Bug 81844 – Actually move WTF files to their new home

• Component: New Bugs
• Status: RESOLVED FIXED

[Eric Seidel (no email)]

Actually move WTF files to their new home

[Eric Seidel (no email)]

Sad.  It silently failed to add the patch.

[Eric Seidel (no email)]

Created attachment 133143 [details]
First attempt for the EWS bots

[Eric Seidel (no email)]

I may try creating a github branch instead of trying to use patches, as this is gonna get ridiculous real fast.

[Eric Seidel (no email)]

This is some sort of svn-apply "add" bug again.

[Adam Barth]

I wonder if you should just land the patch and then try to pick up the pieces.  Given that the header change is now landed, it might not be too bad...

[Eric Seidel (no email)]

I suspect it won't be too bad.  I'd been trying to avoid creating ire, but just telling folks when I'm going to land it is probably the better choice at this point.  Our tools are just not designed for this kind of huge move.

[Hajime Morrita]

FYI:
I tried this locally with XCode4.2 + Mac SL, I didn't build.
On WTF, NDEBUG is defined even for the debug build, that makes some symbols unavailable.
I poked around a bit, but couldn't figure out why.

[Carlos Garcia Campos]

I managed to build the gtk port with a few changes in top of the patch and patch attached to bug #81884. I'll make another patch with the changes I made.

[Carlos Garcia Campos]

Created attachment 133212 [details]
Patch to fix GTK+ build

Applied on top of eric's patch and patch attached to bug #81884, this one fixes the GTK+ build.

[Eric Seidel (no email)]

I've fixed Mac locally as well.  I'll update the patch shortly.

As announced on webkit-dev, I still plan to land this around 4PM this afternoon, and debug the buildbots from there.  If others wish to pre-flight their various ports before then, I welcome the additional patches.

[Eric Seidel (no email)]

Comment on attachment 133212 [details]
Patch to fix GTK+ build

View in context: https://bugs.webkit.org/attachment.cgi?id=133212&action=review

> Source/WTF/GNUmakefile.am:35
> +       -I$(srcdir)/Source/WTF/wtf

I would not expect this to be necessary.

[Carlos Garcia Campos]

(In reply to comment #11)
> (From update of attachment 133212 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=133212&action=review
> 
> > Source/WTF/GNUmakefile.am:35
> > +       -I$(srcdir)/Source/WTF/wtf
> 
> I would not expect this to be necessary.

I don't remember exactly where, but there were headers included without the wtf/ inside WTF.

[Eric Seidel (no email)]

Comment on attachment 133212 [details]
Patch to fix GTK+ build

View in context: https://bugs.webkit.org/attachment.cgi?id=133212&action=review

>>> Source/WTF/GNUmakefile.am:35
>>> +       -I$(srcdir)/Source/WTF/wtf
>> 
>> I would not expect this to be necessary.
> 
> I don't remember exactly where, but there were headers included without the wtf/ inside WTF.

Ah, my bad.  Yes, this seems totally necessary for *building* wtf.  I was thinking it was another project trying to include wtf headers.  wtf cpp files can just use "foo.h" and expect it to work.

[Carlos Garcia Campos]

(In reply to comment #13)
> (From update of attachment 133212 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=133212&action=review
> 
> >>> Source/WTF/GNUmakefile.am:35
> >>> +       -I$(srcdir)/Source/WTF/wtf
> >> 
> >> I would not expect this to be necessary.
> > 
> > I don't remember exactly where, but there were headers included without the wtf/ inside WTF.
> 
> Ah, my bad.  Yes, this seems totally necessary for *building* wtf.  I was thinking it was another project trying to include wtf headers.  wtf cpp files can just use "foo.h" and expect it to work.

exactly :-) wtf_cppflags is only used by WTF itself.

[Raphael Kubo da Costa (:rakuco)]

(In reply to comment #10)
> If others wish to pre-flight their various ports before then, I welcome the additional patches.

I couldn't use webkit-patch apply-from-bug to test this patch locally:

Failed to run "[u'/home/rakuco/dev/webkit/WebKit/Tools/Scripts/svn-apply', u'--force']" exit_code: 2 cwd: /home/rakuco/dev/webkit/WebKit/

Last 500 characters of output:
ictionVerifier.h
patching file Source/WTF/wtf/ThreadSafeRefCounted.h
patching file Source/WTF/wtf/ThreadSpecific.h
patching file Source/WTF/wtf/ThreadSpecificWin.cpp
patching file Source/WTF/wtf/Threading.cpp
patching file Source/WTF/wtf/Threading.h
patch: **** Only garbage was found in the patch input.
fatal: pathspec 'Source/WTF/wtf/ThreadingNone.cpp' did not match any files
Failed to git add Source/WTF/wtf/ThreadingNone.cpp. at /home/rakuco/dev/webkit/WebKit/Tools/Scripts/svn-apply line 444.

Failed to run "[u'/home/rakuco/dev/webkit/WebKit/Tools/Scripts/svn-apply', u'--force']" exit_code: 2 cwd: /home/rakuco/dev/webkit/WebKit/

Any hints?

[Csaba Osztrogonác]

(In reply to comment #4)
> This is some sort of svn-apply "add" bug again.

If you remove these line, it will be applieable:
diff --git a/Source/JavaScriptCore/wtf/ThreadingNone.cpp b/Source/JavaScriptCore/wtf/ThreadingNone.cpp
deleted file mode 100644
index e69de29..0000000

[Csaba Osztrogonác]

*** Bug 79783 has been marked as a duplicate of this bug. ***

[Csaba Osztrogonác]

Created attachment 133367 [details]
Qt buildfix for "First attempt for the EWS bots"

[Csaba Osztrogonác]

Created attachment 133369 [details]
Qt buildfix with "First attempt for the EWS bots" for EWS

Additionally I removed the entry I mentioned to make svn-apply happy.

[Csaba Osztrogonác]

I missed to test Qt-WK2 build, I'm on it now.

[Eric Seidel (no email)]

I attempted to incorporate these changes in my local copy.  Landing now.  I will be around fixing bots until they're all building.

I appreciate all of your help in this effort!

[Eric Seidel (no email)]

Committed r111778: <http://trac.webkit.org/changeset/111778>

(In reply to comment #22)
> Committed r111778: <http://trac.webkit.org/changeset/111778>

This appears to have broken the Snow Leopard and Lion builds.

[Eric Seidel (no email)]

Yup.  Investigating.

Committed a build fix in <http://trac.webkit.org/r111782>.

[Csaba Osztrogonác]

Qt buildfixes landed in:
http://trac.webkit.org/changeset/111780
http://trac.webkit.org/changeset/111783

[Eric Seidel (no email)]

I don't know how to fix EFL/Cmake.  It looks like at least JavaScriptCore/CMakeLists.txt needs to remove the ADD_SUBDIRECTORY(wtf) call.  I'm not sure if the root CMakeList is already building Source/WTF and if it knows to build it before building JavaScriptCore.

[Raphael Kubo da Costa (:rakuco)]

(In reply to comment #27)
> I don't know how to fix EFL/Cmake.  It looks like at least JavaScriptCore/CMakeLists.txt needs to remove the ADD_SUBDIRECTORY(wtf) call.  I'm not sure if the root CMakeList is already building Source/WTF and if it knows to build it before building JavaScriptCore.

Hopefully fixed with http://trac.webkit.org/r111799