MiniBrowser fails to build on Symbian as the rules in Tools/MiniBrowser/DerivedSources.pro fails in the Symbian environment. Keith, Siddharth and myself discussed several ideas how this could be resolved; the option I would recommend (that involves the least amount of complexity) is to create two copy of useragentlist.txt under source control - one copy for QtTestBorwser and a separate copy for MiniBrowser.
*** Bug 56322 has been marked as a duplicate of this bug. ***
I am actively working on this bug.
Created attachment 86004 [details] Patch remove-derivedsources.pro
Comment on attachment 86004 [details] Patch OK. rs=me.
Comment on attachment 86004 [details] Patch As a result of this patch the user agent list for the MiniBrowser is empty on Linux. This patch breaks the UA list for MiniBrowser so I change this to r-. According to http://doc.qt.nokia.com/latest/resources.html "The resource files listed in the .qrc file are files that are part of the application's source tree. The specified paths are relative to the directory containing the .qrc file. Note that the listed resource files must be located in the same directory as the .qrc file, or one of its _subdirectories_."
Created attachment 87422 [details] Patch
I was unable to rip out DerivedSources.pro from Tools/MiniBrowser. Even with ripping it out, it would have remained for WebKitTestRunner, so there would not be much complexity reduction anyway. What I've done with this patch is have the source area be polluted for Symbian - copying over QtTestBrowser's useragentlist.txt file to the MiniBrowser source directory, and pointing to the .qrc in the source directory. In all other cases, I do what was done before - copying the QtTestBrowser/useragentlist.txt and the MiniBrowser.qrc files to the build output directory, and pointing at the qrc in the build output directory. I've tested this on Linux, but I have not tested it on Symbian. If someone could test this on Symbian, that would be swell.
(In reply to comment #7) > I was unable to rip out DerivedSources.pro from Tools/MiniBrowser. Even with ripping it out, it would have remained for WebKitTestRunner, so there would not be much complexity reduction anyway. > > What I've done with this patch is have the source area be polluted for Symbian - copying over QtTestBrowser's useragentlist.txt file to the MiniBrowser source directory, and pointing to the .qrc in the source directory. > In all other cases, I do what was done before - copying the QtTestBrowser/useragentlist.txt and the MiniBrowser.qrc files to the build output directory, and pointing at the qrc in the build output directory. > > I've tested this on Linux, but I have not tested it on Symbian. If someone could test this on Symbian, that would be swell. I have another (simple) idea. Let's just duplicate the txt file in the MiniBrowser area (svn copy from QtTestBrowser area of course) , and completely get rid of the MiniBrowser DerivedSources for all ports. Given that it's a rarely updated text file, it's not worth the trouble IMHO.
(In reply to comment #8) > (In reply to comment #7) > > I was unable to rip out DerivedSources.pro from Tools/MiniBrowser. Even with ripping it out, it would have remained for WebKitTestRunner, so there would not be much complexity reduction anyway. > > > > What I've done with this patch is have the source area be polluted for Symbian - copying over QtTestBrowser's useragentlist.txt file to the MiniBrowser source directory, and pointing to the .qrc in the source directory. > > In all other cases, I do what was done before - copying the QtTestBrowser/useragentlist.txt and the MiniBrowser.qrc files to the build output directory, and pointing at the qrc in the build output directory. > > > > I've tested this on Linux, but I have not tested it on Symbian. If someone could test this on Symbian, that would be swell. > > I have another (simple) idea. Let's just duplicate the txt file in the MiniBrowser area (svn copy from QtTestBrowser area of course) , and completely get rid of the MiniBrowser DerivedSources for all ports. > Given that it's a rarely updated text file, it's not worth the trouble IMHO. I agree! ;)
Comment on attachment 87422 [details] Patch r- as we should try the approach first that is recommended by 3 and is described in the original bug description. Please create two copy of useragentlist.txt under source control - one copy for QtTestBorwser and a separate copy for MiniBrowser.
(In reply to comment #8) > (In reply to comment #7) > I have another (simple) idea. Let's just duplicate the txt file in the MiniBrowser area (svn copy from QtTestBrowser area of course) , and completely get rid of the MiniBrowser DerivedSources for all ports. > Given that it's a rarely updated text file, it's not worth the trouble IMHO. I'd say it's (almost) never updated, so I'd also vote for the simple approach of duplicating the file instead of introducing more complexity in the build system. Because the Symbian issues were unknown when I introduced the feature the DerivedSources approach was prefered against duplicating the file, which was my original approach.
Created attachment 87572 [details] Patch
Comment on attachment 87572 [details] Patch r=me. Thanks.
The commit-queue encountered the following flaky tests while processing attachment 87572 [details]: http/tests/navigation/ping-cookie.html bug 57613 (author: japhet@chromium.org) The commit-queue is continuing to process your patch.
Comment on attachment 87572 [details] Patch Clearing flags on attachment: 87572 Committed r82671: <http://trac.webkit.org/changeset/82671>
All reviewed patches have been landed. Closing bug.