Bug 133918 - DFG AI and constant folder should be able to precisely prune MultiGetByOffset/MultiPutByOffset even if the base structure abstract value is not a singleton
Summary: DFG AI and constant folder should be able to precisely prune MultiGetByOffset...
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: JavaScriptCore (show other bugs)
Version: 528+ (Nightly build)
Hardware: All All
: P2 Normal
Assignee: Filip Pizlo
URL:
Keywords:
Depends on: 133935 133947 133985
Blocks: 133229
  Show dependency treegraph
 
Reported: 2014-06-14 14:41 PDT by Filip Pizlo
Modified: 2014-06-18 21:35 PDT (History)
7 users (show)

See Also:


Attachments
work in progress (13.47 KB, patch)
2014-06-14 14:42 PDT, Filip Pizlo
no flags Details | Formatted Diff | Diff
more (19.17 KB, patch)
2014-06-16 21:05 PDT, Filip Pizlo
no flags Details | Formatted Diff | Diff
rebased (21.75 KB, patch)
2014-06-17 20:21 PDT, Filip Pizlo
no flags Details | Formatted Diff | Diff
it compiles! (28.96 KB, patch)
2014-06-17 21:06 PDT, Filip Pizlo
no flags Details | Formatted Diff | Diff
almost done (61.91 KB, patch)
2014-06-18 13:29 PDT, Filip Pizlo
no flags Details | Formatted Diff | Diff
ready for review (61.86 KB, patch)
2014-06-18 13:43 PDT, Filip Pizlo
mhahnenberg: review+
Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Filip Pizlo 2014-06-14 14:41:59 PDT
...
Comment 1 Filip Pizlo 2014-06-14 14:42:47 PDT
Created attachment 233125 [details]
work in progress
Comment 2 Filip Pizlo 2014-06-16 21:05:18 PDT
Created attachment 233214 [details]
more
Comment 3 Filip Pizlo 2014-06-17 20:21:25 PDT
Created attachment 233279 [details]
rebased
Comment 4 Filip Pizlo 2014-06-17 21:06:11 PDT
Created attachment 233283 [details]
it compiles!

And fails all of the interesting tests.
Comment 5 Filip Pizlo 2014-06-18 13:29:30 PDT
Created attachment 233321 [details]
almost done
Comment 6 Filip Pizlo 2014-06-18 13:43:24 PDT
Created attachment 233322 [details]
ready for review
Comment 7 Mark Hahnenberg 2014-06-18 14:25:29 PDT
Comment on attachment 233322 [details]
ready for review

View in context: https://bugs.webkit.org/attachment.cgi?id=233322&action=review

r=me

> Source/JavaScriptCore/dfg/DFGConstantFoldingPhase.cpp:130
> +                    != node->transition()->next)

Weird line break.
Comment 8 Filip Pizlo 2014-06-18 21:35:59 PDT
Landed in http://trac.webkit.org/changeset/170141