[Webkit-unassigned] [Bug 129992] [GTK][stable] JS engine completely broken on ia64

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Wed Apr 9 13:16:50 PDT 2014


https://bugs.webkit.org/show_bug.cgi?id=129992


Émeric MASCHINO <emeric.maschino at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
 Attachment #226259|0                           |1
        is obsolete|                            |




--- Comment #2 from Émeric MASCHINO <emeric.maschino at gmail.com>  2014-04-09 13:17:10 PST ---
Created an attachment (id=228977)
 --> (https://bugs.webkit.org/attachment.cgi?id=228977&action=review)
Second attempt at porting WebKit 1.8.1 01-ia64-wide-ptr.patch to WebKit 2.2.5

This new patch fixes the compilation problems of the initial patch port. As a quick test, MiniBrowser now starts up flawlessly, some URLs can even be browsed, others simply don't work (nothing happens when hitting the Enter key). As I'm not a WebKit developer, it's entirely plausible that I'm doing a lot of things wrong, but without more help from the WebKit community, it would be hard for me to go further. Indeed, in some places of my patch, it's really tries/guesses...

Currently, the impacted files are:
- API/APICast.h
- heap/CopiedBlock.h
- heap/Region.h
- interpreter/CallFrame.h
- interpreter/CallFrameInlines.h
- interpreter/Register.h
- llint/LLIntData.cpp
- llint/LLIntOfflineAsmConfig.h
- llint/LowLevelInterpreter.asm
- llint/LowLevelInterpreter.cpp
- runtime/JSCJSValue.h
- runtime/JSCJSValueInlines.h
- runtime/JSGlobalObjectFunctions.cpp
- runtime/JSWrapperObject.h
- wtf/Platform.h

Since the basic idea of this patch is to introduce another JSVALUE64W value type alongside JSVALUE32_64 and JSVALUE64, does this filelist sound OK or are they obvious missing participants?

One big difference between WebKit 1.8 and 2.x is the low-level interpreter that is used for C Loops when JIT is disabled (this is the case here for ia64). So I'm wondering e.g. if I have to mimic the JSVALUE64W changes at the llint level as I seem to decipher a close pattern between the structures in JSCJSValues.h and LLIntData.cpp for the JSVALUE32_64 and JSVALUE64 cases. Or should I adapt the JSVALUE64W-defined structures in JSCJSValues.h to use the JSVALUE64 path in LLIntData.cpp?

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.


More information about the webkit-unassigned mailing list