Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Fri, 7 Oct 2016 13:30:52 +0200
From: Jakub Narębski <jnareb@...il.com>
To: Ramsay Jones <ramsay@...sayjones.plus.com>,
 Ævar Arnfjörð Bjarmason <avarab@...il.com>,
 Johannes Schindelin <Johannes.Schindelin@....de>, git@...r.kernel.org
Cc: Rich Felker <dalias@...c.org>, Jeff King <peff@...f.net>,
 musl@...ts.openwall.com
Subject: Re: Regression: git no longer works with musl libc's regex impl

W dniu 07.10.2016 o 00:42, Ramsay Jones pisze: 
> On 06/10/16 20:18, Ævar Arnfjörð Bjarmason wrote:
[...]
>> But just to clarify, does anyone have any objection to making our
>> configure.ac compile a C program to check for this sort of thing?
>> Because that seems like the easiest solution to this class of problem.
> 
> Err, you do know that we already do that, right?
> 
> [see commit a1e3b669 ("autoconf: don't use platform regex if it lacks REG_STARTEND", 17-08-2010)]
> 
> In fact, if you run the auto tools on cygwin, you get a different setting
> for NO_REGEX than via config.mak.uname. Which is why I don't run configure
> on cygwin. :-D
> 
> [The issue is exposed by t7008-grep-binary.sh, where the cygwin native
> regex library matches '.' in a pattern with the NUL character. ie the
> test_expect_failure test passes.]

Huh.  So we have NO_REGEX support in ./configure, and people using
Git on untypical architectures and systems *can* make use of it.

It was just described wrongly, so in turn to have the more neutral
description, the same as in Makefile, let's do this:

-------- >8 ---------- >8 ------------- >8 ---------- >8 ----------
Subject: [PATCH] configure.ac: Improve description of NO_REGEX test

The commit 2f8952250a changed description of NO_REGEX build config
variable to be more neutral, and actually say that it is about
support for REG_STARTEND.  Change description in configure.ac to
be the same.

Change also the test message and variable name to match.  The test
just checks that REG_STARTEND is #defined.

Issue-found-by: Ramsay Jones <ramsay@...sayjones.plus.com>
Signed-off-by: Jakub Narębski <jnareb@...il.com>
---
 configure.ac | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/configure.ac b/configure.ac
index aa9c91d..7f39fd0 100644
--- a/configure.ac
+++ b/configure.ac
@@ -835,9 +835,10 @@ AC_CHECK_TYPE([struct addrinfo],[
 ])
 GIT_CONF_SUBST([NO_IPV6])
 #
-# Define NO_REGEX if you have no or inferior regex support in your C library.
-AC_CACHE_CHECK([whether the platform regex can handle null bytes],
- [ac_cv_c_excellent_regex], [
+# Define NO_REGEX if your C library lacks regex support with REG_STARTEND
+# feature.
+AC_CACHE_CHECK([whether the platform regex supports REG_STARTEND],
+ [ac_cv_c_regex_with_reg_startend], [
 AC_EGREP_CPP(yippeeyeswehaveit,
 	AC_LANG_PROGRAM([AC_INCLUDES_DEFAULT
 #include <regex.h>
@@ -846,10 +847,10 @@ AC_EGREP_CPP(yippeeyeswehaveit,
 yippeeyeswehaveit
 #endif
 ]),
-	[ac_cv_c_excellent_regex=yes],
-	[ac_cv_c_excellent_regex=no])
+	[ac_cv_c_regex_with_reg_startend=yes],
+	[ac_cv_c_regex_with_reg_startend=no])
 ])
-if test $ac_cv_c_excellent_regex = yes; then
+if test $ac_cv_c_regex_with_reg_startend = yes; then
 	NO_REGEX=
 else
 	NO_REGEX=YesPlease
-- 
2.10.0



Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.