maemo.org - Talk

maemo.org - Talk (https://talk.maemo.org/index.php)
-   Maemo 5 / Fremantle (https://talk.maemo.org/forumdisplay.php?f=40)
-   -   Marmistrz's failed devel package - unexpected results/conclusions (https://talk.maemo.org/showthread.php?t=83948)

szopin 2012-04-28 13:57

Marmistrz's failed devel package - unexpected results/conclusions
 
So a devel package can cause so much confusion (and reboot loops leading to reflashes for the unlucky guys that have no backupmenu installed, see: http://talk.maemo.org/showthread.php?t=83939).
Was the autobuilder set up in this way or a bug? If setup, was that a concious decision or an overlook?
Are there no backup/safety mechanisms? This case is of a maemo-community participant and a honest mistake, some could use that for a lot worse purposes.
Is this issue for council to decide/fix? Nokia? Community?

Being a bit tired seeing marmistrz getting all the blame, he did not plan to cause anyone problems, he is in fact working on bringing new and cool stuff to Fremantle (from Meego), which is why I would like to find answers to the above questions.

Will OBS fix those???

Hurrian 2012-04-28 14:01

Re: Marmistrz's failed devel package - unexpected results/conclusions
 
First, it means that, even if it's extras-devel, uploading packages to replace system packages (and you are not the maintainer) is stupid.
At least name it libxau6-power or something. Conflicts/replaces libxau6.

Second, it's a reminder that extras-devel is -devel.

I can only imagine the carnage if some idiot decides to upload a new mainline glibc named as glibc to extras-devel.

szopin 2012-04-28 14:04

Re: Marmistrz's failed devel package - unexpected results/conclusions
 
Hurrian: Funny, as that should allow Aapo to bring gcc/++ 4.6 into devel, yet there are blocks preventing him from it. Why some critical/system packages can and some cannot then?

EDIT: As per your edit, glibc, libstdc, gcc... somehow are safe. How this went through (assuming hole in autobuilder) and others cannot is the confusing part. We would have latest GCC by now if it was that easy

Hurrian 2012-04-28 14:06

Re: Marmistrz's failed devel package - unexpected results/conclusions
 
If he does get gcc-4.6 to work, then package it and send into the repo as gcc-4.6, just like Debian. Over time, the gcc metapackage could probably be edited to request 4.6 instead of 4.2.

szopin 2012-04-28 14:08

Re: Marmistrz's failed devel package - unexpected results/conclusions
 
http://maemo.org/packages/view/gcc-4.6/
gets empty (IIRC)

Discussion about this and problems with it:
http://talk.maemo.org/showthread.php?t=81861

tl;dr As autobuilder uses gcc 4.2 this is somehow banned

Also to answer your post, his 4.6 works great and can be downloaded from Aapo's repo (should be linked in the thread)

misiak 2012-04-28 14:36

Re: Marmistrz's failed devel package - unexpected results/conclusions
 
Quote:

Originally Posted by Hurrian (Post 1198684)
First, it means that, even if it's extras-devel, uploading packages to replace system packages (and you are not the maintainer) is stupid.
At least name it libxau6-power or something. Conflicts/replaces libxau6.

That's exactly what I said :D

It's super-amazing that noone got an idea yet to create a package with postinstall script "rm -rf /" and upload it to extras-devel with name maemo-fremantle-pr ;P

Edit: Hurrican, are you kind of super-user? You don't have "Thanks!" button below your posts...

marmistrz 2012-04-28 14:52

Re: Marmistrz's failed devel package - unexpected results/conclusions
 
Quote:

Originally Posted by misiak (Post 1198698)
That's exactly what I said :D

It's super-amazing that noone got an idea yet to create a package with postinstall script "rm -rf /" and upload it to extras-devel with name maemo-fremantle-pr ;P

Edit: Hurrican, are you kind of super-user? You don't have "Thanks!" button below your posts...

He must've not confirmed his email, I had similar problem.

marmistrz 2012-04-28 14:53

Re: Marmistrz's failed devel package - unexpected results/conclusions
 
I found a strange thing in my libxau6 log:
Code:

[2012-04-27 17:52:01] Processing package libxau 1.0.6-1-meego1141+0m6. Uploader: marmistrz, builder: builder1
[2012-04-27 17:52:04] Building libxau 1.0.6-1-meego1141+0m6 for target 'maemo-fremantle-armel-extras-devel'
[2012-04-27 17:53:06] OK
[2012-04-27 17:53:06] ERROR running /etc/buildme.d/check_build: Package libxau provides binary package libxau6 which is also available on the device or Nokia repository. Build of this package has been prevented.
[2012-04-27 17:53:06] Building libxau 1.0.6-1-meego1141+0m6 for target 'maemo-fremantle-i386-extras-devel'
[2012-04-27 17:53:49] Signing build results
[2012-04-27 17:53:49] libxau 2:1.0.6-1-meego1141+0m6 has been queued for loading into fremantle extras-devel repository
[2012-04-27 17:53:49] You can find more info about this build here: https://garage.maemo.org/builder/fremantle/libxau_1.0.6-1-meego1141+0m6/

check_build isn't running properly

szopin 2012-04-28 15:07

Re: Marmistrz's failed devel package - unexpected results/conclusions
 
Quote:

Originally Posted by misiak (Post 1198698)
That's exactly what I said :D

It's super-amazing that noone got an idea yet to create a package with postinstall script "rm -rf /" and upload it to extras-devel with name maemo-fremantle-pr ;P

Edit: Hurrican, are you kind of super-user? You don't have "Thanks!" button below your posts...

That would be instantly called out on this forum. Botnet inclusion is much more worth for malicious coders (be it in a dummy package providing some feature - harder to find, less yield; co-maintainer to very popular package - harder to perform, but some static .so inclusion that fixes a bug included previously should work easily). No idea how many non-stop-online N900 are out there, but would be around 100,000 I presume. Not bad for a bot-net (a mobile one at that)

marmistrz 2012-04-28 15:12

Re: Marmistrz's failed devel package - unexpected results/conclusions
 
Quote:

Originally Posted by szopin (Post 1198706)
That would be instantly called out on this forum. Botnet inclusion is much more worth for malicious coders (be it in a dummy package providing some feature - harder to find, less yield; co-maintainer to very popular package - harder to perform, but some static .so inclusion that fixes a bug included previously should work easily). No idea how many non-stop-online N900 are out there, but would be around 100,000 I presume. Not bad for a bot-net (a mobile one at that)

There was some app maemo-profiler: without a good .desktop file (named QtApplication or something) and manual running causes nothing.


All times are GMT. The time now is 10:25.

vBulletin® Version 3.8.8