Optimize parsing of regular expressions

[serhiy-storchaka]

BPO	19380
Nosy	@pitrou, @vstinner, @ezio-melotti, @serhiy-storchaka, @MojoVampire
Files	re_parse.patch re_parse_2.patch re_parse_3.patch re_parse_4.patch re_parse_5.patch

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = 'https://github.com/serhiy-storchaka'
closed_at = <Date 2014-10-10.08:46:47.543>
created_at = <Date 2013-10-24.20:14:19.696>
labels = ['expert-regex', 'library', 'performance']
title = 'Optimize parsing of regular expressions'
updated_at = <Date 2014-10-10.08:46:47.541>
user = 'https://github.com/serhiy-storchaka'

bugs.python.org fields:

activity = <Date 2014-10-10.08:46:47.541>
actor = 'serhiy.storchaka'
assignee = 'serhiy.storchaka'
closed = True
closed_date = <Date 2014-10-10.08:46:47.543>
closer = 'serhiy.storchaka'
components = ['Library (Lib)', 'Regular Expressions']
creation = <Date 2013-10-24.20:14:19.696>
creator = 'serhiy.storchaka'
dependencies = []
files = ['32341', '32343', '36649', '36818', '36843']
hgrepos = []
issue_num = 19380
keywords = ['patch', 'needs review']
message_count = 13.0
messages = ['201177', '201183', '201191', '201192', '201227', '206557', '227032', '227041', '227053', '228605', '228838', '228964', '228971']
nosy_count = 7.0
nosy_names = ['pitrou', 'vstinner', 'ezio.melotti', 'mrabarnett', 'python-dev', 'serhiy.storchaka', 'josh.r']
pr_nums = []
priority = 'normal'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'performance'
url = 'https://bugs.python.org/issue19380'
versions = ['Python 3.5']

[serhiy-storchaka]

Proposed patch optimizes parsing of regular expressions. Total time of re unittests decreased by 10%.

[pitrou]

I don't think "+=" speeds up anything for ints, you might as well minimize code churn by avoiding such changes.

[serhiy-storchaka]

Done.

[pitrou]

Do you have any benchmark figures (apart from the time of re unittests)?

[serhiy-storchaka]

### regex_compile ###
Min: 2.897919 -> 2.577488: 1.12x faster
Avg: 3.066306 -> 2.681966: 1.14x faster
Significant (t=26.77)
Stddev: 0.08789 -> 0.05085: 1.7283x smaller

[serhiy-storchaka]

Could someone please make a review?

[serhiy-storchaka]

Actually "if x:" is slightly faster than "if x is not None:" on current implementation.

[pitrou]

"is not None" is more readable, though. When using plain boolean testing, it's never obvious whether you can have a zero-length string, a null number, etc.

[serhiy-storchaka]

Well, then please look at re_parse_2.patch (it is still applied cleanly).

[serhiy-storchaka]

Here is a patch which addresses Yury's and Josh's comments. Also discarded few minor changes.

[serhiy-storchaka]

Updated patch implements Antoine's suggestions.

[python-dev]

New changeset 1adeac2a8714 by Serhiy Storchaka in branch 'default':
Issue bpo-19380: Optimized parsing of regular expressions.
https://hg.python.org/cpython/rev/1adeac2a8714

[serhiy-storchaka]

Thank you for your reviews Yury, Josh, and Antoine.