Re: r150960 changed ltmain.sh and broke the build

[Dave Korn]

Ralf Wildenhues wrote:
> [ trimmed Cc:, added autoconf-patches; followups can remove gcc@ ]
> [ http://thread.gmane.org/gmane.comp.gcc.devel/108348 ]
> 
> * Dave Korn wrote on Wed, Aug 26, 2009 at 08:16:11PM CEST:
>> Ralf Wildenhues wrote:
>>>      Patterns should not include the separator (unless escaped), even
>>>      as part of a character class.  In conformance with Posix, the Cray
>>>      `sed' rejects `s/[^/]*$//': use `s,[^/]*$,,'.
>>> [...]
>>>      Portable `sed' regular expressions should use `\' only to escape
>>>      characters in the string `$()*.0123456789[\^n{}'.
>>   Ah, I didn't read those two conditions as applying simultaneously.
> 
> I don't think they apply simultaneously in the way you interpreted that.
> However, in the  s,x\{1\,\},y,  example, the comma is both a separator
> and another metacharacter.

  That seems strange to me, are you saying the comma is a metachar because it
is a component of the curly-brace number-of-repetitions modifier?

>  Hmm, maybe those sentences don't really cover this case.
> 
> How about
> 
>      Patterns should not include the separator (unless escaped), even
>      as part of a character class.  In conformance with Posix, the Cray
>      `sed' rejects `s/[^/]*$//': use `s,[^/]*$,,'.  Even escaped,
>      patterns should not include separators that are also used as
>      metacharacters.  For example, GNU sed 3.02 rejects `s,x\{1\,\},,',
>      and is used on MinGW.

  Well, modulo my surprise that a comma is considered a metachar, that makes
sense.

    cheers,
      DaveK