Re: [Interest] Compile Speed Question regarding MOC
Till Oliver Knoll till.oliver.kn...@gmail.com wrote: With regards to 1): A ful rebuild takes the time it takes, no matter what. You can speed that up things as others have suggested (#include the mocced files, AUTOMOC), but that's that. What about using pre-compiled headers (pch) with MSVC? I imagine the total time for a full rebuild with MSVC (using PCH and multiple .cc-files in one go) is much less than a partial build using e.g. gcc/MingW. I'm not sure how good PCH in gcc has become. I ask since I've not tried it. I'm not sure it's possible to use MSVC's PCH with Qt. --gv ___ Interest mailing list Interest@qt-project.org http://lists.qt-project.org/mailman/listinfo/interest
Re: [Interest] libeay32.dll - The Ordinal 4369 could not be located
Richard Moore r...@kde.org wrote: On 24 January 2014 12:20, Phil Hannent p...@hannent.co.uk wrote: However could you explain how QtWebkit is able to view https pages when the libeasy32.dll is not in my path or next to the application? It must be located in one of the many locations that are searched. On linux I'd suggest you use strace to check where you're actually getting it from, I'm not sure of the best equivalent on windows though this tool from sysinternals might work. http://technet.microsoft.com/en-gb/sysinternals/bb896656 I'd suggest Process Monitor instead: http://technet.microsoft.com/en-gb/sysinternals/bb896645 It's complicated to setup a filter that shows only LoadLibrary(ssleay32.dll) calls, but this tool can show in much more details what a process is doing. --gv ___ Interest mailing list Interest@qt-project.org http://lists.qt-project.org/mailman/listinfo/interest
[Interest] 4NT shell and qmake
I just installed QtCreator 3.0.3 along with the latest Qt 5.2.0 SDK using the online installer. I loaded up one of the project in the QtCreator example-pane. The YouTubeView project. I found that the bug I reported for Qt4 [1+2] involving nmake (or jom?) trying a del non-existant-file is still present in Qt 5. Here is the step where (the qmake generated) nmake is trying to add a manifest (edited lines for clarity): link /NOLOGO /DYNAMICBASE /NXCOMPAT /DEBUG /SUBSYSTEM:WINDOWS /MANIFESTDEPENDENCY:type='win32' name='Microsoft.Windows.Common-Controls' version='6.0.0.0' publicKeyToken='6595b64144ccf1df' language='*' processorArchitecture='*' /MANIFEST /MANIFESTFILE:debug\youtubeview.exe.embed.manifest /OUT:debug\youtubeview.exe @F:\TEMP\youtubeview.exe.8172.1718.jom if exist debug\youtubeview.exe_manifest.bak fc /b debug\youtubeview.exe.embed.manifest debug\youtubeview.exe_manifest.bak NUL || del debug\youtubeview.exe_manifest.bak 4NT: (Sys) Systemet finner ikke angitt fil. ! error 2 = File not found (ERROR_FILE_NOT_FOUND) G:\Qt\5.2\5.2.0\mingw48_32\examples\webkitqml\build-youtubeview-Desktop_Qt_5_2_0_MSVC2010_32bit_ OpenGL- Debug\debug\youtubeview.exe_manifest.bak 0 files deleted jom: G:\Qt\5.2\5.2.0\mingw48_32\examples\webkitqml\build-youtubeview-Desktop_Qt_5_2_0_MSVC2010_32bit_OpenGL-Debug\Makefile.Debug [debug\youtubeview.exe] Error 2 jom: G:\Qt\5.2\5.2.0\mingw48_32\examples\webkitqml\build-youtubeview-Desktop_Qt_5_2_0_MSVC2010_32bit_OpenGL-Debug\Makefile [debug] Error 2 23:04:09: The process G:\Qt\5.2\Tools\QtCreator\bin\jom.exe exited with code 2. - Pushing the Build project once more, the project builds since now (it seems the .bak) file is present. I use 4NT 5.0 Unideo shell on Win-XP SP3. Please don't advice me to use the inferior cmd shell :-) [1] https://bugreports.qt-project.org/browse/QTBUG-29698 [2] http://lists.qt-project.org/pipermail/interest/2013-February/005990.html This bug-report of mine seems closed. But the bug is still there. --gv ___ Interest mailing list Interest@qt-project.org http://lists.qt-project.org/mailman/listinfo/interest
Re: [Interest] 4NT shell and qmake
Thiago Macieira thiago.macie...@intel.com wrote: You can hack qmake or the mkspec so that it works. We need a command that deletes a file if it exists or exits without error if it doesn't. The equivalent of rm -f on Unix. I edited my %QT_INSTALL_PREFIX\mkspecs\common\shell-win32.conf and replaced: QMAKE_DEL_FILE = del with: QMAKE_DEL_FILE = del /e /k This works fine. Thanks. Are the shell-win32.conf commands used universally now? AFAICR some commands where hardcoded into qmake previously. PS. 4NT docs sys: /E (No error messages) Suppress all non-fatal error messages, such as File Not Found. Fatal error messages, such as Drive not ready, will still be displayed. This option is most useful in batch files and aliases. /K Physically delete files instead of sending them to the Windows Recycle Bin. This option overrides the default RecycleBin=Yes .INI setting --gv ___ Interest mailing list Interest@qt-project.org http://lists.qt-project.org/mailman/listinfo/interest
Re: [Interest] 4NT shell and qmake
Bob Hood bho...@comcast.net wrote: I edited my %QT_INSTALL_PREFIX\mkspecs\common\shell-win32.conf and replaced: QMAKE_DEL_FILE = del with: QMAKE_DEL_FILE = del /e /k This works fine. Thanks. Thanks for that, Gisle. I can use my precious TCCLE (a.k.a., 4NT) again. :) I spoke too soon. del /e /k isn't used in (allways?) QtCreator. I've checked all my shell-win32.conf to figure out which one QtCreator is using. Seems none of them and the del non-existant-file cannot be overridden manually. Yikes! But 4NT/TCC does have global aliases. So I tried alias del=*del /e /k followed by: shralias in the hope this will be picked up by the 4NT-shell QtCreator is using, but no luck. I must be misunderstanding 'shralias'. --gv ___ Interest mailing list Interest@qt-project.org http://lists.qt-project.org/mailman/listinfo/interest
[Interest] How to keep the response files?
Is it possible to tell qmake to generate a KEEP inside the Nmake generated makefiles? Some times I need to see what's been feed to the compiler and linker. Like so: $(DESTDIR_TARGET): $(OBJECTS) . $(LINK) $(LFLAGS) /OUT:$(DESTDIR_TARGET) @ $(OBJECTS) $(LIBS) KEEP Alternatively, is it possible to not use @resp-files? But rather show the complete cmd-line as nmake progresses. --gv ___ Interest mailing list Interest@qt-project.org http://lists.qt-project.org/mailman/listinfo/interest
Re: [Interest] Buggy Makefile for config.tests\arch?
Joerg Bornemann joerg.bornem...@digia.com wrote: Even if the file does not exist, del won't set ERRORLEVEL. Try this Makefile: first: del thisdoesnotexist echo still working... Does del return with an exit code != 0 on your system? It sure does (as it should IMHO). I use 4NT 5.0U. nmake -nologo: del thisdoesnotexist 4NT: (Sys) Systemet finner ikke angitt fil. ERROR_FILE_NOT_FOUND = 2 C:\TEMP\tmp\thisdoesnotexist 0 files deleted NMAKE : fatal error U1077: 'del' : return code '0x2' Stop. -- But cmd doesn't set errorlevel. cmd /c nmake -nologo del thisdoesnotexist Finner ikke C:\TEMP\tmp\thisdoesnotexist echo still working... still working... But modifying my shell-win32.conf doesn't seems to help; The del command there isn't used in all places by qmake. So only that config test barfs up and Qt builds fine after this? No. I expect the whole build of Qt5 three to break. I'm sticking to Qt 4.8.2 that works with my shell (with some tweaking of some .conf- files). PS. The 4NT docs on ERRORLEVEL says: Various implementations of Microsoft's CMD.EXE also set ERRORLEVEL for some but not all internal commands. Our products attempt to emulate that behavior, but... My emphasis on not all internal commands. Not sure I can turn off errorlevel or OS-error codes for del. A fix for me is to use: alias del=rm -f --gv ___ Interest mailing list Interest@qt-project.org http://lists.qt-project.org/mailman/listinfo/interest
Re: [Interest] Buggy Makefile for config.tests\arch?
Joerg Bornemann joerg.bornem...@digia.com wrote: Note that this is a standard qmake generated Makefile and nothing special. arch.exe.embed.manifest is generated by the linker and usually stems from a build you did before. This is my first build. So how can arch.exe.embed.manifest already be present? Okay, I retested this with content in g:\Qt\5.0\qtbase\config.tests\arch only containing arch.cpp + arch.pro. I.e. a nmake distclean del arch.exe.manifest. Then with my env-vars set to: QMAKESPEC=win32-msvc2010 QTDIR=g:\Qt\5.0 QT_BUILD_DIR=g:\Qt\5.0 QT_INSTALL_EXAMPLES=g:\Qt\5.0\qtbase\examples QT_SOURCE_TREE=g:\Qt\5.0 Correct? So w/o patching arch.pro as you tipped me about below, issuing the commands qmake nmake -d produces this: on, feb 13 2013 14:32:44 arch.cpp on, feb 13 2013 14:33:06 G:\Qt\5.0\qtbase\src\corelib\global\qprocessordetection.h arch.obj target does not exist on, feb 13 2013 14:32:44 arch.cpp ** .\arch.cpp newer than arch.obj cl -c -nologo -Zm200 -Zc:wchar_t -Zi -MDd -W3 -w34100 -w34189 -EHsc -DUNICODE -DWIN32 -DQT_NO_DYNAMIC_CAST -IG:\Qt\5.0\qtbase\mkspecs\win32-msvc2010 -Fo @F:\TEMP\nm4263.tmp arch.cpp arch.exe target does not exist on, feb 13 2013 14:32:44 arch.cpp ** arch.cpp newer than arch.exe echo 1 /* CREATEPROCESS_MANIFEST_RESOURCE_ID */ 24 /* RT_MANIFEST */ ... arch.exe.embed.manifestarch.exe_manifest.rc if not exist arch.exe del arch.exe.embed.manifestNUL 21 NMAKE : fatal error U1077: 'if' : return code '0x2' Stop. - The test: if not exist arch.exe del arch.exe.embed.manifestNUL 21 is performed prior to the link step. So this generated Makefile is breaking causality AFAICS. I'm really tired of nmake (not even line numbers for these errors), but I just love GNU make. Is it not possible to use mingw32-make to build a MSVC2010 version of Qt5? Looks like the link call did not produce a manifest file. Can you confirm this? To work around this issue you could add CONFIG -= embed_manifest_exe to config.tests/arch/arch.pro That works fine. Thanks. --gv ___ Interest mailing list Interest@qt-project.org http://lists.qt-project.org/mailman/listinfo/interest