Re: Apologies for breaking the freeze and a suggestion
On Wed, Jul 11, 2012 at 5:06 AM, Allen Winter wrote: > On Tuesday, July 10, 2012 06:32:04 PM Aurélien Gâteau wrote: >> Hi, >> >> This morning I worked on two bug fixes for Gwenview which I pushed to the >> KDE/4.9 branch. Only later in the afternoon did I check email and realized I >> should not have pushed those changes as we were freezing for RC2. I want to >> apologize for that. >> >> Since a "git push" is a bit too easy to do, I was wondering whether it would >> be possible to have a git hook which would disallow pushing in freeze mode. >> The hook could be designed to only allow commits if they have a >> "APPROVED_BY_RELEASE_TEAM" keyword in the commit message. >> >> This would prevent clueless people like me from pushing when they should >> not. Do you think this is doable? >> > > I don't know if it's doable. But I like the idea. > > Our awesome sysadmins might have some thoughts. Whilst this would be possible, it would require maintaining a list of all KDE SC repositories which are affected by this freeze, and making changes around each freeze period to switch it on and off again. > ___ > release-team mailing list > release-team@kde.org > https://mail.kde.org/mailman/listinfo/release-team Regards, Ben ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: KSecrets State?
On 07/10/2012 10:33 PM, Albert Astals Cid wrote: El Dimarts, 10 de juliol de 2012, a les 10:01:53, Rex Dieter va escriure: On 07/10/2012 09:56 AM, Jeremy Whiting wrote: It seems it has been moved to playground from kdelibs, but the application in kdeutils is still sitting there. Shouldn't that be moved to playground also until it works? yes, +1 If noone disagrees, I'll ask sysadmin on thursday to move the app to playground too. No problem for this and sorry for not having time for this these last months. I plan to resume it by this autumn eventually. Cheers, -- Valentin Rusu (IRC valir, KDE vrusu) KSecretsService (former KSecretService, KWallet replacement) ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: KSecrets State?
El Dimarts, 10 de juliol de 2012, a les 10:01:53, Rex Dieter va escriure: > On 07/10/2012 09:56 AM, Jeremy Whiting wrote: > > It seems it has been moved to playground from kdelibs, but the > > application in kdeutils is still sitting there. Shouldn't that be > > moved to playground also until it works? > > yes, +1 If noone disagrees, I'll ask sysadmin on thursday to move the app to playground too. Cheers, Albert > > -- rex > ___ > release-team mailing list > release-team@kde.org > https://mail.kde.org/mailman/listinfo/release-team ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: Review Request: refactoring of import wizard to be able to show errors
--- This is an automatically generated e-mail. To reply, visit: http://svn.reviewboard.kde.org/r/7003/#review10874 --- Ship it! Code looks sane and you got no objection from i18n teams, feel free to commit and if you're fast enough it'll be part of RC2 - Albert Astals Cid On July 2, 2012, 4:55 p.m., Ralf Habacker wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > http://svn.reviewboard.kde.org/r/7003/ > --- > > (Updated July 2, 2012, 4:55 p.m.) > > > Review request for Release Team. > > > Description > --- > > The appended git patch (uploaded as file) refactors the code import wizard to > be able to show import errors. This patch contains i18n string changes. > > > This addresses bug https://bugs.kde.org/show_bug.cgi?id=302671. > > https://bugs.kde.org/show_bug.cgi?id=https://bugs.kde.org/show_bug.cgi?id=302671 > > > Diffs > - > > > Diff: http://svn.reviewboard.kde.org/r/7003/diff/ > > > Testing > --- > > works well in trunk > > > Thanks, > > Ralf Habacker > > ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: Review Request: Include Review Requests in Changelog
--- This is an automatically generated e-mail. To reply, visit: http://svn.reviewboard.kde.org/r/6936/#review10873 --- Problem is that it only works with git reviewboard, not svn reviewboard, but probably we can cope with that? - Albert Astals Cid On April 19, 2012, 8:54 a.m., Martin Gräßlin wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > http://svn.reviewboard.kde.org/r/6936/ > --- > > (Updated April 19, 2012, 8:54 a.m.) > > > Review request for Release Team. > > > Description > --- > > This extends the changelog XML to contain "review" in bugfix entries and co. > XSL is adjusted to generate links to the review requests. > > > Diffs > - > > trunk/www/sites/www/announcements/changelogs/changelog.dtd 1287767 > trunk/www/sites/www/announcements/changelogs/changelog.xsl 1287767 > > Diff: http://svn.reviewboard.kde.org/r/6936/diff/ > > > Testing > --- > > example changelog: > > review="12345 123456">Ensure that KConfig update script does not overwrite > custom settings. > bugno="80897" review="45678">Lower windows does not change focus for Focus > (Strictly) Under Mouse Focus Policies. > review="12345">Fix Window Shading for XRender compositing. > review="22345">Desktop windows which do not annouce the virtual desktop they > belong to are set on all desktops. > review="12345 21463 2308">Zoom effect zooms towards the mouse cursor if mouse > tracking is disabled. > review="12346">XRender support for magnifier effect. > review="38269 23986">Implement xrender support for snaphelper > effect. > class="normal" bugno="243423" review="12121">Fix restore size for shaded > windows. > class="normal" bugno="261799" review="13213 23414">Select minimum stacking > order layer per screen. > class="normal" bugno="295449" review="12311">Save restore size before > removing the decoration for borderless maximized windows. > class="normal" review="121312 12313">Do not animate the highlight the first > time the tabbox is shown. > > > Generated HTML: > kwin > > Features: > > Save restore size before removing the decoration > for borderless maximized windows. Fixes bug href="http://bugs.kde.org/show_bug.cgi?id=295449";>295449. Discussed with > Review Request https://reviewboard.kde.org/r/12311";>12311. See > Git commit href="http://commits.kde.org/kde-workspace/677915da069b4d563b40250c4b69943af53a70bd";>677915d. > > Do not animate the highlight the first time the > tabbox is shown. Discussed with Review Requests href="https://reviewboard.kde.org/r/121312";>121312 and href="https://reviewboard.kde.org/r/12313";>12313. See Git commit href="http://commits.kde.org/kde-workspace/56345a9f158790e5f7102618c55c600cc688c855";>56345a9. > > > Improvements: > > XRender support for magnifier effect. > Discussed with Review Request href="https://reviewboard.kde.org/r/12346";>12346. See Git commit href="http://commits.kde.org/kde-workspace/c63944ee67906b49b75856d394e580def609c64a";>c63944e. > > Implement xrender support for snaphelper > effect. Discussed with Review Requests href="https://reviewboard.kde.org/r/38269";>38269 and href="https://reviewboard.kde.org/r/23986";>23986. See Git commit href="http://commits.kde.org/kde-workspace/4abba97f727faa6b801e7ceb73819e7f20cc59b8";>4abba97. > > > Optimizations: > Fix restore size for shaded windows. Fixes bug > http://bugs.kde.org/show_bug.cgi?id=243423";>243423. Discussed > with Review Request https://reviewboard.kde.org/r/12121";>12121. > See Git commit href="http://commits.kde.org/kde-workspace/5b11bf56780928cd531143ad6675bbe775548f92";>5b11bf5. > > Select minimum stacking order layer per screen. > Fixes bug http://bugs.kde.org/show_bug.cgi?id=261799";>261799. > Discussed with Review Requests href="https://reviewboard.kde.org/r/13213";>13213 and href="https://reviewboard.kde.org/r/23414";>23414. See Git commit href="http://commits.kde.org/kde-workspace/d854fc98630bcdbc0d450ae7bc642240f5f535ad";>d854fc9. > > > Bugfixes: > > Ensure that KConfig update script does not > overwrite custom settings. Discussed with Review Requests href="https://reviewboard.kde.org/r/12345";>12345 and href="https://reviewboard.kde.org/r/123456";>123456. See Git commit href="http://commits.kde.org/kde-workspace/05e1ee03e8b2dcd8e7a7563ca36f4fc01299e570";>05e1ee0. > > Lower windows does not change focus for > Focus (Strictly) Under Mouse Focus Policies. Fixes bug href="http://bugs.kde.org/show_bug.cgi?id=80897";>80897. Discus
Re: Review Request: last minutes major fixes for RC1
--- This is an automatically generated e-mail. To reply, visit: http://svn.reviewboard.kde.org/r/7007/#review10872 --- Ship it! Looks sane and you're using it already, don't see why we should not make it part of RC2 (you have like 2 hours to commit) - Albert Astals Cid On July 10, 2012, 7:15 p.m., Ralf Habacker wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > http://svn.reviewboard.kde.org/r/7007/ > --- > > (Updated July 10, 2012, 7:15 p.m.) > > > Review request for Release Team. > > > Description > --- > > Here are a few patches for bugs fixed in the last days. It would be good to > have them in the RC1 release. There are no i18n changes. > > > Diffs > - > > > Diff: http://svn.reviewboard.kde.org/r/7007/diff/ > > > Testing > --- > > works well in trunk > > > Thanks, > > Ralf Habacker > > ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: Apologies for breaking the freeze and a suggestion
El Dimarts, 10 de juliol de 2012, a les 13:06:17, Allen Winter va escriure: > On Tuesday, July 10, 2012 06:32:04 PM Aurélien Gâteau wrote: > > Hi, > > > > This morning I worked on two bug fixes for Gwenview which I pushed to the > > KDE/4.9 branch. Only later in the afternoon did I check email and realized > > I should not have pushed those changes as we were freezing for RC2. I > > want to apologize for that. > > > > Since a "git push" is a bit too easy to do, I was wondering whether it > > would be possible to have a git hook which would disallow pushing in > > freeze mode. The hook could be designed to only allow commits if they > > have a > > "APPROVED_BY_RELEASE_TEAM" keyword in the commit message. > > > > This would prevent clueless people like me from pushing when they should > > not. Do you think this is doable? > > I don't know if it's doable. But I like the idea. > > Our awesome sysadmins might have some thoughts. Are you asking them? (not sure they read our lists) Cheers, Albert > ___ > release-team mailing list > release-team@kde.org > https://mail.kde.org/mailman/listinfo/release-team ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Review Request: last minutes major fixes for RC1
--- This is an automatically generated e-mail. To reply, visit: http://svn.reviewboard.kde.org/r/7007/ --- Review request for Release Team. Description --- Here are a few patches for bugs fixed in the last days. It would be good to have them in the RC1 release. There are no i18n changes. Diffs - Diff: http://svn.reviewboard.kde.org/r/7007/diff/ Testing --- works well in trunk Thanks, Ralf Habacker ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: Review Request: Fix crash in nepomuk services after a timeout
--- This is an automatically generated e-mail. To reply, visit: http://git.reviewboard.kde.org/r/105507/#review15632 --- Ship it! Could you please add a FIXME in the comment? Just so that it becomes more obvious. - Vishesh Handa On July 10, 2012, 3:57 p.m., David Faure wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > http://git.reviewboard.kde.org/r/105507/ > --- > > (Updated July 10, 2012, 3:57 p.m.) > > > Review request for Nepomuk and Release Team. > > > Description > --- > > Fix crash due to other threads using the model at the time it gets deleted. > No proper solution found though, with the current API. > > > Diffs > - > > libnepomukcore/resource/nepomukmainmodel.cpp ba2626f > > Diff: http://git.reviewboard.kde.org/r/105507/diff/ > > > Testing > --- > > Running nepomukqueryservice in valgrind gave this: > > "/d/kde/inst/kde4.9/bin/nepomukservicestub(19732)" Soprano: "Timeout (5)": > "Command timed out." > ==19732== Thread 3: > ==19732== Invalid read of size 8 > ==19732==at 0xC5AB6DA: > QPointer::operator->() const > (qpointer.h:74) > ==19732==by 0xC5AA70A: Soprano::Client::ClientModel::executeQuery(QString > const&, Soprano::Query::QueryLanguage, QString const&) const > (clientmodel.cpp:106) > ==19732==by 0x68E9319: Nepomuk2::MainModel::executeQuery(QString const&, > Soprano::Query::QueryLanguage, QString const&) const > (nepomukmainmodel.cpp:182) > ==19732==by 0x12BDDEED: Nepomuk2::Query::SearchRunnable::run() > (searchrunnable.cpp:89) > ==19732==by 0x4ECA2FC: QThreadPoolThread::run() (qthreadpool.cpp:107) > ==19732==by 0x4EDAF27: QThreadPrivate::start(void*) (qthread_unix.cpp:307) > ==19732==by 0x53CCF04: start_thread (in /lib64/libpthread-2.14.1.so) > ==19732==by 0x86CD10C: clone (in /lib64/libc-2.14.1.so) > ==19732== Address 0x117af938 is 72 bytes inside a block of size 80 free'd > ==19732==at 0x4C28706: operator delete(void*) (vg_replace_malloc.c:457) > ==19732==by 0xC5AA2ED: Soprano::Client::ClientModel::~ClientModel() > (clientmodel.cpp:58) > ==19732==by 0x68E8BAB: (anonymous > namespace)::GlobalModelContainer::init(bool) (nepomukmainmodel.cpp:82) > ==19732==by 0x68E8E58: (anonymous > namespace)::GlobalModelContainer::model() (nepomukmainmodel.cpp:100) > ==19732==by 0x68E932D: Nepomuk2::MainModel::executeQuery(QString const&, > Soprano::Query::QueryLanguage, QString const&) const > (nepomukmainmodel.cpp:183) > ==19732==by 0x12BDDEED: Nepomuk2::Query::SearchRunnable::run() > (searchrunnable.cpp:89) > ==19732==by 0x4ECA2FC: QThreadPoolThread::run() (qthreadpool.cpp:107) > ==19732==by 0x4EDAF27: QThreadPrivate::start(void*) (qthread_unix.cpp:307) > ==19732==by 0x53CCF04: start_thread (in /lib64/libpthread-2.14.1.so) > ==19732==by 0x86CD10C: clone (in /lib64/libc-2.14.1.so) > ==19732== > > No crash anymore after this "fix"... > > > Thanks, > > David Faure > > ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: Review Request: Disable autocompletion-via-nepomuk
--- This is an automatically generated e-mail. To reply, visit: http://git.reviewboard.kde.org/r/105508/#review15630 --- Ship it! Until we find a method to fix it. It's a good idea to put this patch in 4.9. - Laurent Montel On July 10, 2012, 4:06 p.m., David Faure wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > http://git.reviewboard.kde.org/r/105508/ > --- > > (Updated July 10, 2012, 4:06 p.m.) > > > Review request for Akonadi, Nepomuk, Release Team, Volker Krause, Till Adam, > and Laurent Montel. > > > Description > --- > > Disable autocompletion-via-nepomuk until it is fast and does not block > further nepomuk queries (such as the ones run when clicking Send in kmail > composer). > > Without this patch: up to 30 minutes for the composer window to disappear > With this patch: immediate. > YYMV depending on the size of your nepomuk/virtuoso DB (here it's 3GB, with > 400k contacts, autogenerated by the akonadi-nepomuk-feeder). > > > Diffs > - > > libkdepim/addresseelineedit.cpp 5fab510 > > Diff: http://git.reviewboard.kde.org/r/105508/diff/ > > > Testing > --- > > > Thanks, > > David Faure > > ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: 4.9.0 RC 2 Tagging Freeze Tomorrow
On Monday 09 July 2012 22:30:43 Albert Astals Cid wrote: > El Dilluns, 9 de juliol de 2012, a les 10:57:22, David Faure va escriure: > > On Sunday 08 July 2012 22:22:21 you wrote: > > > Tomorrow 9 July 2012 at 23:59 UTC, the following freeze kicks in > > > > > > * KDE 4.9 Tagging Freeze for Release Candidate 2 > > > During the time in between of the RC2 tagging and the RC2 release. Only > > > compilation fixes for all platforms are allowed to be committed. > > > Everything else (even showstopper fixes) have to be run through > > > reviewboard, with the release-team and the affected maintainers as > > > reviewer. > > > > I have two patches waiting for approval on the nepomuk list, plus the main > > issue (sending email blocks, waiting for nepomuk) still isn't fixed. I > > strongly recommend that we don't release until this is fixed. > > Agreed, let's do this: > * Still freeze tonight > * Wait until nepomul+main issue be fixed > * After that decide further course of action Nepomuk issue is "fixed" for now, cf reviewboard emails to this list. > Anyhow, why didn't we know about this bad thing earlier? I don't know. Because most kde developers who experience trouble sending emails are told to "just disable nepomuk"? (that's what I was told initially, by several people; I had to insist that I wanted to make KDE better for everyone, not just fix my own issues). Or, worse, because most kde developers are not actually using KDE/4.9 or master on an every day basis? (I'm guilty for switching too late, clearly). Or, more likely, that most people who do just assume "it's someone else's problem" and just kill nepomuk or virtuoso when trouble happens (people admitted to doing that). In the end I spend akademy debugging the actual issues that end users will face with 4.9. (Granted, it seems it completely depends on the amount of emails in akonadi, and allegedly old indexers created more crap than they do now, but a quick poll on irc showed other people with the same issue and the same amount of contacts in virtuoso so it's not just me). To be clear, I'm not complaining that I have to debug stuff, it's my job, but last-minute fire-fighting in various layers of KDE isn't a well-organized process for ensuring quality. All in all, I hope the currently-being-formed QA team will be able to pick up such issues and force developers to fix them before the release... -- David Faure, fa...@kde.org, http://www.davidfaure.fr Sponsored by Nokia to work on KDE, incl. KDE Frameworks 5 ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: KDE 4.8.5 planning
On Tuesday, July 10, 2012 06:52:44 PM David Faure wrote: > On Sunday 08 July 2012 13:17:41 Michael Jansen wrote: > > Btw. is there already any time of idea when frameworks will be ready? > > There are more people asking this question, than people actually working on > making it happen. > > Too blunt? :) No ... and good for me. Because that means i have more time. Mike -- Michael Jansen http://michael-jansen.biz___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: Apologies for breaking the freeze and a suggestion
On Tuesday, July 10, 2012 06:32:04 PM Aurélien Gâteau wrote: > Hi, > > This morning I worked on two bug fixes for Gwenview which I pushed to the > KDE/4.9 branch. Only later in the afternoon did I check email and realized I > should not have pushed those changes as we were freezing for RC2. I want to > apologize for that. > > Since a "git push" is a bit too easy to do, I was wondering whether it would > be possible to have a git hook which would disallow pushing in freeze mode. > The hook could be designed to only allow commits if they have a > "APPROVED_BY_RELEASE_TEAM" keyword in the commit message. > > This would prevent clueless people like me from pushing when they should > not. Do you think this is doable? > I don't know if it's doable. But I like the idea. Our awesome sysadmins might have some thoughts. ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: KDE 4.8.5 planning
On Sunday 08 July 2012 13:17:41 Michael Jansen wrote: > Btw. is there already any time of idea when frameworks will be ready? There are more people asking this question, than people actually working on making it happen. Too blunt? :) -- David Faure, fa...@kde.org, http://www.davidfaure.fr Sponsored by Nokia to work on KDE, incl. KDE Frameworks 5 ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Apologies for breaking the freeze and a suggestion
Hi, This morning I worked on two bug fixes for Gwenview which I pushed to the KDE/4.9 branch. Only later in the afternoon did I check email and realized I should not have pushed those changes as we were freezing for RC2. I want to apologize for that. Since a "git push" is a bit too easy to do, I was wondering whether it would be possible to have a git hook which would disallow pushing in freeze mode. The hook could be designed to only allow commits if they have a "APPROVED_BY_RELEASE_TEAM" keyword in the commit message. This would prevent clueless people like me from pushing when they should not. Do you think this is doable? Aurélien ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: Review Request: Disable autocompletion-via-nepomuk
--- This is an automatically generated e-mail. To reply, visit: http://git.reviewboard.kde.org/r/105508/#review15629 --- Ship it! Ship It! - Till Adam On July 10, 2012, 4:06 p.m., David Faure wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > http://git.reviewboard.kde.org/r/105508/ > --- > > (Updated July 10, 2012, 4:06 p.m.) > > > Review request for Akonadi, Nepomuk, Release Team, Volker Krause, Till Adam, > and Laurent Montel. > > > Description > --- > > Disable autocompletion-via-nepomuk until it is fast and does not block > further nepomuk queries (such as the ones run when clicking Send in kmail > composer). > > Without this patch: up to 30 minutes for the composer window to disappear > With this patch: immediate. > YYMV depending on the size of your nepomuk/virtuoso DB (here it's 3GB, with > 400k contacts, autogenerated by the akonadi-nepomuk-feeder). > > > Diffs > - > > libkdepim/addresseelineedit.cpp 5fab510 > > Diff: http://git.reviewboard.kde.org/r/105508/diff/ > > > Testing > --- > > > Thanks, > > David Faure > > ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: Review Request: Disable autocompletion-via-nepomuk
--- This is an automatically generated e-mail. To reply, visit: http://git.reviewboard.kde.org/r/105508/#review15628 --- Ship it! looks good to me. - Volker Krause On July 10, 2012, 4:06 p.m., David Faure wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > http://git.reviewboard.kde.org/r/105508/ > --- > > (Updated July 10, 2012, 4:06 p.m.) > > > Review request for Akonadi, Nepomuk, Release Team, Volker Krause, Till Adam, > and Laurent Montel. > > > Description > --- > > Disable autocompletion-via-nepomuk until it is fast and does not block > further nepomuk queries (such as the ones run when clicking Send in kmail > composer). > > Without this patch: up to 30 minutes for the composer window to disappear > With this patch: immediate. > YYMV depending on the size of your nepomuk/virtuoso DB (here it's 3GB, with > 400k contacts, autogenerated by the akonadi-nepomuk-feeder). > > > Diffs > - > > libkdepim/addresseelineedit.cpp 5fab510 > > Diff: http://git.reviewboard.kde.org/r/105508/diff/ > > > Testing > --- > > > Thanks, > > David Faure > > ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: Review Request: Disable autocompletion-via-nepomuk
--- This is an automatically generated e-mail. To reply, visit: http://git.reviewboard.kde.org/r/105508/ --- (Updated July 10, 2012, 4:06 p.m.) Review request for Akonadi, Nepomuk, Release Team, Volker Krause, Till Adam, and Laurent Montel. Changes --- Sorry, didn't intend to include the query changes. Description --- Disable autocompletion-via-nepomuk until it is fast and does not block further nepomuk queries (such as the ones run when clicking Send in kmail composer). Without this patch: up to 30 minutes for the composer window to disappear With this patch: immediate. YYMV depending on the size of your nepomuk/virtuoso DB (here it's 3GB, with 400k contacts, autogenerated by the akonadi-nepomuk-feeder). Diffs (updated) - libkdepim/addresseelineedit.cpp 5fab510 Diff: http://git.reviewboard.kde.org/r/105508/diff/ Testing --- Thanks, David Faure ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: Review Request: Fix crash in nepomuk services after a timeout
--- This is an automatically generated e-mail. To reply, visit: http://git.reviewboard.kde.org/r/105507/ --- (Updated July 10, 2012, 3:57 p.m.) Review request for Nepomuk and Release Team. Changes --- Sorry, wrong patch/description combination. Fixed. Summary (updated) - Fix crash in nepomuk services after a timeout Description (updated) --- Fix crash due to other threads using the model at the time it gets deleted. No proper solution found though, with the current API. Diffs - libnepomukcore/resource/nepomukmainmodel.cpp ba2626f Diff: http://git.reviewboard.kde.org/r/105507/diff/ Testing (updated) --- Running nepomukqueryservice in valgrind gave this: "/d/kde/inst/kde4.9/bin/nepomukservicestub(19732)" Soprano: "Timeout (5)": "Command timed out." ==19732== Thread 3: ==19732== Invalid read of size 8 ==19732==at 0xC5AB6DA: QPointer::operator->() const (qpointer.h:74) ==19732==by 0xC5AA70A: Soprano::Client::ClientModel::executeQuery(QString const&, Soprano::Query::QueryLanguage, QString const&) const (clientmodel.cpp:106) ==19732==by 0x68E9319: Nepomuk2::MainModel::executeQuery(QString const&, Soprano::Query::QueryLanguage, QString const&) const (nepomukmainmodel.cpp:182) ==19732==by 0x12BDDEED: Nepomuk2::Query::SearchRunnable::run() (searchrunnable.cpp:89) ==19732==by 0x4ECA2FC: QThreadPoolThread::run() (qthreadpool.cpp:107) ==19732==by 0x4EDAF27: QThreadPrivate::start(void*) (qthread_unix.cpp:307) ==19732==by 0x53CCF04: start_thread (in /lib64/libpthread-2.14.1.so) ==19732==by 0x86CD10C: clone (in /lib64/libc-2.14.1.so) ==19732== Address 0x117af938 is 72 bytes inside a block of size 80 free'd ==19732==at 0x4C28706: operator delete(void*) (vg_replace_malloc.c:457) ==19732==by 0xC5AA2ED: Soprano::Client::ClientModel::~ClientModel() (clientmodel.cpp:58) ==19732==by 0x68E8BAB: (anonymous namespace)::GlobalModelContainer::init(bool) (nepomukmainmodel.cpp:82) ==19732==by 0x68E8E58: (anonymous namespace)::GlobalModelContainer::model() (nepomukmainmodel.cpp:100) ==19732==by 0x68E932D: Nepomuk2::MainModel::executeQuery(QString const&, Soprano::Query::QueryLanguage, QString const&) const (nepomukmainmodel.cpp:183) ==19732==by 0x12BDDEED: Nepomuk2::Query::SearchRunnable::run() (searchrunnable.cpp:89) ==19732==by 0x4ECA2FC: QThreadPoolThread::run() (qthreadpool.cpp:107) ==19732==by 0x4EDAF27: QThreadPrivate::start(void*) (qthread_unix.cpp:307) ==19732==by 0x53CCF04: start_thread (in /lib64/libpthread-2.14.1.so) ==19732==by 0x86CD10C: clone (in /lib64/libc-2.14.1.so) ==19732== No crash anymore after this "fix"... Thanks, David Faure ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Review Request: Disable autocompletion-via-nepomuk
--- This is an automatically generated e-mail. To reply, visit: http://git.reviewboard.kde.org/r/105508/ --- Review request for Akonadi, Nepomuk, Release Team, Volker Krause, Till Adam, and Laurent Montel. Description --- Disable autocompletion-via-nepomuk until it is fast and does not block further nepomuk queries (such as the ones run when clicking Send in kmail composer). Without this patch: up to 30 minutes for the composer window to disappear With this patch: immediate. YYMV depending on the size of your nepomuk/virtuoso DB (here it's 3GB, with 400k contacts, autogenerated by the akonadi-nepomuk-feeder). Diffs - libkdepim/addresseelineedit.cpp 5fab510 Diff: http://git.reviewboard.kde.org/r/105508/diff/ Testing --- Thanks, David Faure ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: KSecrets State?
On 07/10/2012 09:56 AM, Jeremy Whiting wrote: It seems it has been moved to playground from kdelibs, but the application in kdeutils is still sitting there. Shouldn't that be moved to playground also until it works? yes, +1 -- rex ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: KSecrets State?
It seems it has been moved to playground from kdelibs, but the application in kdeutils is still sitting there. Shouldn't that be moved to playground also until it works? Every time I build with kdesrc-build here I get ksecretsserviced which doesn't work, causing havoc with gnome-keyring which does. thx, Jeremy On Mon, May 21, 2012 at 12:26 AM, Valentin Rusu wrote: > Hello Anne-Marie, Rex > > > On 05/18/2012 02:37 PM, Rex Dieter wrote: >> >> On 05/17/2012 07:01 AM, Anne-Marie Mahfouf wrote: >> >>> A few weeks ago, KSecrets working state was challenged and considered as >>> still alpha but as we could not agree on removing it from 4.8 releases, >>> it stayed there. >>> Now is the question on whether it improved for 4.9 and should be kept in >>> the 4.9 release series or if it should be put in a development area >>> again. >>> Valentin, what's your opinion? Did people try it through the 4.8 >>> releases and reported problems that you improved? What is your vision >>> for the future? >> >> > > Well, unfortunately too few people attempted to test the service so I > haven't got the amount of feedback I was looking for. > > >> >> Without any positive feedback on either issue, it seems to me that >> ksecrets is doing more harm than good atm, and removing from 4.9 should be >> considered. > > > I think that'll be better that way. I'll have some more time to work on it > the fall of the summer, when the "strategic project" I'm working on for my > employer will go into production. > > Regards, > > > -- > Valentin Rusu (IRC valir, KDE vrusu) > KSecretsService (former KSecretService, KWallet replacement) > ___ > release-team mailing list > release-team@kde.org > https://mail.kde.org/mailman/listinfo/release-team ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
Re: Branch kdelibs KDE/4.10 now?
On Monday, July 09, 2012 08:59:27 PM David Faure wrote: > On Monday 09 July 2012 20:09:26 Albert Astals Cid wrote: > > El Dissabte, 7 de juliol de 2012, a les 19:38:47, Allen Winter va escriure: > > > Howdy, > > > > > > Why not make the KDE/4.10 branch for kdelibs now? > > > > > > That would make sense, no? > > > > > > Then we tell developers that master (the future KDE SC 4.10) builds > > > against > > > kdelibs KDE/4.10 > > > > This is not what we did during the 4.8 and 4.9 cycles. I understand it's > > confusing but probably people already got used to it. > > > > Anyway I'm not for nor against, I'm for what dfaure as maintainer says, > > afair the reason not to have more branches was to ease merging of stuff to > > frameworks, if we add another branch it'll probably be more work again. > > > > David? > > Can this wait until 4.8.5 is released? > > Otherwise we'll really have too many branches. Certainly. Not a big deal. ___ release-team mailing list release-team@kde.org https://mail.kde.org/mailman/listinfo/release-team
