FS#15144 - [makepkg] improve installed dependency removal
Attached to Project:
Pacman
Opened by Allan McRae (Allan) - Wednesday, 17 June 2009, 08:01 GMT
Last edited by Allan McRae (Allan) - Tuesday, 11 May 2010, 22:56 GMT
Opened by Allan McRae (Allan) - Wednesday, 17 June 2009, 08:01 GMT
Last edited by Allan McRae (Allan) - Tuesday, 11 May 2010, 22:56 GMT
|
Details
Summary and Info:
When using makepkg with "-sr" and a provider package is installed, makepkg can not remove it at the end of the transaction. |
This task depends upon
Closed by Allan McRae (Allan)
Tuesday, 11 May 2010, 22:56 GMT
Reason for closing: Implemented
Additional comments about closing: In git: http://projects.archlinux.org/pacman.git /commit/?id=ccbef232
Tuesday, 11 May 2010, 22:56 GMT
Reason for closing: Implemented
Additional comments about closing: In git: http://projects.archlinux.org/pacman.git /commit/?id=ccbef232
Comment by Nagy Gabor (combo) -
Thursday, 18 June 2009, 10:49 GMT
Comment by Allan McRae (Allan) -
Friday, 23 October 2009, 05:51 GMT
Comment by Dan McGee (toofishes) -
Tuesday, 01 December 2009, 03:15 GMT
Comment by Allan McRae (Allan) -
Tuesday, 01 December 2009, 03:22 GMT
Comment by Dan McGee (toofishes) -
Tuesday, 01 December 2009, 03:28 GMT
Comment by xduugu (xduugu) -
Wednesday, 20 January 2010, 14:09 GMT
See also:
http://www.archlinux.org/pipermail/pacman-dev/2008-April/006143.html
Patch:
http://mailman.archlinux.org/pipermail/pacman-dev/2009-October/009873.html
Does this make sense? Say I have a package that takes a long time
to build. I have it install some deps, and then I go and install a
completely unrelated package during the hour it is building. It
would then get uninstalled once the package is built unless I'm
reading this incorrectly.
Good point... it should be safe to move the detection of packages
installed post-dep handling to directly after the dependencies are
installed.
Yeah if we grab the list right after the pacman install completes,
that would at least make the window for a race condition a lot
shorter. I'd be much happier with that.
Latest patch:
http://mailman.archlinux.org/pipermail/pacman-dev/2009-December/010239.html