Changeset 97240 in webkit
- Timestamp:
- Oct 12, 2011 1:01:11 AM (13 years ago)
- Location:
- trunk/Source/JavaScriptCore
- Files:
-
- 2 edited
Legend:
- Unmodified
- Added
- Removed
-
trunk/Source/JavaScriptCore/ChangeLog
r97235 r97240 1 2011-10-12 Filip Pizlo <fpizlo@apple.com> 2 3 Layout tests crashing in DFG JIT code 4 https://bugs.webkit.org/show_bug.cgi?id=69897 5 6 Reviewed by Gavin Barraclough. 7 8 Abstract value filtration didn't take into account cases where a structure 9 set filter, combined with predicted type knowledge, could lead to a stronger 10 filter for the structure abstract value. 11 12 This bug would have been benign in release builds; it would have just meant 13 that the analysis was less precise and some optimization opportunities would 14 be missed. I have an ASSERT that is meant to catch such cases, and it was 15 triggering sporadically in one of the LayoutTests. 16 17 * dfg/DFGAbstractValue.h: 18 (JSC::DFG::AbstractValue::filter): 19 1 20 2011-10-11 Gavin Barraclough <baraclough@apple.com> 2 21 -
trunk/Source/JavaScriptCore/dfg/DFGAbstractValue.h
r97218 r97240 403 403 m_type &= other.predictionFromStructures(); 404 404 m_structure.filter(other); 405 406 // It's possible that prior to the above two statements we had (Foo, TOP), where 407 // Foo is a PredictedType that is disjoint with the passed StructureSet. In that 408 // case, we will now have (None, [someStructure]). In general, we need to make 409 // sure that new information gleaned from the PredictedType needs to be fed back 410 // into the information gleaned from the StructureSet. 411 m_structure.filter(m_type); 405 412 } 406 413
Note: See TracChangeset
for help on using the changeset viewer.