[Webkit-unassigned] [Bug 232933] FindGObjectIntrospection.cmake: prefix variables obtained from pkg-config with PKG_CONFIG_SYSROOT_DIR

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=232933

Adrian Perez <aperez at igalia.com> changed:

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

--- Comment #8 from Adrian Perez <aperez at igalia.com> ---
(In reply to Michael Catanzaro from comment #6)
> Comment on attachment 452238 [details]
> Patch
> 
> This looks fine to me, but Adrian mentioned something about not liking this.
> Adrian?

So what's suspicious about this is that when cross-compiling, the sysroot
refers to the location of the *target platform* files (libraries, headers,
etc.) but in the case of programs to execute during the build, we cannot
(typically, at least) run them on the *build host*. Prefixing the sysroot
is the wrong thing to do in the general case, what one does is passing
-DCMAKE_PROGRAM_PATH= to tell if where to look for programs used during
compilation and that will run in the *build host*.

Now, I do know that GI is cross-compilation trouble, and at least Yocto
and Buildroot use qemu's userspace emulation to run the tools; but how to
treat the particularities of GI tooling is the choice of each particular
build system. I would rather not have a workaround for some compilation
system in the WebKit source tree—such a think should not be needed, and
proof of that is Buildroot can builds WebKitGTK with introspection enabled
without a patch like this:

  https://git.busybox.net/buildroot/tree/package/webkitgtk/webkitgtk.mk?h=2022.02.x#n68

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20220405/afda72f4/attachment-0001.htm>