Issue 11438042: Allow lazy compilation (and thus optimisation) of functions inside eval.

Issue 11438042: Allow lazy compilation (and thus optimisation) of functions inside eval. (Closed)

Created:
8 years ago by rossberg

Modified:
8 years ago

Reviewers:
Michael Starzinger

CC:
v8-dev

Base URL:
https://v8.googlecode.com/svn/branches/bleeding_edge

Visibility:
Public.

More Reviews

Description

Allow lazy compilation (and thus optimisation) of functions inside eval. For strict-mode eval, this requires _disabling_ lazy parsing of inner functions, because we need to collect their free variables to do allocation for the eval scope properly. R=mstarzinger@chromium.org BUG=v8:2315 Committed: https://code.google.com/p/v8/source/detail?r=13161

Patch Set 1 #

Patch Set 2 : Forgot regression test #

Patch Set 3 : Fixed copyright date #

Total comments: 6

Patch Set 4 : Addressed comments #

Created: 8 years ago

Download [raw] [tar.bz2]

	Unified diffs	Side-by-side diffs	Delta from patch set	Stats (+74 lines, -32 lines)			Patch
M	src/contexts.h	View		1 chunk	+5 lines, -1 line	0 comments	Download
M	src/lithium.h	View		1 chunk	+2 lines, -2 lines	0 comments	Download
M	src/parser.cc	View	1 2 3	3 chunks	+9 lines, -5 lines	0 comments	Download
M	src/preparser.h	View		1 chunk	+2 lines, -2 lines	0 comments	Download
M	src/scopes.cc	View		1 chunk	+3 lines, -8 lines	0 comments	Download
A +	test/mjsunit/regress/regress-2315.js	View	1 2 3	1 chunk	+6 lines, -7 lines	0 comments	Download
M	test/mjsunit/regress/regress-492.js	View	1 2 3	2 chunks	+39 lines, -1 line	0 comments	Download
M	test/mjsunit/regress/regress-crbug-135066.js	View	1 2 3	1 chunk	+8 lines, -6 lines	0 comments	Download

Messages

Total messages: 3 (0 generated)

Expand Messages | Collapse Messages

Michael Starzinger

LGTM (with micro-nits). I also verified that it doesn't tank performance on Octane's code load ...

8 years ago (2012-12-07 09:50:17 UTC) #2

rossberg

8 years ago (2012-12-07 10:23:33 UTC) #3

https://codereview.chromium.org/11438042/diff/8001/src/parser.cc
File src/parser.cc (right):

https://codereview.chromium.org/11438042/diff/8001/src/parser.cc#newcode638
src/parser.cc:638: mode = PARSE_EAGERLY;
On 2012/12/07 09:50:17, Michael Starzinger wrote:
> Curly braces around body.

Done.

https://codereview.chromium.org/11438042/diff/8001/test/mjsunit/regress/regre...
File test/mjsunit/regress/regress-2315.js (right):

https://codereview.chromium.org/11438042/diff/8001/test/mjsunit/regress/regre...
test/mjsunit/regress/regress-2315.js:34: for (var i = 0; i < 10; ++i) foo();
On 2012/12/07 09:50:17, Michael Starzinger wrote:
> Is there a particular reason we need two separate loops here?

Now using %OptimizeFunctionOnNextCall instead.

https://codereview.chromium.org/11438042/diff/8001/test/mjsunit/regress/regre...
File test/mjsunit/regress/regress-492.js (right):

https://codereview.chromium.org/11438042/diff/8001/test/mjsunit/regress/regre...
test/mjsunit/regress/regress-492.js:90: for (var i = 0; i < 10; ++i) foo();
On 2012/12/07 09:50:17, Michael Starzinger wrote:
> Is there a particular reason we need two loops here?

Removed.

Expand Messages | Collapse Messages