Pre-cache PLY lex table and use optimized mode (5% build time improvement)

Pre-cache PLY lex table and use optimized mode (5% build time improvement)

PLY lex table creation is (obviously) expensive.
Pre-caching this in a separate build step improves build time.

This is much less of an improvement than the parse table,
as the lexical grammar is much simpler than the phrase grammar,
but:
* it's a quick win, and
* will allow us to use Python's optimized mode (hopefully more win).

Improves build time (user time) by 5% on my Linux box:
59s => 56s

Followup to:
Pre-cache PLY yacc parse table and use optimized mode (45% build time improvement)
https://codereview.chromium.org/184233005/

BUG=341748
R=haraken

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=168630

[Nils Barth (inactive), 2014-03-05 09:25:52]

Another quick win.

After this I'll try Python's optimized mode
(and executing the bytecode directly)
and see if there's any more easy meat on this bone,
otherwise I'll call it a day, since it's already rather fast.

[Nils Barth (inactive), 2014-03-06 07:55:11]

*ping*

[haraken, 2014-03-06 07:58:40]

Sorry for the review delay. LGTM.

https://codereview.chromium.org/183013013/diff/10001/Source/bindings/scripts/...
File Source/bindings/scripts/blink_idl_lexer.py (right):

https://codereview.chromium.org/183013013/diff/10001/Source/bindings/scripts/...
Source/bindings/scripts/blink_idl_lexer.py:88: def __init__(self, debug=False,
outputdir=None, optimize=True):

Do we need to distinguish |debug| and |optimize| ?

[Nils Barth (inactive), 2014-03-06 08:09:46]

https://codereview.chromium.org/183013013/diff/10001/Source/bindings/scripts/...
File Source/bindings/scripts/blink_idl_lexer.py (right):

https://codereview.chromium.org/183013013/diff/10001/Source/bindings/scripts/...
Source/bindings/scripts/blink_idl_lexer.py:88: def __init__(self, debug=False,
outputdir=None, optimize=True):
On 2014/03/06 07:58:41, haraken wrote:
> Do we need to distinguish |debug| and |optimize| ?

Yes, in general: these 3 are possible: FT, FF, TF
optimize reads/writes the lex table, and skips error checking;
debug turns on logging.
You may want to turn off optimize but have no debug if you don't want
any extra files created (no lextab.py file or debug file).
Also, this makes it consistent with the underlying lex library
(see the lex.lex line below).

[Nils Barth (inactive), 2014-03-06 08:10:05]

On 2014/03/06 07:58:40, haraken wrote:
> Sorry for the review delay.

n/p -- enough backup so no hurry ;)

[Nils Barth (inactive), 2014-03-06 11:12:58]

The CQ bit was checked by nbarth@chromium.org

[commit-bot: I haz the power, 2014-03-06 11:13:14]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/nbarth@chromium.org/183013013/30001

[commit-bot: I haz the power, 2014-03-06 12:14:12]

Change committed as 168630

[Nils Barth (inactive), 2014-03-07 09:29:58]

FWIW, while this is a pretty minor improvement on Linux, it's apparently a
significant improvement on Windows (>30%)!