[Webkit-unassigned] [Bug 37945] CMake buildsystem

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=37945





--- Comment #31 from Patrick R. Gansterer <paroga at paroga.com>  2010-05-13 11:20:15 PST ---
(In reply to comment #30)
> Guys, we partially agree with you, but it's hard to say someone is 100% right in this case. Some remarks on our side:
>   - JIT's CPU detection in general CMakeLists.txt does make sense, at least so far. So with Patrick here as well.
> 
>   - DerivedSources does not exist in source, just in build tree, so no option to ADD_DIRECTORY(). If you mean to have it in root or inside cmake/, then I disagree as it will be confusing. So we'd rather stay with these port-independent CMakeList.txt
I startet to port the WebCore CMakeLists.txt to Windows. I moved most of you stuff from WebKitGenerators into the CMakeLists.txt. My idea behind a additional DerivedSources.cmake file was to remove the generators into a own file (one per library). This should be included per INCLUDE() in the CMakeLists.txt.

>   - FeatureFOO.cmake is doubtful, as usually the shared bit is a single CPP definition. The dependencies will likely differ a lot, some will not even have dependencies and it is just a feature to reduce code size.  Leandro's WEBKIT_FEATURE() handles that nicely, ports can define it. If you want we could add more logic there to allow ports to define the ${DEFAULT_${FEATURE}} and maybe add a keyword to block it, in this case WEBKIT_FEATURE() could be moved to common files, with these values being defined in the port -- but we rather keep with the simpler yet efficient solution by Leandro.
I think that this FeatureXXX files mostly define only one cpp too. So I'm not sure if this will be the best solution IMHO. I would prefere the feature dedection/definition in the PortXXX.cmake.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.