Bug 106058 - DFG::ByteCodeCache serves little or no purpose ever since we decided to keep bytecode around permanently
Summary: DFG::ByteCodeCache serves little or no purpose ever since we decided to keep ...
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:
Blocks:
 
Reported: 2013-01-03 15:49 PST by Filip Pizlo
Modified: 2013-01-03 16:29 PST (History)
7 users (show)

See Also:

Attachments
the patch (15.84 KB, patch)
2013-01-03 15:50 PST, Filip Pizlo 		msaboff: review+
Details | Formatted Diff | Diff
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Filip Pizlo 2013-01-03 15:49:06 PST 
ByteCodeCache used to be great because we would by default throw away the instruction stream of code blocks at the earliest opportunity.  ByteCodeCache would allow the DFG to reparse code blocks on demand, regenerating the instruction stream if necessary.

But now we always keep the instruction stream around.  Hence, all we really need to be doing is asking the executable for the baseline code block.  The bytecode parser already does this, and its insistence on distinguishing between the code block and the profiled block is unnecessary for inlinees.
Comment 1 Filip Pizlo 2013-01-03 15:50:36 PST 
Created attachment 181241 [details]
the patch
Comment 2 Filip Pizlo 2013-01-03 16:29:46 PST 
Landed in http://trac.webkit.org/changeset/138763