[Webkit-unassigned] [Bug 89551] DFG should be able to print disassembly interleaved with the IR

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=89551





--- Comment #6 from Filip Pizlo <fpizlo at apple.com>  2012-06-20 10:45:02 PST ---
(In reply to comment #5)
> (From update of attachment 148519 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=148519&action=review
> 
> r=me, but please double-check dominators removal, and add a comment to the ChangeLog explaining why.
> 
> > Source/JavaScriptCore/dfg/DFGDriver.cpp:-95
> > -    dfg.m_dominators.compute(dfg);
> 
> Was this on purpose? Seems wrong.

It was, but then I forgot to finish the job.

We have a couple of things in the DFG code-base that were done as first steps to SSA but that we never used.  They should either get killed or find new life.

One is the redundant Phi elimination phase, which is just now wrong.  The code is sitting there but isn't called from anywhere.

The other is dominators.  I thought I'd use them for arguments simplification, but then I decided not to.  But, in the meantime, I found dominators to be useful for IR dumps.  They make it easier for me to figure out block relationships.  So, what this patch was supposed to do, is remove unconditional dominator computation, and instead call Dominators::computeIfNecessary() inside of Disassembler.

I'll land with that change.

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