Title: [92732] trunk/Source/_javascript_Core
Revision
92732
Author
fpi...@apple.com
Date
2011-08-09 17:30:50 -0700 (Tue, 09 Aug 2011)

Log Message

DFG JIT passes the this argument to constructors even though
it's not necessary
https://bugs.webkit.org/show_bug.cgi?id=65943

Reviewed by Gavin Barraclough.

* dfg/DFGJITCodeGenerator.cpp:
(JSC::DFG::JITCodeGenerator::emitCall):

Modified Paths

Diff

Modified: trunk/Source/_javascript_Core/ChangeLog (92731 => 92732)


--- trunk/Source/_javascript_Core/ChangeLog	2011-08-10 00:28:48 UTC (rev 92731)
+++ trunk/Source/_javascript_Core/ChangeLog	2011-08-10 00:30:50 UTC (rev 92732)
@@ -1,3 +1,14 @@
+2011-08-09  Filip Pizlo  <fpi...@apple.com>
+
+        DFG JIT passes the this argument to constructors even though
+        it's not necessary
+        https://bugs.webkit.org/show_bug.cgi?id=65943
+
+        Reviewed by Gavin Barraclough.
+
+        * dfg/DFGJITCodeGenerator.cpp:
+        (JSC::DFG::JITCodeGenerator::emitCall):
+
 2011-08-09  Chao-ying Fu  <f...@mips.com>
 
         Fix one MIPS instruction to call JITStubThunked_##op

Modified: trunk/Source/_javascript_Core/dfg/DFGJITCodeGenerator.cpp (92731 => 92732)


--- trunk/Source/_javascript_Core/dfg/DFGJITCodeGenerator.cpp	2011-08-10 00:28:48 UTC (rev 92731)
+++ trunk/Source/_javascript_Core/dfg/DFGJITCodeGenerator.cpp	2011-08-10 00:30:50 UTC (rev 92732)
@@ -999,7 +999,7 @@
     m_jit.storePtr(MacroAssembler::TrustedImmPtr(JSValue::encode(jsNumber(numArgs))), addressOfCallData(RegisterFile::ArgumentCount));
     m_jit.storePtr(GPRInfo::callFrameRegister, addressOfCallData(RegisterFile::CallerFrame));
     
-    for (int argIdx = 0; argIdx < numArgs; argIdx++) {
+    for (int argIdx = (node.op == Call ? 0 : 1); argIdx < numArgs; argIdx++) {
         NodeIndex argNodeIndex = m_jit.graph().m_varArgChildren[node.firstChild() + 1 + argIdx];
         JSValueOperand arg(this, argNodeIndex);
         GPRReg argGPR = arg.gpr();
_______________________________________________
webkit-changes mailing list
webkit-changes@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-changes

Reply via email to