/O thing. Of course, I also modified files suggested by freemangordon, but I use slightly different settings (tweaked to my liking):
echo 5 > /proc/sys/vm/page-cluster
echo 256 > /sys/block/mmcblk0/queue/nr_requests
echo 512 > /sys/block/mmcblk1/queue/nr_requests
echo 7 > /sys/block/mmcblk0/queue/iosched/slice_idle
echo 7 > /sys/block/mmcblk1/queue/iosched/slice_idle
echo 4 > /sys/block/mmcblk0/queue/iosched/quantum
echo 4 > /sys/block/mmcblk1/queue/iosched/quantum
echo 125 > /sys/block/mmcblk0/queue/iosched/fifo_expire_sync
echo 125 > /sys/block/mmcblk1/queue/iosched/fifo_expire_sync
echo 250 > /sys/block/mmcblk0/queue/iosched/fifo_expire_async
echo 250 > /sys/block/mmcblk1/queue/iosched/fifo_expire_async
I must add, that combination of swap on (class 2) microSD and above settings, not only device doesn't reboot during heavy I/O, but is also fully responsible (seriously, @900 mhz, even during packaging ED image with lzma and compressions (1 above default), I can do other things on device without much problem!).
/Estel
Tested copying 3GB file from NTFS disk on key to MyDocs and the opposite with your settings - works great - device is fully responsive during the copy process.
Is there a thread for results of these values manipulations? are they explained somewhere or just google one by one :)
The latter, I'm afraid. It's quite required to understand what they're doing, as setting them properly is strictly correlated with Your usage patterns. BTW, that's why the default Nokia ones didn't worked well - we're using device way different, from what they though we may/will :)
Estel - i made a few performance tests - switched back and fourth from using the mounted image vs using directly from the partition.
I can say that launching debian applications from the partition is faster in about 5% - 10%, but there is one strange thing i found - chromium is VERY slow and almost non responsive. can that be related to these settings somehow?
when i used your previous image (chromium had a white theme) it worked much faster.
Do you have any idea why?
No idea, unfortunately... Working fine here. Maybe You've hit swap fragmentation? Tried after fresh reboot?
yep - didn't help. i'll play with it a little and see what i can find.
Thanks.
While trying to use apt-get update i get the following errors - this prevents from installing anything now:
-------------------------------
Setting up libmagickcore3 (8:6.6.0.4-3+squeeze3) ...
/var/lib/dpkg/info/libmagickcore3.postinst: line 5: /sbin/ldconfig: cannot execute binary file
dpkg: error processing libmagickcore3 (--configure):
subprocess installed post-installation script returned error exit status 126
dpkg: dependency problems prevent configuration of libmagickwand3:
libmagickwand3 depends on libmagickcore3 (>= 8:6.6.0.4); however:
Package libmagickcore3 is not configured yet.
dpkg: error processing libmagickwand3 (--configure):
dependency problems - leaving unconfigured
configured to not write apport reports
configured to not write apport reports
Errors were encountered while processing:
libmagickcore3
libmagickwand3
E: Sub-process /usr/bin/dpkg returned an error code (1)
-------------------------------
What is this about?
That looks odd. Can you install any other packages?
Is /sbin/ldconfig even in your image and if it is, is it executable?
That looks odd. Can you install any other packages?
Is /sbin/ldconfig even in your image and if it is, is it executable?
/sbin/ldconfig exists and yes - it has all executable permissions.
I am not able to install anything because i'm getting this error - i intended to install apache (which had failed of course) when i found this.
it seems i have mistaken - apache DID install, but it had a few errors. i checked again and it seems i am able to install/remove packages but this error always pops up at the end.
Haven't You though about starting from fresh image again? Lately, You seem to have many problems, that are hard to solve without knowing their source.
/Estel
Haven't You though about starting from fresh image again? Lately, You seem to have many problems, that are hard to solve without knowing their source.
/Estel
I had only just one more issue - which i did solve (with chromium being very slow).
Eventually if i won't have a choice i will try from the beginning but as a last resort since i will have to install/set up everything again and it'll take some work.
I hoped someone here did cross the issue/knows how to deal with such an error since i'm not too good at these deployment issues.
What was causing the Chromium slowdown issue for You? You haven't reported back about fixing it.
What was causing the Chromium slowdown issue for You? You haven't reported back about fixing it.
You're right - i forgot to update - this is generally what solved the problem (after many changes i tried that failed):
1. I erased the "/home/user/.config/chromium" folder and let the application create it from scratch.
2. Overwritten the default hosts file with mine (to avoid commercials).
3. Removed the offer to translate pages.
That's what happens when you move to a new image - you forget these little tweaks that you did long ago - that's why i'm trying to avoid it.
I just started to backup the debian partition on weekly basis - like maemo to avoid taking any chances, but with the image that you uploaded i did not do anything which should have caused the issue i'm now having with the installation - at least nothing that i know of.
I'll probably have to revert back to the original image, but that'll have to wait unfortunately because lack of time...
D@vIcHoJD
2012-05-07, 01:11
I have a question, if I delete the directory .config, not interfere with other applications or damaged?
Another question with the previous image, when I opened a file in OpenOffice or any other program could access to the directories (all folders with point i /home/user/Mydocs).documents, .sounds, now I can not access these just the other folders. As I can give permissions to access these from LXDE, or is an error in the image. I have debian-m5-estel.img.ext3.lzma.:confused::confused:
Thanks in advance.
D@vIcHoJD
2012-05-07, 21:36
Someone can help me with my questions, thanks
I will, but please do *not* double post like that without good reason, or Your posts will get ignored in future.
As for .config, it depends. If You have /home/ shared with Maemo, deleting .config *will* hurt for sure. If You have independent Maemo and ED /home/, deleting .config will not hurt Maemo, but will hurt Your ED :) unless You really know what You're doing and that it is what You need to achieve <something>.
/Estel
I will, but please do *not* double post like that without good reason, or Your posts will get ignored in future.
As for .config, it depends. If You have /home/ shared with Maemo, deleting .config *will* hurt for sure. If You have independent Maemo and ED /home/, deleting .config will not hurt Maemo, but will hurt Your ED :) unless You really know what You're doing and that it is what You need to achieve <something>.
/Estel
i i just realized i wrote "erase the .config folder" - i meant the .config/chromium of course :)
D@vIcHoJD
2012-05-08, 07:56
If I did so for chromium and LibreOffice, thank you very much for the help. Problem solved.
It seems to me that the md5sum in imagelist-fremantle-01 are not up to date.
I got for debian-m5-sulu.img.ext2.lzma
21c11b164acd0a730f740ceecd612da4
and for debian-m5-estel.img.ext3.lzma
cea873aa149dba449d3c54e7152fe12c
wheras in imagelist-fremantle-01
3,Estel,http://qole.org/files/,debian-m5-estel.img.ext3,lzma,Large Debian 6 with LibreOffice,75d4d0a24d1b62424919accd0a1be9af
4,sulu,http://qole.org/files/,debian-m5-sulu.img.ext2,lzma,Debian 6 image by sulu (DE),90dea3dad97b336fa82a14be50e8fce4
I had the same problem with this file and wrong md5sums. I got around it by disabling some lines in img-install/debian-installer.sh with a "#"
I disabled the line
mv $IMGLIST $IMGLIST.old
up to the line
exit
I then put the right md5sums in the file before starting the installer. Then it worked
But still the question remains: are the md5sums in the file up to date or aren't they?
Thanks, berny, I fixed the md5sums.
ArchiMark
2012-05-10, 17:44
Got ED working now using Estel's image file....(thanks, Estel!)....
Few questions please....
1. ED Wiki page says to initialize sound to do in terminal:
debbie
xmms2d -o pulse
I get 'debbie: command not found....
Any suggestions?
Thanks!
debbie works only from within Maemo (to go into chroot). If You're already inside ED chroot, don't use it.
/Estel
ArchiMark
2012-05-10, 19:48
debbie works only from within Maemo (to go into chroot). If You're already inside ED chroot, don't use it.
/Estel
Thanks for input....
So, I assume you mean that I use terminal in Maemo and issue the commands that way....
i just started to use a fresh image "debian-m5-estel.img" and without installing anything, just wanted to update all packages, so i used the following:
apt-get update
apt-get upgrade -f
apt-get upgrade --fix-missing (was suggested to run this in output of previous command)
now when i run "apt-get update" -> no errors - finishes with no suggested updates
running apt-get upgrade generates the following:
---
You might want to run 'apt-get -f install' to correct these.
The following packages have unmet dependencies:
file : Depends: libmagic1 (= 5.04-5+squeeze2) but 5.04-5 is installed
E: Unmet dependencies. Try using -f.
----
running "apt-get upgrade -f" generates the following:
----
Reading package lists... Done
Building dependency tree
Reading state information... Done
Correcting dependencies... Done
The following packages have been kept back:
debian-multimedia-keyring libsmbclient pulseaudio samba-common smbclient ttf-opensymbol
The following packages will be upgraded:
libmagic1 libssl0.9.8
2 upgraded, 0 newly installed, 0 to remove and 6 not upgraded.
70 not fully installed or removed.
Need to get 0 B/1,121 kB of archives.
After this operation, 0 B of additional disk space will be used.
Do you want to continue [Y/n]? y
E: read, still have 12 to read but none left
E: Error reading archive member header
E: Prior errors apply to /var/cache/apt/archives/libmagic1_5.04-5+squeeze2_armel.deb
debconf: apt-extracttemplates failed: Bad file descriptor
Preconfiguring packages ...
(Reading database ... 65036 files and directories currently installed.)
Preparing to replace libssl0.9.8 0.9.8o-4squeeze7 (using .../libssl0.9.8_0.9.8o-4squeeze12_armel.deb) ...
Unpacking replacement libssl0.9.8 ...
dpkg: error processing /var/cache/apt/archives/libssl0.9.8_0.9.8o-4squeeze12_armel.deb (--unpack):
corrupted filesystem tarfile - corrupted package archive
configured to not write apport reports
dpkg-deb: subprocess paste killed by signal (Broken pipe)
Preparing to replace libtiff4 3.9.4-5+squeeze4 (using .../libmagic1_5.04-5+squeeze2_armel.deb) ...
Unpacking replacement libtiff4 ...
Errors were encountered while processing:
/var/cache/apt/archives/libssl0.9.8_0.9.8o-4squeeze12_armel.deb
E: Sub-process /usr/bin/dpkg returned an error code (1)
----
Am i doing something wrong?
Need to get 0 B/1,121 kB of archives.
After this operation, 0 B of additional disk space will be used.
Do you want to continue [Y/n]? y
E: read, still have 12 to read but none left
E: Error reading archive member header
E: Prior errors apply to /var/cache/apt/archives/libmagic1_5.04-5+squeeze2_armel.deb
debconf: apt-extracttemplates failed: Bad file descriptorIt seems like apt thinks that you've already downloaded libmagic1 but it can't find the package.
Please runapt-get cleanand try to upgrade again!
I got little different question. One of Easy Debian dependencies is Murina Orange theme (for Maemo). It's great for many purposes - i.e. running Debian programs from within Maemo, via debbie - but it got big drawback. Somehow, it replaces "default" theme in situations, where Hildon GTK2 theme is unset - so, murina orange is seen in Truecrypt, or any other application with hildon theme unset before execution.
As Murina actually makes interface elements bigger, the result is, that many times important elements are "hidden", overlapped etc.
Any idea how to disable it for such situations, or at least, easily disable/enable it when needed?
/Estel
It seems like apt thinks that you've already downloaded libmagic1 but it can't find the package.
Please runapt-get cleanand try to upgrade again!
Thanks - problem solved.
Haven't You though about starting from fresh image again? Lately, You seem to have many problems, that are hard to solve without knowing their source.
/Estel
Started from a fresh image, installed all updates first and then installed my wanted packages.
No problems with the updates anymore.
ArchiMark
2012-05-11, 17:19
Two more ED questions please....using Estel's image file....
1. Tried watching a youtube video, first by using Epiphany browser....browser crashed, ie, window disappeared as it tried loading webpage with video....then tried Chromium browser, page loads, but don't see video, just black window with the little spinning circle thingie.....left window for long time (about 15-20 minutes...), but it never loads video....
Is this problem a function of limited functionality of ED and N900?
Anyway to get this working?
2. In order to log in to some webpages, I need to be able to type in my email address. However, when I try typing the '@' symbol, nothing happens....
How do I get this fixed?
Thanks!
Two more ED questions please....using Estel's image file....
1. Tried watching a youtube video, first by using Epiphany browser....browser crashed, ie, window disappeared as it tried loading webpage with video....then tried Chromium browser, page loads, but don't see video, just black window with the little spinning circle thingie.....left window for long time (about 15-20 minutes...), but it never loads video....
Is this problem a function of limited functionality of ED and N900?
Anyway to get this working?
There are reports of bringing it to working state, but it's still too much memory hog. For watching flash videos, microB is still recommended.
2. In order to log in to some webpages, I need to be able to type in my email address. However, when I try typing the '@' symbol, nothing happens....
How do I get this fixed?
Thanks!
copy Your (Maemo's) /usr/share/X11/xkb/symbols/nokia_vndr/rx-51 file into Your *debian's* [b]/usr/share/X11/xkb/symbols/nokia_vndr. From within Maemo, when Debian is mounted and .chrooted, it's /.debian//usr/share/X11/xkb/symbols/nokia_vndr, where You should copy file to. This way, you'll have exactly same hardware keyboard mapping, as in Maemo - of course, if You setxkbmap <your language> inside ED, or via Ed autostart script.
In layout that is found default in ED (before You copy rx-51), @ is just blue arrow + shift + space. Blue arrow + space is TAB. Default config is actually very useful, and contain many symbols assigned to blue arrow + shift + key, but I think that everyone will like own mapping's most.
/Estel
ArchiMark
2012-05-11, 21:30
There are reports of bringing it to working state, but it's still too much memory hog. For watching flash videos, microB is still recommended.
OK, thanks for explaining....
copy Your (Maemo's) /usr/share/X11/xkb/symbols/nokia_vndr/rx-51 file into Your *debian's* [b]/usr/share/X11/xkb/symbols/nokia_vndr. From within Maemo, when Debian is mounted and .chrooted, it's /.debian//usr/share/X11/xkb/symbols/nokia_vndr, where You should copy file to.
OK, done....
This way, you'll have exactly same hardware keyboard mapping, as in Maemo - of course, if You setxkbmap <your language> inside ED, or via Ed autostart script.
By 'ED autostart script', do you refer to
'/etc/xdg/lxsession/LXDE/autostart' or ?
If you meant in '/etc/xdg/lxsession/LXDE/autostart', then would I enter a line like this:
@setxkbmap us
Or ?
In layout that is found default in ED (before You copy rx-51), @ is just blue arrow + shift + space. Blue arrow + space is TAB. Default config is actually very useful, and contain many symbols assigned to blue arrow + shift + key, but I think that everyone will like own mapping's most.
/Estel
Thanks for the info...it worked, but as you say, would prefer US keyboard mapping as I'm used to it now...
Yes, I mean file You've mentioned. It isn't working for You, or You've asked just in case?
ArchiMark
2012-05-12, 21:12
Yes, I mean file You've mentioned. It isn't working for You, or You've asked just in case?
Thanks for confirming file....
I just wanted to make sure that I am doing the right thing....trying to avoid making more mistakes and screw-ups than I already do..... :rolleyes:
So, I edited that file as I noted in my previous post and keyboard mapping is good no, ie, same as in Maemo....
Thanks!
ArchiMark
2012-05-13, 03:22
Got another issue now...hope someone can please advise....
ED (Estel's image file...) was running OK....until I did 'apt-get upgrade'....seemed to go OK, but while it was in process of unpacking and installing some items, I left the room for a few minutes...and when I returned N900 had rebooted into backupmenu....so, I assume that ED crashed and N900 reboot or ?
:confused:
In any event, when I rebooted, it booted up to Maemo OK....everything seems OK....
However, now when I click on 'Debian LXDE' it looks like it's going to open....ie, display goes black, I see 'Debian LXDE' at top of display with little spinning wheel thingie next to it....does that for about 15 seconds and black window disappears and I'm back at Maemo desktop.....
Any suggestions as to how to get Debian working again???
Thanks!
try reinstalling debian choot this may help .i hope
i think maybe you did something wrong to the image if reinstalling DC didnt solve the problem then you may need a new image or wait for Estel's reply he may got a solution for you
./sifo
ArchiMark
2012-05-13, 04:36
try reinstalling debian choot this may help .i hope
i think maybe you did something wrong to the image if reinstalling DC didnt solve the problem then you may need a new image or wait for Estel's reply he may got a solution for you
./sifo
Thanks for your help.....tried reinstalling Debian chroot, but still no go.....
???
It doewsn't have anything to do with chroot, neither any package was broken. It's jsut that, due to bad (Maemo, not Debian - on last few pages of this thread You've listed swap settings You should use, both Maemo and ED benefit from it) swap-related settings, + swap on eMMC instead of microSD.
This made device unresponsive for too long during Debian's apt-get upgrade, and rebooted device. In turn, this FCKD Your ED image. You may try to fix it with fsck.ext* /dev/mmcblk*p* (put number instead of *, corresponding to ext type You've used for Debian image, and provide proper storage/partition numbers) - remember, to try it on *unmounted* debian image.
In case fsck doesn't help, You must start with new image from scratch. Unfortunately, it's common thing to happen - although, after unexpected reboot, I perform fsck from backupmenu before booting (I got ED started at every boot, via event.d script), thus I've never had image corruption that would destroy my ED image.
Generally, if - after crash - You fsck image before running Ed again, 99% of times You'll be fine. ho ever, if You mounted it already and tried to proceed, it may be screwed beyond repair. Just try, and report back.
/Estel
ArchiMark
2012-05-13, 18:48
It doewsn't have anything to do with chroot, neither any package was broken. It's jsut that, due to bad (Maemo, not Debian - on last few pages of this thread You've listed swap settings You should use, both Maemo and ED benefit from it) swap-related settings, + swap on eMMC instead of microSD.
This made device unresponsive for too long during Debian's apt-get upgrade, and rebooted device. In turn, this FCKD Your ED image. You may try to fix it with fsck.ext* /dev/mmcblk*p* (put number instead of *, corresponding to ext type You've used for Debian image, and provide proper storage/partition numbers) - remember, to try it on *unmounted* debian image.
In case fsck doesn't help, You must start with new image from scratch. Unfortunately, it's common thing to happen - although, after unexpected reboot, I perform fsck from backupmenu before booting (I got ED started at every boot, via event.d script), thus I've never had image corruption that would destroy my ED image.
Generally, if - after crash - You fsck image before running Ed again, 99% of times You'll be fine. ho ever, if You mounted it already and tried to proceed, it may be screwed beyond repair. Just try, and report back.
/Estel
Thanks for your detail reply...
Decided to be safe, so, I deleted current image file....then copied over the file again from my microSD card....however, it started copying file and since it's such a large file, I stepped away from my N900 for a few minutes...came back and I saw the backupmenu again!
So, something happened and N900 rebooted again....
I choose backupmenu option to do fsck, it found some errors, did it again, and this time no errors......then rebooted...
N900 booted up OK to Maemo...I see that 108.3MB of image file was copied over before N900 had crashed...
Copied over estel image file again from microSD to N900...tried ED and voila! it is working again....
Regarding swap, you are saying that I should use eMMC for swap space, not microSD, correct?
What settings would you recommend for this? (you mention that I had stated my swap settings on recent previous post, but I think that was probably someone else....)
If I have swap set up OK, then it should be OK to do 'apt-get upgrade'? (upgrade was downloading about 220MB....)
Thanks!
Nope, I recommend You to use microSD for swap, and absolutely *no* swap on eMMC.
As for recommended settings, impeham asked for them just few pages ago, please search in this very thread.
/Estel
ArchiMark
2012-05-13, 21:22
Nope, I recommend You to use microSD for swap, and absolutely *no* swap on eMMC.
OK, got it now....somehow I thought you meant the opposite.... :o
As for recommended settings, impeham asked for them just few pages ago, please search in this very thread.
/Estel
Yes, saw that and this too:
http://wiki.maemo.org/Swap_on_microSD
Thanks again for all your great help!
I tried to run gCompris via Easy Debian - my Son loves to play it, and Maemo version is quite outdated. It works ok (although, loads quite slow initially), but sound isn't working.
Here is error message, as seen in console:
(gst-plugin-scanner:10549): GStreamer-WARNING **: Failed to load plugin '/usr/lib/gstreamer-0.10/libgstpulse.so': /usr/lib/gstreamer-0.10/libgstpulse.so: symbol pa_stream_cancel_write, version PULSE_0 not defined in file libpulse.so.0 with link time reference
(gcompris:10548): Gtk-CRITICAL **: gtk_window_set_type_hint: assertion `!gtk_widget_get_mapped (GTK_WIDGET (window))' failed
Any ideas, how to fix it? I would really, really be grateful for any advices. Other than sound problems, it's really comfortable to play.
/Estel
There are reports of bringing it to working state, but it's still too much memory hog. For watching flash videos, microB is still recommended.The main problem is that ED has no HW acceleration for video decoding (and never will have). Therefore video in ED will never actually work because the Cortex A8 is just too slow to decode videos in real time (maybe except for some 240p videos).
Generally, if - after crash - You fsck image before running Ed again, 99% of times You'll be fine.Just a suggestion for the next image:
Having an ext3 file system is not optimal. Use ext2 if you want to reduce flash write cycles caused by journaling or use ext4 if you want journaling. fsck.ext4 is way faster than fsck.ext3.
I don't know if the stock Maemo kernel fully supports ext4, but I guess most - if not all - ED users use KP anyway.
I tried to run gCompris via Easy Debian - my Son loves to play it, and Maemo version is quite outdated. It works ok (although, loads quite slow initially), but sound isn't working.
Here is error message, as seen in console:
(gst-plugin-scanner:10549): GStreamer-WARNING **: Failed to load plugin '/usr/lib/gstreamer-0.10/libgstpulse.so': /usr/lib/gstreamer-0.10/libgstpulse.so: symbol pa_stream_cancel_write, version PULSE_0 not defined in file libpulse.so.0 with link time reference
(gcompris:10548): Gtk-CRITICAL **: gtk_window_set_type_hint: assertion `!gtk_widget_get_mapped (GTK_WIDGET (window))' failed
Any ideas, how to fix it? I would really, really be grateful for any advices. Other than sound problems, it's really comfortable to play.
/EstelI think I can clear this up. Do you remember that I had to macgyver a fake Squeeze pulseaudio package that actually contained the Lenny version to stay compatible to Maemo's pulseaudio protocol version?
So what you have now is a Lenny pulseaudio with a Squeeze gcompris that thinks it can rely on a Squeeze pulseaudio, which in this case causes trouble.
I believe I have a better idea for a new pulseaudio package which actually contains the correct pulseaudio version and is compatible to Maemo. But in order to do that I have to recompile pulseaudio, which I can't do on the N900 due to the watchdog. I already have an armel qemu VM set up but that comes with huge performance penalties (8-10% of the x86 host CPU seems to be reasonable), so I intend to wait with any actual work until Wheezy is frozen, which is scheduled for June.
I will then try to create a minimalistic Wheezy image containing the proper (but recompiled) pulseaudio which will be able to run iceweasel 10 too btw. (I already tried to backport iceweasel 10 to Squeeze by reproducing what the people at mozilla.debian.net are doing for i386 and amd64 but gcc4.4 for armel is too buggy for that and I didn't succeed in compiling gcc4.6 on Squeeze armel either, so I finally gave up on that.)
ArchiMark
2012-05-14, 20:15
Just curious as to what programs people are using besides the ones that come with the ED image file?
Hope it's OK to ask here, if not. I can start a new topic....
FWIW, I installed gcc/g++, make, cmake in ED.
Then I was able to compile the CLI ToDo program 'Task' <taskwarrior.org> on my N900, very cool!
So, what programs do you use in ED?
vBulletin® v3.8.8, Copyright ©2000-2025, vBulletin Solutions, Inc.