These functions are implemented directly in assembly, so they need the proper directives to enable/disable visibility. On ELF systems, it's .hidden, whereas on Mach-O systems (Mac) it's .private_extern. On Windows, it's not necessary since you have to explicitly export. I also implemented the AIX idiom, though it's unlikely anyone will implement AIX/POWER JIT.
Created attachment 42033 [details]
FYI, svn-apply (and thus the commit-queue) will fail to handle this patch until bug 30683 is landed (which will hopefully be soon).
This looks sane to me, but Gavin or Oliver are really your best people to review this.
I'm confused that the ChangeLog says Thiago Macieira, and yet Jocelyn Turcotte attached the patch. Neither of you have committer powers AFAIK, so I would add this to the commit-queue, except that I fear the ChangeLog may list the wrong person?
(In reply to comment #4)
> I'm confused that the ChangeLog says Thiago Macieira, and yet Jocelyn Turcotte
> attached the patch. Neither of you have committer powers AFAIK, so I would add
> this to the commit-queue, except that I fear the ChangeLog may list the wrong
Thiago is a Qt product manager and developer. He has a lot of experience of making Qt compile on all of our supported platforms. As Qt distributes WebKit as part of it's source code, Thiago probably did the patch for Qt and not trunk WebKit.
As many of the Qt developers is not working directly on WebKit, but still make fixes as part of Qt (think documentation as well), Jocelyn is trying to upstream those patches and has thus rebased them to trunk WebKit and added them to Bugzilla.
I think this explains the situation.
Comment on attachment 42033 [details]
Clearing flags on attachment: 42033
Committed r51141: <http://trac.webkit.org/changeset/51141>
All reviewed patches have been landed. Closing bug.