0d7bba65c9
The previous code is below. I could not understand this code. It was written when conflicts were string. It traverses between the conflict strings of the to be installed package and then for every conflict it always looks if the _to be installed package_ is installed. If it is installed then raises conflict with the conflicting package ????
Faik Uygur
2006-10-26 18:30:11 +00:00
ffc64fc06f
* We will pass the conflicting package if it is in the order at the first check, because the package that is installed may be a conflicting one but the same package in the order (that is going to be installed) may not be a conflicting one. The second test shows us that. If it is then nothing can be done.
Faik Uygur
2006-10-26 10:58:39 +00:00
fcc40af294
* this is more real, and asks what i thought :(
Faik Uygur
2006-10-26 10:19:33 +00:00
9c7fdf5195
* scenarios never lies :) this is our real case.
Faik Uygur
2006-10-26 10:06:52 +00:00
4485d78c5f
* rebuildb cause packagedb info has changed. Conflicts are not strings anymore they are objects.
Faik Uygur
2006-10-26 10:53:57 +00:00
291d977ab5
* do not set to SIG_DFL, set it to what handler was used before.
Faik Uygur
2006-10-25 13:27:39 +00:00
f0a78b2d09
* Alright. extract_install may become a longstanding operation so it is not fair to disable keyboard interrupts before that. The real critical path contains store_pisi_files which can lead to a situation that the new package's pisi files extracted and old package's are deleted from /var/lib/package but the db may not be updated. So db thinks the old version is installed but pisi files are deleted without its knowledge. And this leads to that nasty bug.
Faik Uygur
2006-10-24 22:59:22 +00:00
3c1ec46433
patches after builddeps...
Barış Metin
2006-10-24 18:46:15 +00:00
5316400724
Change order of the tags in specfile.py. This change doesn't effect pisi but scripts using specfile can make a little difference now according to common choice of tag ordering in repository.
Barış Metin
2006-10-24 18:20:20 +00:00
1928392b4e
* These errors would become much more seen when people start to play with the package-manager's cancel button. I have seen it a lot while playing with it. And when this happens there is no way to explain this to users.
Faik Uygur
2006-10-24 13:53:58 +00:00
53bc6a6c9b
update from willem
Görkem Çetin
2006-10-23 15:38:05 +00:00
60ad823c90
import glob
İsmail Dönmez
2006-10-23 08:05:33 +00:00
668474d7d3
cleanup, this will work on all arches not just i686
İsmail Dönmez
2006-10-23 08:04:12 +00:00
bb489e6686
* If addrepo operation fails, remove the repository that is left behind. (fixes 3743)
Faik Uygur
2006-10-22 20:30:55 +00:00
365fa59c64
* postpone removing conflicts for upgrade operation until fetching of all the packages finishes and before the actual installation begins. This keeps the system stable while fetching the packages. xorg -> xorg-server case.
Faik Uygur
2006-10-22 19:05:26 +00:00
6b8c854ffe
remove useless stuff
Onur Küçük
2006-10-22 12:30:08 +00:00
a79e1ec65d
copyright
Onur Küçük
2006-10-22 12:26:22 +00:00
7ea3039585
Traceback (most recent call last): File "/usr/kde/3.5/bin/package-manager", line 395, in showBasket basketDialog = BasketDialog.BasketDialog(self, self.basket.packages, self.state) File "/usr/kde/3.5/share/apps/package-manager/BasketDialog.py", line 99, in __init__ self.createExtraPackagesList() File "/usr/kde/3.5/share/apps/package-manager/BasketDialog.py", line 126, in createExtraPackagesList base = pisi.api.generate_base_upgrade(pkgs) File "/usr/lib/pardus/pisi/api.py", line 218, in generate_base_upgrade base = upgrade_base(A) File "/usr/lib/pardus/pisi/operations.py", line 278, in upgrade_base extra_installs = filter(lambda x: not ctx.installdb.is_installed(x), systembase - A) TypeError: unsupported operand type(s) for -: 'set' and 'list'
S.Çağlar Onur
2006-10-22 00:27:11 +00:00
d5552cd77d
bu checkleri ismail.py yapıyor zaten, bu arada epey paslanmış bu betik, şimdilik çalışmaya devam etsin diye çıkardım hata veren kodu.
Gürer Özen
2006-10-21 11:21:08 +00:00
a11e3fb8b2
* The other day I went to Malta. With so much confusion and unhospitality, I broke YALI.
Faik Uygur
2006-10-19 18:57:12 +00:00
3e37e3ded8
Yanlış dizin
Bahadır Kandemir
2006-10-19 13:28:24 +00:00
4da8efe8fb
Gelecekte yapılmış paketler var...
Bahadır Kandemir
2006-10-19 12:53:02 +00:00
074b7b4ae1
* current maintainer thingy
Faik Uygur
2006-10-19 08:00:15 +00:00
3bdc74f8dd
pisi 1.1_rc4 is coming.
Faik Uygur
2006-10-19 07:42:55 +00:00
ec741ffd2b
SpecFile should never ever ruin the generated PSPEC file. If build process needs this, we should do it there...
Barış Metin
2006-10-16 11:53:46 +00:00
2070fba9cd
* add methods that generate install, remove, upgrade order lists to api.
Faik Uygur
2006-10-15 19:56:49 +00:00
f39c1e4dcc
* handle more than 2 dashes gracefully * split returns strings. Make release and build also VersionItems.
Faik Uygur
2006-10-14 21:27:57 +00:00
84f4e1d247
Onur found out this:
Faik Uygur
2006-10-14 20:06:22 +00:00
235a1c2c87
kozmetik
Onur Küçük
2006-10-14 16:49:58 +00:00
600791a141
bana haber vermeden API ile oynamayın lütfen
S.Çağlar Onur
2006-10-13 15:19:12 +00:00
40800630f3
find the correct destination directory for dodir function
Ali Erdinç Köroğlu
2006-10-13 15:16:01 +00:00
7dd358cc9c
* pisi it component now will try to install all the packages of the component and the underlying components' packages.
Faik Uygur
2006-10-13 07:33:29 +00:00
ad3ab31bc7
* there were two ways to get packages of the component.
Faik Uygur
2006-10-13 06:53:37 +00:00
c9033067d5
COMAR betiklerinide derlemeye başlamadan kontrol et, hata varsa derlemeye çalışma
S.Çağlar Onur
2006-10-10 06:54:13 +00:00
1156b72df9
revert last, unsermake will installed into /opt as Onur suggested
S.Çağlar Onur
2006-10-09 16:40:01 +00:00
df30cf708f
Check if any package wants to use unsermake as a build system or whole kde packages try to use this :(
S.Çağlar Onur
2006-10-09 16:36:25 +00:00
60cd28d680
* tell what is being removed
Faik Uygur
2006-10-08 21:20:11 +00:00