Summary: | [Qt] Baseline qt_minimal configuration | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Laszlo Gombos <laszlo.gombos> | ||||||||
Component: | Platform | Assignee: | Nobody <webkit-unassigned> | ||||||||
Status: | RESOLVED FIXED | ||||||||||
Severity: | Normal | CC: | benjamin, commit-queue, ossy, smagnuso | ||||||||
Priority: | P2 | Keywords: | Qt, QtTriaged | ||||||||
Version: | 528+ (Nightly build) | ||||||||||
Hardware: | PC | ||||||||||
OS: | All | ||||||||||
Attachments: |
|
Description
Laszlo Gombos
2010-12-19 17:15:56 PST
Created attachment 76965 [details]
proposed patch
I left the following configurations out from the baseline as these configurations did not resulted in a functional build (QtTestBrowser usually crashed at startup). I will file a separate bugs/patchess (e.g 51221) to try to fix these (or perhaps remove them and declare them unsupported).
- QT_NO_ACTION,
- QT_NO_CONTEXTMENU
- QT_NO_DATESTRING" (autotest framework seems to have a dependency on this)
- QT_NO_DRAGANDDROP
- QT_NO_LIBRARY
- QT_NO_PROPERTIES
- QT_NO_SETTINGS
- QT_NO_WHEELEVENT
Created attachment 76967 [details]
same as previous, but replace += to *= in WebKit.pri
I really like the idea of having that tested on a bot. It would make it easier not to break those configurations by accident. Is QtTestBrowser usable in this qt_minimal configuration? Maybe you could also create an official qconfig with which it is possible to compile Qt with WebKit? This minimal Qt would be used for testing by us and could be used by users asking for minimal QtWebKit footprint. (In reply to comment #3) > Is QtTestBrowser usable in this qt_minimal configuration? Yes. > Maybe you could also create an official qconfig with which it is possible to compile Qt with WebKit? This minimal Qt would be used for testing by us and could be used by users asking for minimal QtWebKit footprint. Let's establish that configuration in WebKit and than perhaps we can push it "down" to Qt. Comment on attachment 76967 [details] same as previous, but replace += to *= in WebKit.pri View in context: https://bugs.webkit.org/attachment.cgi?id=76967&action=review Great patch, thanks for it. ;) Please fix typos and QT_NO_BEARERMANAGEMENT before landing. (I tested the patch, QtTestBrowser works fine.) > ChangeLog:9 > + * WebKit.pri: List the supported QT_NO_FUTURE flags > + under qt_minial configuration. s/QT_NO_FUTURE/QT_NO_FEATURE s/minial/minimal > Tools/QtTestBrowser/launcherwindow.h:79 > +class QPropertyAnimation; > + Why do we need it? > WebKit.pri:118 > + DEFINES *= QT_NO_BEARERMANAGEMENT To use QT_NO_BEARERMANAGEMENT we need one more little patch. Including of moc_NetworkStateNotifierPrivate.cpp should be inside the ENABLE(BEARER) guard: diff --git a/WebCore/platform/network/qt/NetworkStateNotifierQt.cpp b/WebCore/platform/network/qt/NetworkStateNotifierQt.cpp index 959e74a..f3e7023 100644 --- a/WebCore/platform/network/qt/NetworkStateNotifierQt.cpp +++ b/WebCore/platform/network/qt/NetworkStateNotifierQt.cpp @@ -91,6 +91,6 @@ void NetworkStateNotifier::setNetworkAccessAllowed(bool isAllowed) } // namespace WebCore -#endif - #include "moc_NetworkStateNotifierPrivate.cpp" + +#endif I'm going to build a qt_minimal with same QT_NO_FEATURES options, and make the QtWebKit minimal buildbot use this minimal Qt and build QtWebKit with additional CONFIG+=qt_minimal flag. It will guarantee that folks won't break this supported qt_minimal build by accident. Additionally we are planning to start running QtWebKit unittests (run-qtwebkit-tests) on the minimal bot too. (In reply to comment #1) > - QT_NO_DATESTRING" (autotest framework seems to have a dependency on this) It seems it is a bug in Qt 4.7.1, because toString methods should be guarded by #ifndef QT_NO_DATESTRING in qtest.h. I'll file a bug in Qt bug tracker about it. Created attachment 78477 [details]
Patch with proposed minor changes for cq
Comment on attachment 78477 [details] Patch with proposed minor changes for cq Clearing flags on attachment: 78477 Committed r75467: <http://trac.webkit.org/changeset/75467> All reviewed patches have been landed. Closing bug. |