[pacman-dev] Bug in pactest suite
Hi! I noticed, that with fileconflict*.py the .FILELISTs of generated packages isn't generated correctly: the symlinks are missing. So we can ignore the results of these pactests now. Bye
On Dec 23, 2007 6:55 AM, Nagy Gabor
Hi! I noticed, that with fileconflict*.py the .FILELISTs of generated packages isn't generated correctly: the symlinks are missing. So we can ignore the results of these pactests now.
.FILELIST was killed off ages ago. I think even before pacman 3.0 Considering you tracked this bug down, could you maybe... I dunno, even give us some sort of source line or something if you're not going to provide a patch?
On Dec 23, 2007 6:55 AM, Nagy Gabor
wrote: Hi! I noticed, that with fileconflict*.py the .FILELISTs of generated packages isn't generated correctly: the symlinks are missing. So we can ignore the results of these pactests now.
.FILELIST was killed off ages ago. I think even before pacman 3.0
Considering you tracked this bug down, could you maybe... I dunno, even give us some sort of source line or something if you're not going to provide a patch?
Well, I looked into _alpm_pkg_load, and as I see, it uses .FILELIST if that is provided in the archive, and auto-generates it if it is missing. The bug is in the testsuite, and since I am not familiar in python, I cannot provide a patch for fixing this. Steps to reproduce: run pactest on fileconflict002.py, and check the generated package file in pactest/root/tmp. Bye
On Dec 23, 2007 2:01 PM, Aaron Griffin
On Dec 23, 2007 6:55 AM, Nagy Gabor
wrote: Hi! I noticed, that with fileconflict*.py the .FILELISTs of generated packages isn't generated correctly: the symlinks are missing. So we can ignore the results of these pactests now.
.FILELIST was killed off ages ago. I think even before pacman 3.0
Considering you tracked this bug down, could you maybe... I dunno, even give us some sort of source line or something if you're not going to provide a patch?
Um...I've thought it should be killed off, but that really wasn't done. http://projects.archlinux.org/git/?p=pacman.git;a=blob;f=scripts/makepkg.sh.... It is true that we no longer require it for pacman. Has it been long enough then that we can safely remove it from makepkg (meaning there are no holdovers from 2.9.8 that we want to worry about)? -Dan
On Dec 23, 2007 6:23 PM, Dan McGee
On Dec 23, 2007 2:01 PM, Aaron Griffin
wrote: On Dec 23, 2007 6:55 AM, Nagy Gabor
wrote: Hi! I noticed, that with fileconflict*.py the .FILELISTs of generated packages isn't generated correctly: the symlinks are missing. So we can ignore the results of these pactests now.
.FILELIST was killed off ages ago. I think even before pacman 3.0
Considering you tracked this bug down, could you maybe... I dunno, even give us some sort of source line or something if you're not going to provide a patch?
Um...I've thought it should be killed off, but that really wasn't done. http://projects.archlinux.org/git/?p=pacman.git;a=blob;f=scripts/makepkg.sh....
It is true that we no longer require it for pacman. Has it been long enough then that we can safely remove it from makepkg (meaning there are no holdovers from 2.9.8 that we want to worry about)?
Aaron, can you give me some feedback here? -Dan
On Dec 23, 2007 6:23 PM, Dan McGee
On Dec 23, 2007 2:01 PM, Aaron Griffin
wrote: On Dec 23, 2007 6:55 AM, Nagy Gabor
wrote: Hi! I noticed, that with fileconflict*.py the .FILELISTs of generated packages isn't generated correctly: the symlinks are missing. So we can ignore the results of these pactests now.
.FILELIST was killed off ages ago. I think even before pacman 3.0
Considering you tracked this bug down, could you maybe... I dunno, even give us some sort of source line or something if you're not going to provide a patch?
Um...I've thought it should be killed off, but that really wasn't done. http://projects.archlinux.org/git/?p=pacman.git;a=blob;f=scripts/makepkg.sh....
It is true that we no longer require it for pacman. Has it been long enough then that we can safely remove it from makepkg (meaning there are no holdovers from 2.9.8 that we want to worry about)?
Oh yes, I am merging two things here. We removed "filelist" generation from makepkg, and skipped the dependence on .FILELIST in pacman itself. I'd say we can get rid of the generation in makepkg and always generate it on install. It makes more sense to me. Additionally, regarding non-3.0 holdovers: they're screwed anyway with current/core rename and all that, so they're non-issues.
This is something pacman can do on its own straight from the archive, and we
will reduce the chance of problems occurring becuase of inproper FILELIST
generation as we have had in the past with special characters in filenames.
Once we remove it from makepkg. we can remove any usage of it from all of
our other tools, including pacman, pactest, and contrib/ utilities.
Note that removing it from pacman uncovered a few other bugs anyway, so this
was probably a good move.
Signed-off-by: Dan McGee
participants (4)
-
Aaron Griffin
-
Dan McGee
-
Dan McGee
-
Nagy Gabor