Mailing-List:	contact cygwin-apps-help AT sourceware DOT cygnus DOT com; run by ezmlm
list-help:	<mailto:cygwin-apps-help AT sourceware DOT cygnus DOT com>
list-post:	<mailto:cygwin-apps AT sourceware DOT cygnus DOT com>
Sender:	cygwin-apps-owner AT sourceware DOT cygnus DOT com
Delivered-To:	mailing list cygwin-apps AT sourceware DOT cygnus DOT com
Date:	26 May 2000 13:13:29 +0000
Discarded-X400-MTS-Extensions:	(43) (12) (2) (135) (115) (5) (6) (3)
From:	"Michael Ring" <Michael DOT Ring AT t-mobil DOT de>
To:	"earnie_boyd AT yahoo DOT com" <earnie_boyd AT yahoo DOT com>
Cc:	"cygwin-apps AT sourceware DOT cygnus DOT com" <cygwin-apps AT sourceware DOT cygnus DOT com>
Subject:	Re: Additional Patch for bash
Importance:	normal
Autoforwarded:	FALSE
Message-Id:	<C1668IMC39JUE*@MHS>
Original-Encoded-Information-Types:	(1) (0) (10021) (7) (1) (0) (6), (1) (0) (10021) (7) (1) (0) (1)

>> The problem that is fixed by my patch is as follows:
>>	
>> bash-distribution uses an old config.guess which creates results like:
>>	
>> i686-pc-cygwin32	
>>	
>> and then later checks for *cygwin32 in the configure script.
>>	
>> recent versions of config.guess return	
>>	
>> i686-pc-cygwin   (!! no 32 at the end !!)
>>	
>> and, as a result, configure does not react correctly and a bash build on base
>> of
>> this configure stackdumps.
>>	
>> This behaviour is reproduceable with either the original bash-2.04.tar.gz or
>> the
>> bash-2.0.4-src.tar.gz in the LATEST directory on sourceware.
>>	
>> Greetings,
>>	
>> Michael Ring
>
>But, I didn't/don't have your problem.  So, why should we do a patch.  The real
>patch is to replace the config.guess, config.sub in bash after unarchiving the
>source.
>
>Regards,
>

You are right in one point, the patch should also include config.guess and
config.sub to make sense; because if you apply the patch without exchanging
config.guess then bash will also stackdump.

Here's the important line from configure:

-*-cygwin32*)   opt_bash_malloc=no ;;   # Cygnus's CYGWIN32 environment
+*-cygwin*)     opt_bash_malloc=no ;;   # Cygnus's CYGWIN32 environment 

After exchanging config.guess opt_bash_malloc will not be set to no ==>
Stackdump during initialisation of bash.

Did you build bash inside the cygwin sourcetree or standalone ?

I built it standalone; perhaps there is a difference.

Greetings again,

Michael Ring

- Raw text -