View previous topic :: View next topic |
Author |
Message |
Jeekay n00b
Joined: 27 Mar 2003 Posts: 18
|
Posted: Sun Sep 28, 2003 10:05 am Post subject: fixpackages moving the same files over and over |
|
|
I suspect others have asked this question but I couldn't find anything relevant in the search.. my apologies if this is a dupe.
I have `fixpackages' in my FEATURES, so whenever I run emerge sync it also runs fixpackages. Thats fine, etc.
However, I don't understand why it seems to keep moving the same files over and over again. I was under the impression that it moved files around if their package directory had changed, but in such a case I do not understand why two consecutive runs of emerge sync both result in fixpackages apparently moving around 200 files, some of which are from 2002 according to the name of the changes file! It would seem that once it has been moved, its been moved...
On that note, the sheer slowness with which fixpackages works leads me to think it is copying rather than moving, which could explain why it is so dog slow. Can anyone confirm this?
And on a final note, whats with the `destination already exists' errors? |
|
Back to top |
|
|
Genone Retired Dev
Joined: 14 Mar 2003 Posts: 9535 Location: beyond the rim
|
Posted: Sun Sep 28, 2003 10:45 am Post subject: |
|
|
See my fixpackages explanation on the gentoo-user list: http://article.gmane.org/gmane.linux.gentoo.user/46273
Some of the issues should be solved in recent portage versions (IIRC -r5 and later), for others are proposed solutions in bugzilla. |
|
Back to top |
|
|
Jeekay n00b
Joined: 27 Mar 2003 Posts: 18
|
Posted: Sun Sep 28, 2003 10:51 am Post subject: |
|
|
Thankyou for the update, but I still don't understand why consecutive runs of fixpackages update the same files over and over. |
|
Back to top |
|
|
Genone Retired Dev
Joined: 14 Mar 2003 Posts: 9535 Location: beyond the rim
|
Posted: Sun Sep 28, 2003 11:04 am Post subject: |
|
|
The first problem is that fixpackages can't determine if a binary package was already updated or not, so currently we use the brute force approach and fix them even if they are already corrected.
The bigger problem is that somehow the mtime comparison doesn't work right, I'm still investigating why this is the case. Have to discuss this some more with carpaski as we came to different conclusions (see bugzilla for details). |
|
Back to top |
|
|
|