[webkit-reviews] review denied: [Bug 39243] Auto-generate Canvas overloads in JSC : [Attachment 56290] Patch
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Fri Jun 18 15:48:18 PDT 2010
Adam Barth <abarth at webkit.org> has denied Yaar Schnitman <yaar at chromium.org>'s
request for review:
Bug 39243: Auto-generate Canvas overloads in JSC
https://bugs.webkit.org/show_bug.cgi?id=39243
Attachment 56290: Patch
https://bugs.webkit.org/attachment.cgi?id=56290&action=review
------- Additional Comments from Adam Barth <abarth at webkit.org>
In general, I think this is great. However, Darin asked to review each change
separately. Can you break this up into a series of patches:
1) Rebaseline run-bindings-tests to pick up the PERFECT_HASH_SIZE change.
rs=me for landing that.
2) Update the code generator and show the changes in run-bindings-tests output.
3) One patch each for each API you're converting, along with a test that tries
calling the API with each number of arguments and showing that the behavior
either doesn't change or changes in a way that we want.
Discussion in other bugs has uncovered that WebIDL specs that you should throw
exceptions if you get a number of arguments forbidden by the IDL. It seems
this is the behavior we want for new bindings. Whether we want to adopt that
behavior in old bindings is a question that we'll tackle on a case-by-case
basis until we uncover a pattern to our decision making.
More information about the webkit-reviews
mailing list