Re: lib/56024: Recent regex change breaks textproc/link-grammar

[Ryo ONODERA <ryo%tetera.org@localhost>]

The following reply was made to PR lib/56024; it has been noted by GNATS.

From: Ryo ONODERA <ryo%tetera.org@localhost>
To: Christos Zoulas <christos%zoulas.com@localhost>, gnats-bugs%netbsd.org@localhost
Cc: 
Subject: Re: lib/56024: Recent regex change breaks textproc/link-grammar
Date: Sat, 27 Feb 2021 02:40:17 +0900

 Hi,
 
 Christos Zoulas <christos%zoulas.com@localhost> writes:
 
 > What' going on here is that the package was expecting to have gnu regex that supports \w out of the box.
 > Our old regex code just ignored \<characters> that it did not understand, so it was not working correctly either; it was just ignoring the error.
 > The current regex code will work properly if you pass REG_GNU in regcomp(). This is not the default behavior.
 
 Thanks for your explanation.
 I will try to create a patch for textproc/link-grammar.
 
 > Best,
 >
 > christos
 
 -- 
 Ryo ONODERA // ryo%tetera.org@localhost
 PGP fingerprint = 82A2 DC91 76E0 A10A 8ABB  FD1B F404 27FA C7D1 15F3