| Summary: | Add tests for WebIDL type conversions | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | WebKit | Reporter: | Erik Arvidsson <arv> | ||||||||
| Component: | Tools / Tests | Assignee: | Joshua Bell <jsbell> | ||||||||
| Status: | RESOLVED FIXED | ||||||||||
| Severity: | Normal | CC: | abarth, jsbell, roger_fong, sam, webkit.review.bot | ||||||||
| Priority: | P2 | ||||||||||
| Version: | 528+ (Nightly build) | ||||||||||
| Hardware: | Unspecified | ||||||||||
| OS: | Unspecified | ||||||||||
| Attachments: |
|
||||||||||
This seems like a great idea to me. We could also add them to Internals if that's easier. Created attachment 181113 [details]
Patch
First stab at tests. Can land this, then extend it with EnforceRange support and fix long long conversion issues over in bug 96798 Comment on attachment 181113 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=181113&action=review I was imagining having a sub object on window.internals so that we don't pollute it with a bunch of properties that most people do not care about. > LayoutTests/fast/js/webidl-type-mapping.html:3 > +<script src="script-tests/webidl-type-mapping.js"></script> Just inline the js here Comment on attachment 181113 [details] Patch Attachment 181113 [details] did not pass mac-ews (mac): Output: http://queues.webkit.org/results/15627729 New failing tests: fast/js/webidl-type-mapping.html Created attachment 181198 [details]
Patch
(In reply to comment #5) > (From update of attachment 181113 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=181113&action=review > > I was imagining having a sub object on window.internals so that we don't pollute it with a bunch of properties that most people do not care about. Done. > > LayoutTests/fast/js/webidl-type-mapping.html:3 > > +<script src="script-tests/webidl-type-mapping.js"></script> > > Just inline the js here Done. (In reply to comment #6) > (From update of attachment 181113 [details]) > Attachment 181113 [details] did not pass mac-ews (mac): > Output: http://queues.webkit.org/results/15627729 > > New failing tests: > fast/js/webidl-type-mapping.html Apparently we're not consistent. I'm trying to get a WebKit build together to see what the difference is between JSC and V8. Comment on attachment 181198 [details]
Patch
LGTM... Please get a reviewer to r+ this.
Created attachment 181243 [details]
Patch for landing
(In reply to comment #8) > > Apparently we're not consistent. I'm trying to get a WebKit build together to see what the difference is between JSC and V8. +/-Inf/NaN and even Number.MAX_VALUE conversions were different for [unsigned] long long in a handful of cases. So far as I can tell from the WebIDL spec those should all yield 0. For now I just disabled those tests. Comment on attachment 181243 [details] Patch for landing Rejecting attachment 181243 [details] from commit-queue. New failing tests: inspector-protocol/debugger-terminate-dedicated-worker-while-paused.html Full output: http://queues.webkit.org/results/15666201 Comment on attachment 181243 [details] Patch for landing Clearing flags on attachment: 181243 Committed r138836: <http://trac.webkit.org/changeset/138836> All reviewed patches have been landed. Closing bug. http://trac.webkit.org/changeset/138838 broke the windows build http://build.webkit.org/builders/Apple%20Win%20Release%20%28Build%29/builds/41920 reopening until build is fixed (In reply to comment #16) > reopening until build is fixed Landed what I think is the fix in: http://trac.webkit.org/changeset/138864 |
Maybe we could add an object to the testRunner whose sole purpose is to test the type conversions required by WebIDL interface TestWebIDLConversions { long attribute testLong; unsigned long attribute testUnsignedLong; [EnforceRange] long attribute testEnforceRangeLong; ... }