Summary: | [Qt] Remove the definition of WTF_CHANGES guards from the build system | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Laszlo Gombos <laszlo.gombos> | ||||||||
Component: | Platform | Assignee: | QtWebKit Unassigned <webkit-qt-unassigned> | ||||||||
Status: | CLOSED FIXED | ||||||||||
Severity: | Normal | CC: | commit-queue, hausmann, kent.hansen | ||||||||
Priority: | P2 | Keywords: | Qt | ||||||||
Version: | 528+ (Nightly build) | ||||||||||
Hardware: | PC | ||||||||||
OS: | All | ||||||||||
Bug Depends on: | |||||||||||
Bug Blocks: | 35784 | ||||||||||
Attachments: |
|
Description
Laszlo Gombos
2009-11-19 08:56:57 PST
Created attachment 43510 [details]
proposed patch
Created attachment 43511 [details]
fix typo - s/sett/set
Comment on attachment 43511 [details]
fix typo - s/sett/set
It's good to get rid of all the duplication. But this is a step in the wrong direction as far as what Platform.h is supposed to be. It's clearly wrong to define a platform-independent thing such as WTF_CHANGES in there.
Comment on attachment 43511 [details]
fix typo - s/sett/set
Cancel review based on Darin's comment. I might make this into a QtWebkit port-only patch.
As I was working on this, noticed that when JSC is compiled for QtWebkit config.h is coming from WebCore and config.h from JavaScriptCore does not get used to compile the JSC part. This seems wrong and I will try to address it in my next patch.
I believe that is an inherent problem from compiling WebCore and JavaScriptCore at the same time into one shared library. A while ago we had the two separated for a while, with JSC built as static library (PIC) and linked into QtWebKit. That however created the problem that QtWebKit's .prl file included a link reference to libJavaScriptCore.a, which made it impossible to link any application against QtWebKit as it always tried to drag in libJavaScriptCore.a. Now with the recently introduced LIBS_PRIVATE feature in qmake we _might_ now be able to try that earlier approach again. I believe however that the right thing to do - also with the other ports in mind - is to move towards building JavaScriptCore as a shared library. See also bug #27551 for some work in that area. If you agree about the latter approach, then we should probably make this bug depend on #27551. Thanks Simon for the comments - making this bug dependent on 27551. In fact bug 27551 would probably solve the problem mentioned in the title entirely, but we can use this bug to clean up WebCore.pro (as it is proposed in the patch) after 27551 is fixed. in theory we could do some if the clean-up right now, but I consider that risky. Created attachment 51771 [details]
proposed patch
Removing dependency, as in ToT they're built separately now at least. Comment on attachment 51771 [details] proposed patch Clearing flags on attachment: 51771 Committed r56700: <http://trac.webkit.org/changeset/56700> All reviewed patches have been landed. Closing bug. Revision r56700 cherry-picked into qtwebkit-2.0 with commit 8f84e1115edba4a317c3a0d955fa94ba1824171d |