Active Topics
-
Porting apps to Leste (32)
to Maemo 7 / Leste by Arno_11 - 5 days, 17 hrs ago - more...
misiak |
2012-06-14
, 16:05
|
|
Posts: 804 |
Thanked: 1,598 times |
Joined on Feb 2010
@ Gdynia, Poland
|
#111
|
I will quote my post from Meecolay development thread, as marmistrz said it may be useful to you:
Originally Posted by misiak
The site I'm talking about is http://linuxtesting.org/compatibilit...at_report.html . The report with problems for libc.so.6 is accessible here: http://linuxtesting.org/compatibilit..._Risk_Problems (if you click at "+" signs near functions, you will see the exact explanations of differences). This may explain why new libc crashes some apps
| The Following 4 Users Say Thank You to misiak For This Useful Post: | ||
|
|
2012-06-14
, 21:57
|
|
Posts: 2,076 |
Thanked: 3,268 times |
Joined on Feb 2011
|
#112
|
Originally Posted by Android_808
This just runs flash.py with standard images:
Something tells me that someone has already done something like this. I can't remember the exact page I found but this may be helpful:
https://wiki.mozilla.org/ReferencePl...ing/Nokia_N900
At this point you need to obtain the following files:
* moz-n900-vNN.ubi - mozilla custom firmware
* RX-51_2009SE_1.2009.41-1.VANILLA_PR_EMMC_MR0_ARM.bin
* RX-51_2009SE_2.2009.51-1_PR_COMBINED_MR0_ARM.bin
Except for OSX specific moz... it still flashes standard bins, or maybe I missed something. Still would be nice to create flashable images for exchange (full images, here try thumb2 CSSU with KP51testing and full SDK chroot, wait, use this Ubuntu12 with full -dev libs... hope this is not a dream)
| The Following User Says Thank You to szopin For This Useful Post: | ||
|
|
2012-06-18
, 23:25
|
|
Posts: 362 |
Thanked: 426 times |
Joined on Nov 2010
@ Italy, Lombardia
|
#113
|
Originally Posted by Hurrian
UbuntuArm 12.10 (Alpha) which uses GCC 4.7.x is faster then UbuntuArm 12.04 which uses GCC 4.6.x
Side note - GCC 4.7.0 is available. Anyone up to try that?
http://www.phoronix.com/scan.php?pag...210_arm1&num=1
Not all performance increments are due only to newer GCC version, but surely the newer GCC plays a big role
Side note - GCC 4.7.1 with 100+ bug fixes is available
| The Following User Says Thank You to Fabry For This Useful Post: | ||
|
|
2012-06-19
, 16:26
|
|
Posts: 3,328 |
Thanked: 4,476 times |
Joined on May 2011
@ Poland
|
#114
|
I compiled newer libdsme0.2.0, for compatibility with meego.
Can anyone test it?
If there are any problems, you can try commenting out maemo-optify from debian/rules and recompiling it.
It's here: http://marmistrz.net63.net/public/dsme/
Can anyone test it?
If there are any problems, you can try commenting out maemo-optify from debian/rules and recompiling it.
It's here: http://marmistrz.net63.net/public/dsme/
|
|
2012-06-20
, 05:40
|
|
Posts: 2,225 |
Thanked: 3,822 times |
Joined on Jun 2010
@ Florida
|
#115
|
Wow, finally caught up on this thread after my mostly-absense from TMO, and I am fascinated by how much progress has been made so far.
You know, on the one hand, I am disappointed with my failure to get involved in this, on the other hand, it seems like the entire endeavor of getting newer gcc, libc, etc, onto the N900 let alone getting to the point of succesfully replacing the older ones completely, is way out of my league.
I just want you all working on this to know I appreciate your work on this matter. If any of you take donations let me know, and in the meantime if/when I can, I will attempt to jump into this and contribute.
You know, on the one hand, I am disappointed with my failure to get involved in this, on the other hand, it seems like the entire endeavor of getting newer gcc, libc, etc, onto the N900 let alone getting to the point of succesfully replacing the older ones completely, is way out of my league.
I just want you all working on this to know I appreciate your work on this matter. If any of you take donations let me know, and in the meantime if/when I can, I will attempt to jump into this and contribute.
|
|
2012-06-20
, 11:49
|
|
Posts: 838 |
Thanked: 3,384 times |
Joined on Mar 2009
|
#116
|
First testing of gcc-4.7.1. I don't have debs yet. I have only tested: 'gcc --version'.
Run on scratchbox1:
Run on scratchbox1:
Code:
wget http://ftp.de.debian.org/debian/pool/main/g/gcc-4.7/gcc-4.7_4.7.1-1.diff.gz
wget http://ftp.de.debian.org/debian/pool/main/g/gcc-4.7/gcc-4.7_4.7.1.orig.tar.gz
gunzip gcc-4.7_4.7.1-1.diff.gz
tar xvf gcc-4.7_4.7.1.orig.tar.gz
mv gcc-4.7-4.7.1.orig gcc-4.7-4.7.1
patch -p 0 < gcc-4.7_4.7.1-1.diff
cd gcc-4.7-4.7.1/
---
#change distrelease
sed -i 's/:= sid/:= fremantle/g' debian/rules.defs
#no multiarch
sed -i 's/with_multiarch_lib := yes/with_multiarch_lib := no/g' debian/rules.defs
sed -i 's/multiarch_stage1 := yes/multiarch_stage1 := no/g' debian/rules.defs
#do not use gcc-4.4
sed -i 's/gcc-4.4/gcc/g' debian/rules2
#lower some build dependencies
sed -i 's/AUTO_BUILD_DEP += autoconf2.64, automake (>= 1:1.11), automake (<< 1:1.12),/AUTO_BUILD_DEP += autoconf, automake (>= 1:1.10), automake (<< 1:1.12),/g' debian/rules.conf
sed -i 's/make (>= 3.81)/make (>= 3.81) | make-3.81/g' debian/control.m4
sed -i 's/bison (>= 1:2.3)/bison (>= 1:2.3) | bison-2.4/g' debian/control.m4
sed -i 's/g++-4.4/g++/g' debian/control.m4
sed -i 's/autoconf2.64/autoconf/g' debian/control
sed -i 's/automake (>= 1:1.11)/automake/g' debian/control
#lower versions of documentation tools (and drop gsfonts-x11)
sed -i 's/doxygen (>= 1.7.2)/doxygen (>= 1.5.1)/g' debian/rules.conf
sed -i 's/gsfonts-x11, //g' debian/rules.conf
sed -i 's/doxygen (>= 1.7.2)/doxygen (>= 1.5.1)/g' debian/control
sed -i 's/gsfonts-x11, //g' debian/control
#lower some dependency-libraries
sed -i 's/GMP_BUILD_DEP = libgmp-dev (>= 2:5.0.1~),/GMP_BUILD_DEP = libgmp3-dev,/g' debian/rules.conf
sed -i 's/MPFR_BUILD_DEP = libmpfr-dev (>= 3.0.0-9~),/MPFR_BUILD_DEP = libmpfr-dev,/g' debian/rules.conf
sed -i 's/libelfg0-dev (>= 0.8.12),/libelfg0-dev,/g' debian/rules.conf
sed -i 's/libgmp-dev (>= 2:5.0.1~)/libgmp3-dev/g' debian/control
sed -i 's/libmpfr-dev (>= 3.0.0-9~)/libmpfr-dev/g' debian/control
sed -i 's/libelfg0-dev (>= 0.8.12)/libelfg0-dev/g' debian/control
#remove multilib
sed -i 's/GCC_MULTILIB_BUILD_DEP = g++-multilib \[\$(multilib_archs)\]/GCC_MULTILIB_BUILD_DEP =/g' debian/rules.conf
#lower dependencies on part2
sed -i 's/dpkg-dev (>= 1.16.0~ubuntu4)/dpkg-dev/g' debian/rules.conf
sed -i 's/dpkg-dev (>= 1.16.0~ubuntu4)/dpkg-dev/g' debian/control
sed -i 's/debhelper (>= 5.0.62)/debhelper/g' debian/control
#lower dependency of libc6
sed -i 's/libc6-dev (>= 2.13-5)/libc6-dev/g' debian/control
sed -i 's/libc_ver := 2.11/libc_ver := 2.5/g' debian/rules.conf
sed -i 's/libc6-dev (>= 2.13-31) \[armel armhf\]/libc6-dev [armel armhf]/g' debian/control
sed -i 's/libc6-dev (>= 2.13-31) \[armel armhf\]/libc6-dev [armel armhf]/g' debian/rules.conf
#do not use gnu-unique-object (GLIBC 2.11 or above)
sed -i 's/CONFARGS += --enable-gnu-unique-object/CONFARGS += --disable-gnu-unique-object/g' debian/rules2
# This bug is marked invalid on upstream: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52218
# and it is fixed on newer libc: http://sourceware.org/ml/libc-ports/2012-02/msg00079.htlibml
#add block to one existing/used patch-file
echo "
--- a/src/libgo/runtime/proc.c 2012-06-20 14:03:10.721499031 +0300
+++ b/src/libgo/runtime/proc.c 2012-06-20 14:03:41.421651325 +0300
@@ -111,7 +111,15 @@
# else
-# error unknown case for SETCONTEXT_CLOBBERS_TLS
+static inline void
+initcontext(void)
+{
+}
+
+static inline void
+fixcontext(ucontext_t *c __attribute__ ((unused)))
+{
+}
# endif
" >> debian/patches/libgo-fix-arm.diff
#build
fakeroot dpkg-buildpackage
| The Following 7 Users Say Thank You to AapoRantalainen For This Useful Post: | ||
|
|
2012-06-20
, 13:21
|
|
Posts: 3,328 |
Thanked: 4,476 times |
Joined on May 2011
@ Poland
|
#117
|
Originally Posted by Mentalist Traceur
For donation link, look at my signature.
Wow, finally caught up on this thread after my mostly-absense from TMO, and I am fascinated by how much progress has been made so far.
You know, on the one hand, I am disappointed with my failure to get involved in this, on the other hand, it seems like the entire endeavor of getting newer gcc, libc, etc, onto the N900 let alone getting to the point of succesfully replacing the older ones completely, is way out of my league.
I just want you all working on this to know I appreciate your work on this matter. If any of you take donations let me know, and in the meantime if/when I can, I will attempt to jump into this and contribute.
There'll be some more dsme packages, now I'm waiting whether the maintainer of libbme from gitorious adds debian/*
I'm uploading newer libiphb, the same link as here: http://talk.maemo.org/showpost.php?p...&postcount=114
|
|
2012-06-25
, 11:58
|
|
Posts: 8 |
Thanked: 71 times |
Joined on Jun 2012
|
#118
|
I've built scratchbox toolchains for the Maemo5 SDK based on Linaro 4.6/7-2012.06, they're available here:
http://www.snewbury.org.uk//maemo/li...armv7a.tar.bz2
and
http://www.snewbury.org.uk//maemo/li...armv7a.tar.bz2
Just unpack into "/scratchbox/compilers", and they'll appear in the list when setting up a target.
http://www.snewbury.org.uk//maemo/li...armv7a.tar.bz2
and
http://www.snewbury.org.uk//maemo/li...armv7a.tar.bz2
Just unpack into "/scratchbox/compilers", and they'll appear in the list when setting up a target.
| The Following 10 Users Say Thank You to zeq For This Useful Post: | ||
|
|
2012-06-26
, 06:54
|
|
Posts: 8 |
Thanked: 71 times |
Joined on Jun 2012
|
#119
|
Mainline gcc-4.6.3 is now also available at the same location.
| The Following 2 Users Say Thank You to zeq For This Useful Post: | ||
|
|
2012-06-26
, 18:12
|
|
Posts: 3,328 |
Thanked: 4,476 times |
Joined on May 2011
@ Poland
|
#120
|
I'm trying to compile glib 2.28 from Harmattan, but the autotools are detected as too old, even though I installed automake1.10-autoconf264 and autoconf2.64.
What am I doing wrong?
What am I doing wrong?