| Summary: | x11-misc/shared-mime-info-1.2 updates database very slowly | ||
|---|---|---|---|
| Product: | Portage Development | Reporter: | Nikos Chantziaras <realnc> |
| Component: | Core | Assignee: | Freedesktop bugs <freedesktop-bugs> |
| Status: | RESOLVED FIXED | ||
| Severity: | enhancement | CC: | adaptee, david, doug.hunley, gmurray, josef64, marduk, pacho, phantom4, systemd |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | All | ||
| See Also: |
https://bugs.freedesktop.org/show_bug.cgi?id=70366 https://bugs.gentoo.org/show_bug.cgi?id=487626 https://bugs.gentoo.org/show_bug.cgi?id=462118 https://bugs.gentoo.org/show_bug.cgi?id=480616 https://bugs.gentoo.org/show_bug.cgi?id=819783 |
||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
|
Description
Nikos Chantziaras
2013-10-10 10:37:18 UTC
I am updating KDE now and this is making the process unbelievable slow. I was thinking "what the hell happened with my PC, that now it compiles KDE all day instead of a few hours" and then I noticed HDD bzzzzzzzzz and ages at stage Updating shared mime info database ... (In reply to David Kredba from comment #1) > I am updating KDE now and this is making the process unbelievable slow. I > was thinking "what the hell happened with my PC, that now it compiles KDE > all day instead of a few hours" and then I noticed HDD bzzzzzzzzz and ages > at stage Updating shared mime info database ... As a workaround, simply abort the emerge, downgrade to x11-misc/shared-mime-info-1.1 and then do a @world upgrade again. Added workaround to Portage: + 11 Oct 2013; Samuli Suominen <ssuominen@gentoo.org> + shared-mime-info-1.2.ebuild: + Disable use of fdatasync() due to severe time regression wrt #487504 by Nikos + Chantziaras And reported it to upstream: https://bugs.freedesktop.org/show_bug.cgi?id=70366 (In reply to Samuli Suominen from comment #3) > Added workaround to Portage: > > + 11 Oct 2013; Samuli Suominen <ssuominen@gentoo.org> > + shared-mime-info-1.2.ebuild: > + Disable use of fdatasync() due to severe time regression wrt #487504 by > Nikos > + Chantziaras Perhaps it would be useful to do a revision bump to ensure everyone gets the workaround? This is causing major issues for KDE users, since update-mime-database is called for every ebuild by the eclass. (In reply to Michael Palimaka (kensington) from comment #4) > (In reply to Samuli Suominen from comment #3) > > Added workaround to Portage: > > > > + 11 Oct 2013; Samuli Suominen <ssuominen@gentoo.org> > > + shared-mime-info-1.2.ebuild: > > + Disable use of fdatasync() due to severe time regression wrt #487504 by > > Nikos > > + Chantziaras > > Perhaps it would be useful to do a revision bump to ensure everyone gets the > workaround? > > This is causing major issues for KDE users, since update-mime-database is > called for every ebuild by the eclass. Revision bump is required here! I see no bump here? :) ...
> Revision bump is required here!
Yes, please please make a Revision bump.
A sleek kdebase-meta minor update needed here over two hours only for update-mime-database :(
Please make a revision bump. I had to google this issue first just to figure out it could have been easily avoided if you bothered to revbump the package. Thanks! +*shared-mime-info-1.2-r1 (06 Nov 2013) + + 06 Nov 2013; Pacho Ramos <pacho@gentoo.org> +shared-mime-info-1.2-r1.ebuild, + -shared-mime-info-1.2.ebuild: + Ensure people get fix for bug #487504 + CCing dev-portage@ Can we integrate fdo-mime.eclass functionality to the PM and make it run the database updates only once, at the end of emerge, just like eg. ldconfig is ran now As upstream suggests at https://bugs.freedesktop.org/show_bug.cgi?id=70366#c5 Because I consider disabling the fdatasync() a workaround. There is indeed no point in running it so many times. Portage could run /usr/bin/update-desktop-database only when files get installed to /usr/share/applications and /usr/bin/update-mime-database only when files get installed to /usr/share/mime Same applies for gtk icon cache, immodules cache, pango cache, and a few others. Also would be useful for systemd as would allow to , for example, run command for creating new files when tmpfiles.d are installed, running it only one time at the end of emerge (like other PMs like yum do) (In reply to Pacho Ramos from comment #12) > Also would be useful for systemd as would allow to , for example, run > command for creating new files when tmpfiles.d are installed, running it > only one time at the end of emerge (like other PMs like yum do) This probably should be reassigned to portage team (and tagged as enhancement as we don't have patches or similar :( ) This is fixed in Portage by disabling upstream's use of fdatasync() and this will be fixed in shared-mime-info-1.3 by the new -n parameter which I intend to add to fdo-mime.eclass after the version bump (after it's released) http://cgit.freedesktop.org/xdg/shared-mime-info/commit/?id=29a04be6c9cbaf0865c8b57428b7b7c37fbda4c3 http://cgit.freedesktop.org/xdg/shared-mime-info/commit/?id=4b3f9f774da8859d4f1f7e991b12832d6c09b63e So, fix summary away from shared-mime-info if you intend to reuse this bug to fix something else? We have no problem with shared-mime-info, whatsoever, not now, not in future. (In reply to Samuli Suominen from comment #10) > CCing dev-portage@ > > Can we integrate fdo-mime.eclass functionality to the PM and make it run the > database updates only once, at the end of emerge, just like eg. ldconfig is > ran now as in, this is no longer required... upstream fixed this in the application itself for shared-mime-info if you still want to do it, i suppose you still can, but nothing is broken meanwhile Is it OK to close this bug now? (In reply to Samuli Suominen from comment #14) > So, fix summary away from shared-mime-info if you intend to reuse this bug > to fix something else? We have no problem with shared-mime-info, whatsoever, > not now, not in future. We should open a new bug against PMS for this. (In reply to Nikos Chantziaras from comment #16) > Is it OK to close this bug now? Done. |
