| Summary: | app-portage/portage-utils: qlop: args should be processed as atoms and not substring | ||
|---|---|---|---|
| Product: | Portage Development | Reporter: | Jeroen Roovers (RETIRED) <jer> |
| Component: | Third-Party Tools | Assignee: | Portage Utils Team <portage-utils> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | dan, denilsonsa |
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | All | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: | portage-utils-0.1.9-qlop-pkgname-fix.patch | ||
|
Description
Jeroen Roovers (RETIRED)
2005-11-17 06:44:50 UTC
Created attachment 73086 [details, diff]
portage-utils-0.1.9-qlop-pkgname-fix.patch
all q applets suffer from this "feature" we should prob bite the bullet and standardize this stuff ... (In reply to comment #2) > all q applets suffer from this "feature" > > we should prob bite the bullet and standardize this stuff ... Perhaps. The way I did the patch, it would return info on all ambiguous names cat-one/pkgname vs cat-two/pkgname. I might rewrite it to 1) apply to all applets (probably in main.c) and 2) to allow for the inclusion of a category, but not today. :) I think there is some additional brokeness with qlop. The output from this command doesn't seem self-consistent to me: dan@tangier ~ $ sudo qlop -lt links Wed Jun 22 18:58:22 2005 >>> www-client/links-2.1_pre17 Tue Jul 5 09:24:53 2005 >>> www-client/links-2.1_pre17-r1 Mon Nov 21 15:49:38 2005 >>> www-client/links-2.1_pre18 Wed Dec 14 00:46:16 2005 >>> www-client/links-2.1_pre19 links: 98 seconds average for 7 merges onyx ~ # qlop -t portage-utils portage-utils: 21 seconds average for 5 merges onyx ~ # qlop -l portage-utils Thu Nov 10 07:28:07 2005 >>> app-portage/portage-utils-0.1.8 Mon Nov 14 13:58:42 2005 >>> app-portage/portage-utils-0.1.9 Thu Nov 17 12:44:41 2005 >>> app-portage/portage-utils-0.1.9 Fri Nov 18 17:19:38 2005 >>> app-portage/portage-utils-0.1.9 Thu Dec 1 14:22:53 2005 >>> app-portage/portage-utils-0.1.10 onyx ~ # qlop -lt portage-utils Thu Nov 10 07:28:07 2005 >>> app-portage/portage-utils-0.1.8 Mon Nov 14 13:58:42 2005 >>> app-portage/portage-utils-0.1.9 Thu Nov 17 12:44:41 2005 >>> app-portage/portage-utils-0.1.9 Fri Nov 18 17:19:38 2005 >>> app-portage/portage-utils-0.1.9 Thu Dec 1 14:22:53 2005 >>> app-portage/portage-utils-0.1.10 portage-utils: 21 seconds average for 5 merges How would you expect the output to look like when invoking two unrelated options? there is a bug i noticed where qlop gets confused when you run multiple emerges at the sametime ... so in the log file you get events like 'starting emerge; starting emerge; finish emerge; finish emerge' instead of 'starting emerge; finish emerge; starting emerge; finish emerge' I should have been more clear. You can see in the command output in comment #4 that it only lists links being as being merged 4 times yet it says it calculated its average time to merge based on 7 merges. Where did those extra three come from? What else is it counting? Simple way to reproduce and understand this bug: "qdepends -a alsa" shows a different output from "qdepends -a media-sound/ alsa" I can confirm the problem on comment #4 # genlop -t kdelibs * kde-base/kdelibs Fri Nov 25 02:58:21 2005 >>> kde-base/kdelibs-3.4.1-r1 merge time: 5 hours, 9 minutes and 40 seconds. Sat Nov 26 22:03:36 2005 >>> kde-base/kdelibs-3.4.1-r1 merge time: 1 hour, 41 minutes and 1 second. # qlop -lt kdelibs Fri Nov 25 02:58:21 2005 >>> kde-base/kdelibs-3.4.1-r1 Sat Nov 26 22:03:36 2005 >>> kde-base/kdelibs-3.4.1-r1 kdelibs: 6061 seconds average for 1 merges Yes, one of those emerges (could you guess which one? ;) ) was done in parallel with some other(s) emerge(s). (confirming the comment #6) Another case which seems to demonstrate the strange behaviour of qlop:
marge ~ # qlop -gl mysql
Tue May 4 18:19:23 2004 >>> dev-db/mysql-4.0.18-r1
Tue May 4 18:20:57 2004 >>> dev-perl/DBD-mysql-2.1027
Wed May 19 12:36:41 2004 >>> dev-db/mysql-4.0.18-r2
Mon Jul 12 21:03:31 2004 >>> dev-db/mysql-4.0.20
Mon Aug 30 19:04:29 2004 >>> dev-db/mysql-4.0.20-r1
Sun Oct 24 11:59:45 2004 >>> dev-db/mysql-4.0.21
Wed Nov 3 23:00:56 2004 >>> dev-db/mysql-4.0.22
Fri Jan 21 11:52:28 2005 >>> dev-db/mysql-4.0.22-r2
Fri Feb 11 16:08:33 2005 >>> dev-perl/DBD-mysql-2.1027
Thu Mar 17 11:08:11 2005 >>> dev-db/mysql-4.0.24
Wed Apr 27 11:53:07 2005 >>> dev-perl/DBD-mysql-2.9003
Tue Aug 9 19:50:58 2005 >>> dev-perl/DBD-mysql-2.9003
Thu Aug 18 12:10:30 2005 >>> dev-db/mysql-4.0.25-r2
Tue Dec 13 13:15:19 2005 >>> dev-db/mysql-4.1.14
Tue Dec 13 13:16:01 2005 >>> dev-perl/DBD-mysql-2.9007
mysql: Tue May 4 17:49:31 2004: 1792 seconds
mysql: Wed May 19 11:37:07 2004: 3574 seconds
mysql: Mon Jul 12 20:35:21 2004: 1690 seconds
mysql: Mon Aug 30 18:32:18 2004: 1931 seconds
mysql: Fri Jan 21 11:27:04 2005: 1524 seconds
mysql: Thu Mar 17 10:36:17 2005: 1914 seconds
mysql: Thu Aug 18 11:43:43 2005: 1607 seconds
mysql: Tue Dec 13 12:05:49 2005: 4170 seconds
mysql: 8 times
marge ~ #
1st part list all 10 emerge of mysql whereas it reports only 8 in the second part. Also genlop displays all of them:
marge ~ # genlop mysql
* dev-db/mysql
Tue May 4 18:19:23 2004 >>> dev-db/mysql-4.0.18-r1
Wed May 19 12:36:41 2004 >>> dev-db/mysql-4.0.18-r2
Mon Jul 12 21:03:31 2004 >>> dev-db/mysql-4.0.20
Mon Aug 30 19:04:29 2004 >>> dev-db/mysql-4.0.20-r1
Sun Oct 24 11:59:45 2004 >>> dev-db/mysql-4.0.21
Wed Nov 3 23:00:56 2004 >>> dev-db/mysql-4.0.22
Fri Jan 21 11:52:28 2005 >>> dev-db/mysql-4.0.22-r2
Thu Mar 17 11:08:11 2005 >>> dev-db/mysql-4.0.24
Thu Aug 18 12:10:30 2005 >>> dev-db/mysql-4.0.25-r2
Tue Dec 13 13:15:19 2005 >>> dev-db/mysql-4.1.14
marge ~ #
with recent versions, we have the ability to sanely process & compare atoms. so this should be feasible now. the advantage would you could do things like: qlop -l '>sys-devel/gcc-5' qlop -l 'python:2.7' the downside is you'd no longer have substring matching. but maybe that's for the best. |