From: "Tim Van Holder" <tim DOT van DOT holder AT pandora DOT be>
To: <djgpp-workers AT delorie DOT com>
Subject: Re: GCC configure buglet
Date: Tue, 27 May 2003 21:18:46 +0200
Message-ID: <002c01c32484$cbdb3650$2302a8c0@dualzastai>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook, Build 11.0.4920
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165
In-reply-to: <200305221810.h4MIA4aC018594@envy.delorie.com>
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

[seems mailing this from work doesn't reach the list - trying]
[again from home]

> > One caveat - this test (both original & patched) may break if
> > autoconf 2.13 is used to generate configure,
> 
> Assume 2.13, since that's the version being used at the moment.
> 
> > AC_PREREQ(2.13)

I do assume that - I wrote "autoconf > 2.13" but the ">"
unfortunately slipped to the start of the line, making it look
like quoted mail text.

> GCC 3.3 works fine with the 2.57. I have done allot of builds
> this week, proably in the order of 15-20 or more without any
> GCC build issues that I know of.
>
> If the configure did fail then would I be able to do a build?

It's not configure that fails, just the test I'm patching that
will misdetect the feature it's testing.
It creates a file using a shell here document that is later
compared to the output of objdump.  The last line of this file
has trailing spaces, and autoconf 2.57a at least (not sure about
2.57 or others before it) trims that trailing whitespace, so it
doesn't find a match and assumes the eh_frame optimization bug
is present when it isn't.  For DJGPP this currently means nothing
as without my patch it misdetects that anyway, but using an
autoconf other than 2.13 _may_ break the test regardless of the
patch.