Bug 108578

Summary:

[CPP,GObject,ObjC] Add 'static' skip to CodeGenerator{CPP,GObject,ObjC}.pm

Product:

WebKit

Reporter:

Nils Barth <nbarth>

Component:

WebKit Misc.

Assignee:

Nobody <webkit-unassigned>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

abarth, haraken, japhet, ojan.autocc, webkit.review.bot

Priority:

Version:

528+ (Nightly build)

Hardware:

Unspecified

OS:

Unspecified

Attachments:

Description	Flags
Patch	none
ChangeLog wording fix	none

Nils Barth

Reported 2013-01-31 21:42:39 PST

Legacy code generators do not support static attributes. Fix by adding a test to skip static types. This allows us to not use macros like "#if defined(TESTING_JS) || defined(TESTING_V8)" in IDL files. Concretely it's just adding a suitable check to the SkipAttribute functions, and then removing macros from the test IDL file.

Attachments
Patch (5.00 KB, patch) 2013-01-31 23:24 PST, Nils Barth	no flags	Details Formatted Diff Diff
ChangeLog wording fix (5.01 KB, patch) 2013-01-31 23:38 PST, Nils Barth	no flags	Details Formatted Diff Diff
Show Obsolete (1) View All Add attachment proposed patch, testcase, etc.

Nils Barth

Comment 1 2013-01-31 23:24:50 PST

Created attachment 185953 [details] Patch V. simple fix -- just check $attribute->isStatic This lets us remove macros from TestObj.idl and also TestSupplemental.idl

Kentaro Hara

Comment 2 2013-01-31 23:28:10 PST

Comment on attachment 185953 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=185953&action=review > Source/WebCore/ChangeLog:8 > + Since legacy code generators (CodeGenerator{CPP,GObject,ObjC}.pm) Don't call them legacy. Their implementations are just behind.

Nils Barth

Comment 3 2013-01-31 23:30:33 PST

(In reply to comment #2) > (From update of attachment 185953 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=185953&action=review > > > Source/WebCore/ChangeLog:8 > > + Since legacy code generators (CodeGenerator{CPP,GObject,ObjC}.pm) > > Don't call them legacy. Their implementations are just behind. Oops, sorry! Is there a proper name, like "lagging code generators", or should I just call them something like "CPP/GObject/ObjC"?

Kentaro Hara

Comment 4 2013-01-31 23:31:40 PST

(In reply to comment #3) > or should I just call them something like "CPP/GObject/ObjC"? You can just say CPP/GObject/ObjC. We shouldn't break these bindings just like we shouldn't break JSC/V8.

Nils Barth

Comment 5 2013-01-31 23:32:47 PST

(In reply to comment #4) > (In reply to comment #3) > > or should I just call them something like "CPP/GObject/ObjC"? > > You can just say CPP/GObject/ObjC. We shouldn't break these bindings just like we shouldn't break JSC/V8. Ok, will do.

Nils Barth

Comment 6 2013-01-31 23:38:03 PST

Created attachment 185956 [details] ChangeLog wording fix Fix wording re: CPP/GObject/ObjC in ChangeLog

Kentaro Hara

Comment 7 2013-01-31 23:38:38 PST

Comment on attachment 185956 [details] ChangeLog wording fix LGTM

WebKit Review Bot

Comment 8 2013-02-01 00:45:57 PST

Comment on attachment 185956 [details] ChangeLog wording fix Clearing flags on attachment: 185956 Committed r141551: <http://trac.webkit.org/changeset/141551>

WebKit Review Bot

Comment 9 2013-02-01 00:46:01 PST

All reviewed patches have been landed. Closing bug.

Note You need to log in before you can comment on or make changes to this bug.