Re: [Development] Updating the minimal supported version of libxcb
> Another question is what to do with the bundled libraries? The -qt-xcb switch is there to reduce run-time dependencies. There always will be somebody who wants to run the latest Qt version on some old linux distribution. Also I think it was decided that in Qt5 we can remove support for -no-x11extenssion-name switches in the Qt configure script. There is always -qt-xcb available if your distro does not provide the required dependencies. And it does not add any value not to compile-in the extension code. It should be a runtime decision, if X server has an extension available, we use it. From: development-bounces+gatis.paeglis=theqtcompany@qt-project.orgon behalf of Александр Волков Sent: Friday, October 16, 2015 2:02 PM To: Knoll Lars; Thiago Macieira; development@qt-project.org Subject: Re: [Development] Updating the minimal supported version of libxcb 16.10.2015 09:28, Knoll Lars пишет: > On 15/10/15 17:59, "Thiago Macieira" wrote: > >> On Thursday 15 October 2015 18:26:23 Александр Волков wrote: >>> Hi, >>> >>> Currently Qt supports libxcb 1.5, which is very old (it was released on >>> December 3, 2009) >>> It complicates adding new bundled xcb libraries, because they have to be >>> checked >>> for compatibility with this old version and maybe patched. >>> >>> Qt officially supports the following platforms that use libxcb: >>> OpenSuSE 13.1 (libxcb 1.9.1, xcb-proto 1.8) >>> Red Hat Enterprise Linux 6.6 (libxcb 1.9, xcb-proto 1.8) >>> Ubuntu 14.04 - 64bit (libxcb 1.10, xcb-proto 1.10) >>> >>> What do you think about updating the minimal supported version of libxcb >>> to 1.9? >> Unless it's critical, let's do it only in 5.7. > Yes, I don’t know of larger issues that would require a change in 5.6. But > for 5.7 by all means go ahead and update the minimum requirement to 1.9. Another question is what to do with the bundled libraries? With libxcb 1.9 they are all provided by the supported distros. The only exception is xcb-xkb, which we require to be from libxcb 1.10. So my proposal is to remove them from Qt, remove configure options -qt-xcb and -system-xcb, and use the bundled xcb-xkb library if we can't find the appropriate version on the system. ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] Code Coverage Statistics for QtBase
Why can they be ignored? I mean, we use only some of the API from that library, but bundle all of it. If it is not fully covered by tests it is not an issue. From: Hausmann Simon Sent: Thursday, February 19, 2015 4:51 PM To: Paeglis Gatis; Sébastien Fricker; development Subject: Re: [Development] Code Coverage Statistics for QtBase Why can they be ignored? When the users run Qt applications that code is executed, so coverage is important IMO. However if the added code is from build tools (Code generators) then maybe it's not so critical, as long as the generated code is covered. That said, the results look incomplete to me. Qxcbwindow.cpp surely has non-zero coverage when running tests on X11. Simon From: Paeglis Gatis Sent: Thursday, February 19, 2015 23:45 To: Sébastien Fricker; development Subject: Re: [Development] Code Coverage Statistics for QtBase Many of those files are from 3rdparty code, the ones I recognize are from /qtbase/src/3rdparty/xkbcommon, those I think can be ignored when thinking from code coverage point of view. From: development-bounces+gatis.paeglis=theqtcompany@qt-project.org development-bounces+gatis.paeglis=theqtcompany@qt-project.org on behalf of Sébastien Fricker fric...@froglogic.com Sent: Thursday, February 19, 2015 3:51 PM To: development Subject: [Development] Code Coverage Statistics for QtBase Hi, According http://download.froglogic.com/public/qt5-squishcoco-report/ there is a big decrease of the code coverage between Qt5/dev and Qt5.4. The code coverage from QtBase is decreasing from 56% to 42%. The first reason of that, is that approximately 100 new files get added into the project and the code coverage of this set is near to zero (see the HTML attachment and the table below). This new uncovered files are responsible of a lost of the code coverage of 3% approximately. Question: are these files not tested through an unit test? If they are tested, I need to look why they are not taken in account by the code coverage analysis. The main difference are caused by the code coverage of the unit tests itself. The number of unit tests did not decrease, but the code coverage of the tests with the status Unknown decrease from 13% (from 43% to 30%) and this explain the rest of the code coverage lost. These tests are in fact the code coverage of the code which is running before and after a unit test (setup the GUI, common initializations executed before main(), exit code, ...). Question: are they changes between Qt5.4 and Qt5/dev in QTestLib (in the unit tests or in QtBase) which could explain this? Regards, Sébastien List of new files: Source Coverage (1 means 100%) action.c0 ast-build.c 0 atom.c 0 compat.c0 context-priv.c 0 context.c 0 expr.c 0 include.c 0 keycodes.c 0 keymap-dump.c 0 keymap-priv.c 0 keymap.c0 keysym-utf.c0 keysym.c0 keywords.c 0 parser.c0 parser.y0 qbasicfontdatabase.cpp 0 qdbusmenuadaptor.cpp0 qdbusmenuconnection.cpp 0 qdbusmenutypes.cpp 0 qdbusplatformmenu.cpp 0 qdbustrayicon.cpp 0 qdbustraytypes.cpp 0 qdevicediscovery_udev.cpp 0 qeglconvenience.cpp 0 qeglfscontext.cpp 0 qeglfsdeviceintegration.cpp 0 qeglfshooks.cpp 0 qeglfsintegration.cpp 0 qeglfsoffscreenwindow.cpp 0 qeglfsscreen.cpp0 qeglfswindow.cpp0 qeglpbuffer.cpp 0 qeglplatformcontext.cpp 0 qeglplatformcursor.cpp 0 qeglplatformintegration.cpp 0 qeglplatformscreen.cpp 0 qeglplatformwindow.cpp 0 qevdevkeyboardhandler.cpp 0 qevdevkeyboardmanager.cpp 0 qevdevmousehandler.cpp 0 qevdevmousemanager.cpp 0 qevdevtablet.cpp0 qevdevtouch.cpp 0 qeventdispatcher_glib.cpp 0 qfbbackingstore.cpp 0 qfbcursor.cpp 0 qfbscreen.cpp 0 qfbvthandler.cpp0 qfbwindow.cpp 0 qfontconfigdatabase.cpp 0 qfontengine_ft.cpp 0 qfontenginemultifontconfig.cpp 0 qgenericunixeventdispatcher.cpp 0 qgenericunixservices.cpp0 qgenericunixthemes.cpp 0 qglxconvenience.cpp 0 qinputdevicemanager.cpp 0 qopenglcompositor.cpp 0 qopenglcompositorbackingstore.cpp 0 qopenglfunctions_4_4_compatibility.cpp 0 qopenglfunctions_4_4_core.cpp 0 qopenglfunctions_4_5_compatibility.cpp 0 qopenglfunctions_4_5_core.cpp 0 qplatformgraphicsbuffer.cpp 0 qplatformgraphicsbufferhelper.cpp 0 qsslellipticcurve.cpp 0 qsslpresharedkeyauthenticator.cpp 0 qstatusnotifieritemadaptor.cpp 0 qunixeventdispatcher.cpp0 qxcbbackingstore.cpp0 qxcbclipboard.cpp 0 qxcbconnection.cpp 0 qxcbconnection_xi2.cpp 0 qxcbcursor.cpp 0 qxcbdrag.cpp0 qxcbglintegration.cpp 0 qxcbglintegrationfactory.cpp0 qxcbimage.cpp 0 qxcbintegration.cpp 0 qxcbkeyboard.cpp0 qxcbmime.cpp0 qxcbnativeinterface.cpp 0 qxcbnativeinterfacehandler.cpp 0 qxcbscreen.cpp 0 qxcbsessionmanager.cpp 0
Re: [Development] Code Coverage Statistics for QtBase
Many of those files are from 3rdparty code, the ones I recognize are from /qtbase/src/3rdparty/xkbcommon, those I think can be ignored when thinking from code coverage point of view. From: development-bounces+gatis.paeglis=theqtcompany@qt-project.org development-bounces+gatis.paeglis=theqtcompany@qt-project.org on behalf of Sébastien Fricker fric...@froglogic.com Sent: Thursday, February 19, 2015 3:51 PM To: development Subject: [Development] Code Coverage Statistics for QtBase Hi, According http://download.froglogic.com/public/qt5-squishcoco-report/ there is a big decrease of the code coverage between Qt5/dev and Qt5.4. The code coverage from QtBase is decreasing from 56% to 42%. The first reason of that, is that approximately 100 new files get added into the project and the code coverage of this set is near to zero (see the HTML attachment and the table below). This new uncovered files are responsible of a lost of the code coverage of 3% approximately. Question: are these files not tested through an unit test? If they are tested, I need to look why they are not taken in account by the code coverage analysis. The main difference are caused by the code coverage of the unit tests itself. The number of unit tests did not decrease, but the code coverage of the tests with the status Unknown decrease from 13% (from 43% to 30%) and this explain the rest of the code coverage lost. These tests are in fact the code coverage of the code which is running before and after a unit test (setup the GUI, common initializations executed before main(), exit code, ...). Question: are they changes between Qt5.4 and Qt5/dev in QTestLib (in the unit tests or in QtBase) which could explain this? Regards, Sébastien List of new files: Source Coverage (1 means 100%) action.c0 ast-build.c 0 atom.c 0 compat.c0 context-priv.c 0 context.c 0 expr.c 0 include.c 0 keycodes.c 0 keymap-dump.c 0 keymap-priv.c 0 keymap.c0 keysym-utf.c0 keysym.c0 keywords.c 0 parser.c0 parser.y0 qbasicfontdatabase.cpp 0 qdbusmenuadaptor.cpp0 qdbusmenuconnection.cpp 0 qdbusmenutypes.cpp 0 qdbusplatformmenu.cpp 0 qdbustrayicon.cpp 0 qdbustraytypes.cpp 0 qdevicediscovery_udev.cpp 0 qeglconvenience.cpp 0 qeglfscontext.cpp 0 qeglfsdeviceintegration.cpp 0 qeglfshooks.cpp 0 qeglfsintegration.cpp 0 qeglfsoffscreenwindow.cpp 0 qeglfsscreen.cpp0 qeglfswindow.cpp0 qeglpbuffer.cpp 0 qeglplatformcontext.cpp 0 qeglplatformcursor.cpp 0 qeglplatformintegration.cpp 0 qeglplatformscreen.cpp 0 qeglplatformwindow.cpp 0 qevdevkeyboardhandler.cpp 0 qevdevkeyboardmanager.cpp 0 qevdevmousehandler.cpp 0 qevdevmousemanager.cpp 0 qevdevtablet.cpp0 qevdevtouch.cpp 0 qeventdispatcher_glib.cpp 0 qfbbackingstore.cpp 0 qfbcursor.cpp 0 qfbscreen.cpp 0 qfbvthandler.cpp0 qfbwindow.cpp 0 qfontconfigdatabase.cpp 0 qfontengine_ft.cpp 0 qfontenginemultifontconfig.cpp 0 qgenericunixeventdispatcher.cpp 0 qgenericunixservices.cpp0 qgenericunixthemes.cpp 0 qglxconvenience.cpp 0 qinputdevicemanager.cpp 0 qopenglcompositor.cpp 0 qopenglcompositorbackingstore.cpp 0 qopenglfunctions_4_4_compatibility.cpp 0 qopenglfunctions_4_4_core.cpp 0 qopenglfunctions_4_5_compatibility.cpp 0 qopenglfunctions_4_5_core.cpp 0 qplatformgraphicsbuffer.cpp 0 qplatformgraphicsbufferhelper.cpp 0 qsslellipticcurve.cpp 0 qsslpresharedkeyauthenticator.cpp 0 qstatusnotifieritemadaptor.cpp 0 qunixeventdispatcher.cpp0 qxcbbackingstore.cpp0 qxcbclipboard.cpp 0 qxcbconnection.cpp 0 qxcbconnection_xi2.cpp 0 qxcbcursor.cpp 0 qxcbdrag.cpp0 qxcbglintegration.cpp 0 qxcbglintegrationfactory.cpp0 qxcbimage.cpp 0 qxcbintegration.cpp 0 qxcbkeyboard.cpp0 qxcbmime.cpp0 qxcbnativeinterface.cpp 0 qxcbnativeinterfacehandler.cpp 0 qxcbscreen.cpp 0 qxcbsessionmanager.cpp 0 qxcbsystemtraytracker.cpp 0 qxcbwindow.cpp 0 qxcbwmsupport.cpp 0 qxcbxsettings.cpp 0 qxdgnotificationproxy.cpp 0 qxlibeglintegration.cpp 0 rules.c 0 scanner.c 0 state.c 0 symbols.c 0 text.c 0 types.c 0 utf8.c 0 utils.c 0 vmod.c 0 xkb-compat.c0 xkb-keymap.c0 xkbcomp.c 0 forkfd.c0,2841726619 qsharedmemory_systemv.cpp 0,5789473684 qsystemsemaphore_systemv.cpp0,78125 qsslellipticcurve_openssl.cpp 1 ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] optimizing QComposeInputContext / TableGenerator?
Hi, The idea is to move the compose file reading code out of Qt. libxkbcommon 5.0 (released on 2014-10-18 [1]) added support for the compose keys. I did look at the new API already in October [2]. It should be simple enough to remove the table generation code and use xkbcommon's API instead. I have not done any benchmarking because I never got around to finishing that patch. There are several projects that use xkbcommon (Wayland, GTK), so we can benefit from the optimizations done by others if we use this new API. Also if somebody has suggestions for improvements, those should be done in [3]. [1] http://lists.freedesktop.org/archives/wayland-devel/2014-October/017836.html [2] *WIP* https://codereview.qt-project.org/#/c/98062/ [3] https://github.com/xkbcommon/libxkbcommon From: development-bounces+gatis.paeglis=theqtcompany@qt-project.org development-bounces+gatis.paeglis=theqtcompany@qt-project.org on behalf of Thiago Macieira thiago.macie...@intel.com Sent: Monday, January 19, 2015 9:38 PM To: development@qt-project.org Subject: Re: [Development] optimizing QComposeInputContext / TableGenerator? On Monday 19 January 2015 21:17:00 Giuseppe D'Angelo wrote: The best approach of course would be to have a OpenDesktop standard that allows mmapping the compose table in and using it from there. Probably not feasible. Why not? x...@freedesktop.org. I think I can easily convince the EFL folks for this. -- Thiago Macieira - thiago.macieira (AT) intel.com Software Architect - Intel Open Source Technology Center ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] Jira component re-arrangements
I would not be terribly upset if somebody that actually did work on the xcb plugin stepped up and took this one :) I would not say that I know all parts of the XCB plugin (yet), but I guess I feel comfortable enough to be the default assignee if nobody else is volunteering. Gatis. From: development-bounces+gatis.paeglis=digia@qt-project.org [development-bounces+gatis.paeglis=digia@qt-project.org] on behalf of Gunnar Sletta [gunnar.sle...@jolla.com] Sent: Tuesday, May 13, 2014 12:29 PM To: Blasche Alexander Cc: development@qt-project.org Subject: Re: [Development] Jira component re-arrangements On 13 May 2014, at 11:18, Blasche Alexander alexander.blas...@digia.com wrote: Hi, Following some discussions on this list and in Jira I have made changes to Jira components. If you have a filter that explicitly mentions one of the affected components, please adjust them now. Otherwise you filter may not return the correct results going forward. 1.) - Core: Gesture Support - Widget: Gesture Support (rename) - New owner for above component: Marc Mutz For more details see QTJIRA-255 2.) Following new components have been created: GUI: Windows integration - owner: Friedemann Kleint GUI: X11 (xcb) integration - owner: Gunnar Sletta I would not be terribly upset if somebody that actually did work on the xcb plugin stepped up and took this one :) GUI: OS X (cocoa) integration - owner: Morton Sorvig For more details see QTJIRA-256 3.) - Core: Event System - Core: Event loop (rename) For more details see QTJIRA-257 4.) - GUI: Input methods -GUI: Complex Input methods (rename) - GUI: Basic Input System (keyboard, mouse, touch) (created) - New owner for Basic Input System: Gunnar Sletta For more details see QTJIRA-258 -- Alex ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] qt5.3 beta - xkbcommon: ERROR
The latest change in that area was https://codereview.qt-project.org/#change,79342 , it requires re-configuring Qt. Did you do that? Also when using the bundled version of libkxbcommon, make sure that in config.summary you get: xkbcommon .. yes (bundled copy, xkb-config-root some_path_here) From: development-bounces+gatis.paeglis=digia@qt-project.org [development-bounces+gatis.paeglis=digia@qt-project.org] on behalf of Martin Koller [kol...@aon.at] Sent: Tuesday, March 25, 2014 9:15 PM To: development@qt-project.org Subject: [Development] qt5.3 beta - xkbcommon: ERROR I have installed qt5.3 pre21.1 today from openSuse repo. Now whenever I start a simple Qt5 app, I get: xkbcommon: ERROR: failed to add default include path auto Qt: Failed to create XKB context and I can not enter any text. any idea what the problem could be ? Is than an openSuse problem and should be reported there ? -- Best regards/Schöne Grüße Martin A: Because it breaks the logical sequence of discussion Q: Why is top posting bad? () ascii ribbon campaign - against html e-mail /\- against proprietary attachments Geschenkideen, Accessoires, Seifen, Kulinarisches: www.bibibest.at ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] Wiki: 3rd party in Qt
Regarding the xkbcommon, we are bundling 0.2.0, not 0.3.1. I am waiting for the next release of xkbcommon, which will provide few new functions that I want to use in the xcb platform plugin. That is when I am planning to update the bundled version as well. Gatis. From: development-bounces+gatis.paeglis=digia@qt-project.org [development-bounces+gatis.paeglis=digia@qt-project.org] on behalf of Liang Qi [cavendish...@gmail.com] Sent: Wednesday, November 20, 2013 3:03 PM To: development@qt-project.org Subject: [Development] Wiki: 3rd party in Qt Hi, all, I just spent some time to gather the information of 3rd party in Qt(5). http://qt-project.org/wiki/Third_Party_In_Qt Please help to update or maintain it. Maybe there is a legal issue about wintab, more details in their webpage: http://www.pointing.com/Wintab.html Regards, Liang ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] Changing keyboard layouts in Qt 5.1 apps with GNOME
Its probably because of the issue described in: https://bugreports.qt-project.org/browse/QTBUG-30911 https://bugs.freedesktop.org/show_bug.cgi?id=23202 Switching between layouts should work in all cases, but adding a new layout might not be seen until the application is restarted. From: development-bounces+gatis.paeglis=digia@qt-project.org [development-bounces+gatis.paeglis=digia@qt-project.org] on behalf of Thiago Macieira [thiago.macie...@intel.com] Sent: Saturday, August 31, 2013 6:35 PM To: development@qt-project.org Subject: Re: [Development] Changing keyboard layouts in Qt 5.1 apps with GNOME On sábado, 31 de agosto de 2013 14:51:58, Petko Ditchev wrote: I need some help troubleshooting a problem I've been having : for a week now (I think since some updates to the keyboard layout settings in gnome) I can't change the keyboard layout in QtCreator (built with Qt5.1) and in my app that is on the same lib . Otherwise everything's ok , but Qt5.1 apps stick with the layout they are launched with . I'm sending this to both lists because the bug affects GNOME , but not Unity , and it affects Qt5.1 , but not Qt4 . Can you provide us with the output of setxkbmap -print before and after the keyboard change? It would be useful to get the xev output for a keystroke that changed too (before and after). Finally, can you make the same test by switching keyboard layouts with setxkbmap? -- Thiago Macieira - thiago.macieira (AT) intel.com Software Architect - Intel Open Source Technology Center ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] Why I can't use ibus Input Method for type Chinese character?
Did you set the environmental variable to QT_IM_MODULE=ibus? From: development-bounces+gatis.paeglis=digia@qt-project.org [development-bounces+gatis.paeglis=digia@qt-project.org] on behalf of Yuchen Deng [loa...@gmail.com] Sent: Saturday, May 11, 2013 2:54 PM To: development Subject: [Development] Why I can't use ibus Input Method for type Chinese character? Hi, there! I want help because I can't input any Chinese character for Qt5 based application, include QtCreator when build with Qt5.1. I just know I need build D-Bus for plugins/platforminputcontexts/ibus. Then I did, and I can sure QtCreator loaded the ibus plugins. Here is the full log: http://pastebin.mozilla.org/2393461 But I still can't input Chinese in QtCreator, because ibus tell me there `No input window`. I asked in #qt, #qt-labs, and #qt-cn, but seems nobody know why and how. So I think it's maybe is a bug. Any comments are welcome. Help needed. thanks a lot! -- Best Regards Yuchen ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] Why I can't use ibus Input Method for type Chinese character?
Why Qt5 need it? In Qt5 this variable is read to determine which input context plugin to use, by default it is set to compose. [1] Other keys are ibus and maliit. [1] https://codereview.qt-project.org/#change,51555 From: Yuchen Deng [loa...@gmail.com] Sent: Saturday, May 11, 2013 4:18 PM To: Paeglis Gatis Cc: development Subject: Re: [Development] Why I can't use ibus Input Method for type Chinese character? 2013/5/11 Paeglis Gatis gatis.paeg...@digia.commailto:gatis.paeg...@digia.com QT_IM_MODULE=ibus I did not set this environment variable. and I can sure after set this environment variable to ibus, then the issue just gone! But why? no need this environment, all other applications works fine with ibus. Why Qt5 need it? -- Best Regards Yuchen ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] include X11/keysym.h
It is a leftover, I have submitted a change: https://codereview.qt-project.org/#change,52599 From: development-bounces+gatis.paeglis=digia@qt-project.org [development-bounces+gatis.paeglis=digia@qt-project.org] on behalf of Thiago Macieira [thiago.macie...@intel.com] Sent: Thursday, March 28, 2013 4:02 PM To: development@qt-project.org Subject: Re: [Development] include X11/keysym.h On quinta-feira, 28 de março de 2013 13.35.22, Thomas Senyk wrote: Question on the side: This is a commit for stable, right? ... It can be considered a build-bug-fix... or not? Stable, it's a build fix. -- Thiago Macieira - thiago.macieira (AT) intel.com Software Architect - Intel Open Source Technology Center ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
[Development] input method support in Qt5 (linux)
After looking at Qt4 / Qt5 source code and doing some internet search I see that ibus plugin has become a part of Qt, where in Qt4 it was a separate plugin hosted somewhere else.. Qt4 offered two options XIM and IBUS, selectable by using 'qtconfig' tool, currently Qt5 has support only for ibus as input method framework (IMF) , which does look like a great IM framework, but is it complete enough to be the only available option in Qt for desktop? It's very important to the Qt project that input works well for all users without exhausting tweaking, but at the present it is quite frustrating with ibus-setup GUI tool which almost never does what you tell it to do. It seems that the latest ibus release supports good additions like: - Improve IME switch UI - Integrate IME and XKB keyboard layouts. Now ibus can manage both IME and XKB layouts. etc. AFAIK linux distros doesn't ship by default this version yet. I don't know if there has been any decision regarding this.. I was wondering if we should bring back xinput support in XCB plugin? [1] [1] http://lists.freedesktop.org/archives/xcb/2013-January/008059.html ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development