Issue 21002: _sre.SRE_Scanner object should have a fullmatch() method (original) (raw)

Created on 2014-03-21 01:31 by Gareth.Gouldstone, last changed 2022-04-11 14:58 by admin. This issue is now closed.

Messages (7)
msg214317 - (view) Author: Gareth Gouldstone (Gareth.Gouldstone) Date: 2014-03-21 01:31
I believe that the SRE_Scanner object should have a .fullmatch() method for consistency with other re pattern-matching behaviour. >>> rex = re.compile('([^\\W\\d_]{1,2}[0-9]{1,2}[^\\d\\W_]?)[ \\t]*([0-9][^\\d\\W_]{2})') >>> rex.scanner('bn20bs') <_sre.SRE_Scanner object at 0x102006400> >>> rex.scanner('bn20bs').search() <_sre.SRE_Match object; span=(0, 6), match='bn20bs'> >>> rex.scanner('bn20bs').match() <_sre.SRE_Match object; span=(0, 6), match='bn20bs'> >>> rex.scanner('bn20bs').fullmatch() Traceback (most recent call last): File "", line 1, in AttributeError: '_sre.SRE_Scanner' object has no attribute 'fullmatch'
msg214320 - (view) Author: Gareth Gouldstone (Gareth.Gouldstone) Date: 2014-03-21 01:45
I also encountered issue 20998, which explains the convoluted [^\W\d_] in place of [a-zA-Z] as I was investigating why case-insensitivity and quantifiers would not work together.
msg214323 - (view) Author: R. David Murray (r.david.murray) * (Python committer) Date: 2014-03-21 03:09
Scanner isn't a public interface, so the real enhancement here would be to make it one, in which case adding fullmatch would make sense. I don't know if making it public is a good idea or not.
msg214337 - (view) Author: Gareth Gouldstone (Gareth.Gouldstone) Date: 2014-03-21 08:54
Scanner may not be a public interface but it is widely documented, not least on page 67 of the O'reilly Python Cookbook. Even if Scanner is not made public, then surely it should maintain consistency with the public interfaces?
msg355468 - (view) Author: Lewis Gaul (LewisGaul) * Date: 2019-10-27 14:26
Scanner is not documented in CPython's own re documentation: https://docs.python.org/3/library/re.html#regular-expression-objects, so I'd say there's no obligation to maintain consistency with public APIs. Is anyone able to make a call on whether this issue should be closed, or alternatively give some guidance on what work this issue should encompass?
msg355489 - (view) Author: Kyle Stanley (aeros) * (Python committer) Date: 2019-10-27 18:17
> Is anyone able to make a call on whether this issue should be closed, or alternatively give some guidance on what work this issue should encompass? Added Serhiy to the nosy list, since he's an active maintainer for the re module.
msg355500 - (view) Author: Serhiy Storchaka (serhiy.storchaka) * (Python committer) Date: 2019-10-27 19:56
I concur with David. The SRE_Scanner object is not public. Its search() method is used for implementing finditer(). Its match() method could be useful for implementing a functional equivalent of \G in Perl. But I do not know a use case for its fullmatch() method. If anybody provide a use case for SRE_Scanner.fullmatch() this issue can be reopened.
History
Date User Action Args
2022-04-11 14:58:00 admin set github: 65201
2019-10-27 19:56:21 serhiy.storchaka set status: open -> closedresolution: rejectedmessages: + stage: resolved
2019-10-27 18:17:00 aeros set nosy: + aerosmessages: +
2019-10-27 18:14:33 aeros set nosy: + serhiy.storchaka
2019-10-27 14:26:02 LewisGaul set nosy: + LewisGaulmessages: +
2014-03-21 08:54:46 Gareth.Gouldstone set messages: + versions: + Python 3.4
2014-03-21 03:09:03 r.david.murray set versions: + Python 3.5, - Python 3.4nosy: + r.david.murray, twoutersmessages: + type: behavior -> enhancement
2014-03-21 01:45:26 Gareth.Gouldstone set messages: +
2014-03-21 01:31:17 Gareth.Gouldstone create