[Webkit-unassigned] [Bug 144561] [GTK] Re-enable Quartz backend on cmake build system

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=144561

Michael Catanzaro <mcatanzaro at igalia.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mcatanzaro at igalia.com

--- Comment #6 from Michael Catanzaro <mcatanzaro at igalia.com> ---
I think the option should be defined with WEBKIT_OPTION_DEFINE in PlatformGTK.cmake, off by default and private (invisible/unsupported). The option should have appropriate conflicts on ENABLE_WAYLAND_TARGET or ENABLE_X11_TARGET as needed.

(In reply to comment #5)
> Here you should give an error, because we want the build to produce errors
> instead of silently disabling things now.

Well, it should be an error indeed, but in PlatformGTK.cmake, not there. The find module shouldn't have anything to do with which features to enable: it should just find things. It should export a variable, say GTK_QUARTZ_FOUND, to indicate if it is available. Then in PlatformGTK.cmake, if ENABLE_QUARTZ_TARGET is set but GTK_QUARTZ_FOUND is not, the build should fail using message(FATAL_ERROR ...).

But this patch is only copying the existing format of FindGTK3.cmake, which touches both ENABLE_X11_TARGET and ENABLE_WAYLAND_TARGET currently. I didn't realize it was doing that (since it is just a find module and I didn't anticipate it having side effects). Anyway, that won't work as expected anymore since this code runs after the values of those settings have already gone into the FEATURE_DEFINES list, so you would wind up with e.g. ENABLE_X11_TARGET turned off for most of the build, but on in IDL files (which use FEATURE_DEFINES). (I think ENABLE_QUARTZ_TARGET doesn't make it into FEATURE_DEFINES at all, since it's not defined with WEBKIT_OPTION_DEFINE.) I should file another bug for that.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20150504/1d967e4b/attachment-0001.html>