bash 3 trap usage message when configuring gcc-3.3.3
christopher at bergeron.com
Thu Aug 12 16:22:21 PDT 2004
Matt - just for reference, I'm using the 5.1.1 stable book. My _host_ is
running bash-3.0. I'm not sure if it's relevant, but I just thought I'd
mention it in case it's useful.
Matthew Burgess wrote:
>On Thu, 12 Aug 2004 13:19:46 -0400
>Christopher M Bergeron <christopher at bergeron.com> wrote:
>>When trying to configure gcc-3.3.3 (pass 1 and pass 2) the last
>>message that is displayed is a "trap" usage explanation.
>>I'm using bash v3 and I think the trap syntax might be different from
>>bash v2. Has anyone else seen this message? Is it an error or
>>something I should be concerned about? I was able to build gcc (both
>>passes) successfully, but I'm wondering if I should be concerned about
>>a mis-configuration that might bite me in the arse later.
>>Anyone have any info?
>Right, I've done some tests (as have others on IRC) and this isn't a
>problem if you follow either of the books (unstable or testing) that
>have bash-3.0 in it. The problem occurs with the configure scripts in
>gcc-3.3.x tarballs. I got it with 3.3.4, you got it with 3.3.3, Moody
>(on IRC) got it with 3.3.2. Running gcc-3.4.1's configure script
>completes successfully, although that's largely because it's multi-level
>(like binutils), not all subdirectories get configured immediately.
>Running through 'make' results in a fully configured tree, and 'make'
>completes successfully :) By the way, if you're wondering what causes
>older gcc's to not configure properly there's a thread here -
>http://lists.gnu.org/archive/html/bug-bash/2004-07/msg00279.html - which
>shows it's due to stricter POSIX compliance on the part of bash-3.0.
>Thanks again for your report - at least we know the book's instructions
More information about the lfs-support