D15420: Intialize m_lastPosition

jtamate created this revision. jtamate added reviewers: KTextEditor, cullmann. Herald added projects: Kate, Frameworks. Herald added subscribers: kde-frameworks-devel, kwrite-devel. jtamate requested review of this revision. REVISION SUMMARY Avoid uninitialized usage of m_lastPosition. TEST

D12016: [ktexteditor] much faster positionFromCursor

jtamate added a comment. In D12016#323833 , @volkov wrote: > I get the following warning from valgrind because of this change: Fix in D15420 REPOSITORY R39 KTextEditor REVISION DETAIL

D15422: [server] Fix remote access buffer handlig when output not bound

davidedmundson added inline comments. INLINE COMMENTS > remote_access_interface.cpp:217 > +// buffer was not requested by any client > +close(buf->fd()); > +delete buf; Would it be better to emit bufferReleased? Otherwise we're sometimes doing this deletion in the

D11055: Pack python bindings into right dirs

McPain abandoned this revision. Herald edited subscribers, added: kde-buildsystem, kde-frameworks-devel; removed: Frameworks, Build System. REPOSITORY R240 Extra CMake Modules REVISION DETAIL https://phabricator.kde.org/D11055 To: McPain, vkrause Cc: kde-frameworks-devel, kde-buildsystem,

D15422: [server] Fix remote access buffer handling when output not bound

romangg edited the summary of this revision. REPOSITORY R127 KWayland REVISION DETAIL https://phabricator.kde.org/D15422 To: romangg, #kwin, #frameworks Cc: davidedmundson, kde-frameworks-devel, michaelh, ngraham, bruns

D15426: Avoid QByteArray::remove in AccessManagerReply::readData

fvogt created this revision. fvogt added reviewers: Frameworks, elvisangelaccio. Herald added a project: Frameworks. fvogt requested review of this revision. REVISION SUMMARY It copies the remaining data to the beginning of the buffer, which can be really expensive. BUG: 375765 TEST PLAN

D15406: Manually resize KCMUtilDialog to sizeHint()

acrouthamel added a comment. In D15406#324063 , @ngraham wrote: > @acrouthamel, could you test this patch to make sure it doesn't regress anything for High DPI users? I can test it out tonight. REPOSITORY R295 KCMUtils BRANCH

D15417: [AppImage Thumbnailer] Avoid creating QTemporaryFile

TheAssassin accepted this revision. This revision is now accepted and ready to land. REPOSITORY R320 KIO Extras REVISION DETAIL https://phabricator.kde.org/D15417 To: broulik, #frameworks, TheAssassin, anthonyfieroni

D15421: [server] Add surface data proxy mechanism

romangg added a dependency: D15023: [server] Add selectionChanged signal. REPOSITORY R127 KWayland REVISION DETAIL https://phabricator.kde.org/D15421 To: romangg, #kwin Cc: kde-frameworks-devel, michaelh, ngraham, bruns

D15421: [server] Add surface data proxy mechanism

romangg set the repository for this revision to R127 KWayland. romangg added a project: Frameworks. Herald added a subscriber: kde-frameworks-devel. REPOSITORY R127 KWayland REVISION DETAIL https://phabricator.kde.org/D15421 To: romangg, #kwin Cc: kde-frameworks-devel, michaelh, ngraham,

D15023: [server] Add selectionChanged signal

romangg added a dependent revision: D15421: [server] Add surface data proxy mechanism. REPOSITORY R127 KWayland REVISION DETAIL https://phabricator.kde.org/D15023 To: romangg, #kwin Cc: kde-frameworks-devel, michaelh, ngraham, bruns

D15421: [server] Add surface data proxy mechanism

romangg added a comment. Thinking about it again it makes probably more sense to only allow one proxy surface per remote. Otherwise a drag remote can change mid-drag when a user presses a second button while the drag is going on. REPOSITORY R127 KWayland REVISION DETAIL

D15422: [server] Fix remote access buffer handlig when output not bound

romangg created this revision. romangg added reviewers: KWin, Frameworks. Herald added a project: Frameworks. Herald added a subscriber: kde-frameworks-devel. romangg requested review of this revision. REVISION SUMMARY If a client has not bound a certain output do not directly return but try

D15406: Manually resize KCMUtilDialog to sizeHint()

ngraham added a comment. @acrouthamel, could you test this patch to make sure it doesn't regress anything for High DPI users? REPOSITORY R295 KCMUtils BRANCH master REVISION DETAIL https://phabricator.kde.org/D15406 To: valeriymalov, #frameworks, ngraham Cc: acrouthamel,

D12016: [ktexteditor] much faster positionFromCursor

volkov added a comment. Herald added subscribers: kde-frameworks-devel, kwrite-devel. I get the following warning from valgrind because of this change: ==3621== Conditional jump or move depends on uninitialised value(s) ==3621==at 0x4FF3605:

D15422: [server] Fix remote access buffer handlig when output not bound

romangg added inline comments. INLINE COMMENTS > davidedmundson wrote in remote_access_interface.cpp:217 > Would it be better to emit bufferReleased? > > Otherwise we're sometimes doing this deletion in the framework, sometimes in > kwin True, is nicer. REPOSITORY R127 KWayland REVISION

D15422: [server] Fix remote access buffer handlig when output not bound

romangg updated this revision to Diff 41400. romangg marked an inline comment as done. romangg added a comment. - Emit bufferReleased signal instead of cleaning up directly REPOSITORY R127 KWayland CHANGES SINCE LAST UPDATE https://phabricator.kde.org/D15422?vs=41396=41400 BRANCH

D15422: [server] Fix remote access buffer handling when output not bound

romangg retitled this revision from "[server] Fix remote access buffer handlig when output not bound" to "[server] Fix remote access buffer handling when output not bound". REPOSITORY R127 KWayland REVISION DETAIL https://phabricator.kde.org/D15422 To: romangg, #kwin, #frameworks Cc:

D15422: [server] Fix remote access buffer handling when output not bound

jgrulich added a comment. Both versions solve issue I have and KWin no longer opens dozens of fds and abort when reaches the limit. REPOSITORY R127 KWayland REVISION DETAIL https://phabricator.kde.org/D15422 To: romangg, #kwin, #frameworks Cc: jgrulich, davidedmundson,

D15406: Manually resize KCMUtilDialog to sizeHint()

valeriymalov added a comment. neither `QWidget::adjustSize()` nor `QWidget::adjustedSize()` seem to use `devicePixelRatio` (https://code.qt.io/cgit/qt/qtbase.git/tree/src/widgets/kernel/qwidget.cpp#n8665), but I can't test that to vouch that it works on HiDPI as intended I can land it as

D15417: [AppImage Thumbnailer] Avoid creating QTemporaryFile

broulik created this revision. broulik added reviewers: Frameworks, TheAssassin, anthonyfieroni. broulik requested review of this revision. REVISION SUMMARY Instead use new API that returns a buffer to avoid writing to the file system TEST PLAN Built libappimage with

D15420: Intialize m_lastPosition

cullmann accepted this revision. cullmann added a comment. This revision is now accepted and ready to land. Reasonable fix. Thanks! REPOSITORY R39 KTextEditor REVISION DETAIL https://phabricator.kde.org/D15420 To: jtamate, #ktexteditor, cullmann Cc: kwrite-devel, kde-frameworks-devel,

kio slaves + kcrash + drkonqi

Hey I am a bit confused about crash handling in kio slaves. As in: it's not working the way it should be working and I don't know why it is the way it is. The debugging faq [1] suggest that this should behave like KCrash. by default all crashes get drkonqi'd unless you set KDE_DEBUG at which

D15023: [server] Add selectionChanged signal

romangg updated this revision to Diff 41386. romangg added a comment. Rebase on master. REPOSITORY R127 KWayland CHANGES SINCE LAST UPDATE https://phabricator.kde.org/D15023?vs=40302=41386 BRANCH 0selectionChangeXwl REVISION DETAIL https://phabricator.kde.org/D15023 AFFECTED FILES

D15420: Intialize m_lastPosition

volkov accepted this revision. REPOSITORY R39 KTextEditor REVISION DETAIL https://phabricator.kde.org/D15420 To: jtamate, #ktexteditor, cullmann, volkov Cc: kwrite-devel, kde-frameworks-devel, michaelh, kevinapavew, ngraham, bruns, demsking, cullmann, sars, dhaumann

D15422: [server] Fix remote access buffer handling when output not bound

romangg marked an inline comment as done. REPOSITORY R127 KWayland REVISION DETAIL https://phabricator.kde.org/D15422 To: romangg, #kwin, #frameworks Cc: davidedmundson, kde-frameworks-devel, michaelh, ngraham, bruns

D15422: [server] Fix remote access buffer handling when output not bound

davidedmundson accepted this revision. This revision is now accepted and ready to land. REPOSITORY R127 KWayland BRANCH remoteBufferClientsUnbindOutput REVISION DETAIL https://phabricator.kde.org/D15422 To: romangg, #kwin, #frameworks, davidedmundson Cc: jgrulich, davidedmundson,

D15426: Avoid QByteArray::remove in AccessManagerReply::readData

fvogt added a comment. Depending on how AccessManagerReply is used, it might be necessary to do `m_data.remove(0, m_offset); m_offset = 0;` if `m_offset` grows too large to not leak memory. Can a KIO expert answer this? REPOSITORY R241 KIO REVISION DETAIL

D15406: Manually resize KCMUtilDialog to sizeHint()

acrouthamel added a comment. I've tested this patch on my system. Although it does not resolve the scaling issue I experience, it does not cause any regressions I see. Ship it! :) REPOSITORY R295 KCMUtils BRANCH master REVISION DETAIL https://phabricator.kde.org/D15406 To:

D6513: Add support for Attica tags support

ngraham added a comment. In D6513#322167 , @leinir wrote: > In D6513#322078 , @ngraham wrote: > > > So is this enough for people to start tagging KDE4 content as such? Or is anything else still

Re: Quality of Frameworks announcements

On Sunday, 9 September 2018 16:07:43 CEST Yuri Chornoivan wrote: > Hi, > > I think that several latest Frameworks announcements are not of even medium > quality. It is hard to translate them because the commit messages list looks > like some kind of puzzle. This is a good point, but the i18n

D15406: Manually resize KCMUtilDialog to sizeHint()

This revision was automatically updated to reflect the committed changes. Closed by commit R295:cc4ecfdcd48a: Manually resize KCMUtilDialog to sizeHint() (authored by valeriymalov). REPOSITORY R295 KCMUtils CHANGES SINCE LAST UPDATE https://phabricator.kde.org/D15406?vs=41344=41446

Re: Quality of Frameworks announcements

Hi, if you have any commits where the heading is a very technical description, consider using the CHANGELOG: tag, for instance Look up foo directly instead of doing bar When doing foo we needlessly ended up doing bar which comes with a huge overhead. CHANGELOG: Significantly sped up doing

D15443: [server] Allow multiple touch interfaces per client and remove fallback code

romangg created this revision. romangg added reviewers: KWin, Frameworks. Herald added a project: Frameworks. Herald added a subscriber: kde-frameworks-devel. romangg requested review of this revision. REVISION SUMMARY The touch related code in the seat interface class has been for no apparent

D15422: [server] Fix remote access buffer handling when output not bound

romangg updated this revision to Diff 41422. romangg added a comment. Add autotest. Needed larger refactoring to allow multiple clients with different outputs being bound. REPOSITORY R127 KWayland CHANGES SINCE LAST UPDATE https://phabricator.kde.org/D15422?vs=41400=41422 BRANCH

KDE CI: Frameworks » kwayland » kf5-qt5 SUSEQt5.10 - Build # 84 - Fixed!

BUILD SUCCESS Build URL https://build.kde.org/job/Frameworks/job/kwayland/job/kf5-qt5%20SUSEQt5.10/84/ Project: kf5-qt5 SUSEQt5.10 Date of build: Tue, 11 Sep 2018 15:40:28 + Build duration: 4 min 41 sec and counting JUnit Tests Name: (root)

D15023: [server] Add selectionChanged signal

romangg updated this revision to Diff 41424. romangg added a comment. Rebase on master. REPOSITORY R127 KWayland CHANGES SINCE LAST UPDATE https://phabricator.kde.org/D15023?vs=41386=41424 BRANCH 0selectionChangeXwl REVISION DETAIL https://phabricator.kde.org/D15023 AFFECTED FILES

D15422: [server] Fix remote access buffer handling when output not bound

This revision was automatically updated to reflect the committed changes. Closed by commit R127:1faa629d2791: [server] Fix remote access buffer handling when output not bound (authored by romangg). REPOSITORY R127 KWayland CHANGES SINCE LAST UPDATE

D15431: Make PersonPluginManager API public

poboiko created this revision. poboiko added reviewers: KDE PIM, dvratil, apol. poboiko added a project: KDE PIM. Herald added a project: Frameworks. Herald added a subscriber: kde-frameworks-devel. poboiko requested review of this revision. REVISION SUMMARY This patch makes PersonPluginManager

D15422: [server] Fix remote access buffer handling when output not bound

romangg added a comment. At several places only one of two signal spies were waiting, what can lead to flickering tests. Fixed this with an if (spy.size() == 0) { spy.wait(); } on the second spy. Is there a nicer solution? REPOSITORY R127 KWayland BRANCH

D15420: Intialize m_lastPosition

This revision was automatically updated to reflect the committed changes. Closed by commit R39:3348145dbfbf: Intialize m_lastPosition (authored by jtamate). REPOSITORY R39 KTextEditor CHANGES SINCE LAST UPDATE https://phabricator.kde.org/D15420?vs=41393=41419 REVISION DETAIL

D15422: [server] Fix remote access buffer handling when output not bound

romangg updated this revision to Diff 41423. romangg added a comment. - Some more cleanup of test. REPOSITORY R127 KWayland CHANGES SINCE LAST UPDATE https://phabricator.kde.org/D15422?vs=41422=41423 BRANCH remoteBufferClientsUnbindOutput REVISION DETAIL

KDE CI: Frameworks » kjsembed » kf5-qt5 WindowsMSVCQt5.11 - Build # 4 - Failure!

BUILD FAILURE Build URL https://build.kde.org/job/Frameworks/job/kjsembed/job/kf5-qt5%20WindowsMSVCQt5.11/4/ Project: kf5-qt5 WindowsMSVCQt5.11 Date of build: Tue, 11 Sep 2018 16:45:18 + Build duration: 39 sec and counting CONSOLE OUTPUT

KDE CI: Frameworks » kjs » kf5-qt5 WindowsMSVCQt5.11 - Build # 4 - Failure!

BUILD FAILURE Build URL https://build.kde.org/job/Frameworks/job/kjs/job/kf5-qt5%20WindowsMSVCQt5.11/4/ Project: kf5-qt5 WindowsMSVCQt5.11 Date of build: Tue, 11 Sep 2018 16:14:43 + Build duration: 57 sec and counting CONSOLE OUTPUT

D15426: Avoid QByteArray::remove in AccessManagerReply::readData

fvogt updated this revision to Diff 41427. fvogt added a comment. - Actually make it work - Free memory if m_offset is half of the array size - Bail out if maxSize < 0 REPOSITORY R241 KIO CHANGES SINCE LAST UPDATE https://phabricator.kde.org/D15426?vs=41406=41427 BRANCH master

D15426: Avoid QByteArray::remove in AccessManagerReply::readData

fvogt edited the summary of this revision. fvogt edited the test plan for this revision. REPOSITORY R241 KIO REVISION DETAIL https://phabricator.kde.org/D15426 To: fvogt, #frameworks, elvisangelaccio Cc: kde-frameworks-devel, michaelh, ngraham, bruns

D15406: Manually resize KCMUtilDialog to sizeHint()

davidedmundson added a comment. > neither QWidget::adjustSize() nor QWidget::adjustedSize() seem to use devicePixelRatio They shouldn't. Both qwidget sizes and qscreen sizes are in logical pixels. REPOSITORY R295 KCMUtils BRANCH master REVISION DETAIL

D15426: Avoid QByteArray::remove in AccessManagerReply::readData

fvogt edited the summary of this revision. REPOSITORY R241 KIO REVISION DETAIL https://phabricator.kde.org/D15426 To: fvogt, #frameworks, elvisangelaccio Cc: kde-frameworks-devel, michaelh, ngraham, bruns

KDE CI: Frameworks » kpackage » kf5-qt5 AndroidQt5.11 - Build # 18 - Still Failing!

BUILD FAILURE Build URL https://build.kde.org/job/Frameworks/job/kpackage/job/kf5-qt5%20AndroidQt5.11/18/ Project: kf5-qt5 AndroidQt5.11 Date of build: Tue, 11 Sep 2018 18:03:59 + Build duration: 35 sec and counting CONSOLE OUTPUT [...truncated

KDE CI: Frameworks » kpackage » kf5-qt5 SUSEQt5.9 - Build # 50 - Still Unstable!

BUILD UNSTABLE Build URL https://build.kde.org/job/Frameworks/job/kpackage/job/kf5-qt5%20SUSEQt5.9/50/ Project: kf5-qt5 SUSEQt5.9 Date of build: Tue, 11 Sep 2018 18:03:58 + Build duration: 6 min 24 sec and counting JUnit Tests Name: (root)

KDE CI: Frameworks » kservice » kf5-qt5 WindowsMSVCQt5.11 - Build # 5 - Failure!

BUILD FAILURE Build URL https://build.kde.org/job/Frameworks/job/kservice/job/kf5-qt5%20WindowsMSVCQt5.11/5/ Project: kf5-qt5 WindowsMSVCQt5.11 Date of build: Tue, 11 Sep 2018 18:38:44 + Build duration: 2 min 36 sec and counting CONSOLE OUTPUT

D15426: Avoid QByteArray::remove in AccessManagerReply::readData

fvogt added inline comments. INLINE COMMENTS > bruns wrote in accessmanagerreply_p.cpp:156 > Wouldn't it be better to trim the buffer in slotData, at least in the > non-trivial case? For the non-trivial case it shouldn't make a difference really. For the trivial one it does, as otherwise it

KDE CI: Frameworks » kpackage » kf5-qt5 SUSEQt5.10 - Build # 72 - Still Unstable!

BUILD UNSTABLE Build URL https://build.kde.org/job/Frameworks/job/kpackage/job/kf5-qt5%20SUSEQt5.10/72/ Project: kf5-qt5 SUSEQt5.10 Date of build: Tue, 11 Sep 2018 18:03:59 + Build duration: 4 min 30 sec and counting JUnit Tests Name: (root)

D13700: implement reading of the replaygain tags

astippich added inline comments. INLINE COMMENTS > bruns wrote in taglibextractor.cpp:943 > You should check if the value can be converted to double, and probably also > if it is in a sane range. I added the checks for a successful conversion, but in my opinion kfilemetadata should not judge

D13700: implement reading of the replaygain tags

astippich updated this revision to Diff 41429. astippich added a comment. - add checks for conversion to double REPOSITORY R286 KFileMetaData CHANGES SINCE LAST UPDATE https://phabricator.kde.org/D13700?vs=41119=41429 BRANCH replaygain REVISION DETAIL

D15426: Avoid QByteArray::remove in AccessManagerReply::readData

bruns added inline comments. INLINE COMMENTS > accessmanagerreply_p.cpp:156 > > -if (length) { > -memcpy(data, m_data.constData(), length); > -m_data.remove(0, length); > +// Remove the stale data before m_offset only if it grows to half > +// of the total size, to

D15426: Avoid QByteArray::remove in AccessManagerReply::readData

bruns added a comment. For the trivial case, do the clear in `readData()`. For the non-trivial case: 1. in `readData()`, no memmoves were ever done. Currently, if you have e.g. 50 kB in m_data, you read 2 * 16 kB, and move the remaining 18 kB to the front. You can instead just read