Re: Despite text in gnu bash manual, quote removal appears to be performed on case pattern

[Stahlman Family]

----- Original Message ----- 
From: "Eric Blake" <ebb9@byu.net>
To: "Stahlman Family" <brettstahlman@comcast.net>
Cc: <bug-bash@gnu.org>
Sent: Sunday, December 10, 2006 3:40 PM
Subject: Re: Despite text in gnu bash manual, quote removal appears to be 
performed on case pattern


> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> According to Stahlman Family on 12/10/2006 8:11 AM:
> > Mingw (Msys) Bash 2.04.0(1)-release
>
> Consider upgrading.  Bash is now at 3.2, with a number of bugs fixed in
> the meantime.

I'd like to, but I think the Msys maintainers perform some mingw-specific tweaks on the cygwin sources before building, and the Msys 
distribution is still at 2.04. I have a newer Cygwin version installed (3.00) but rarely use it anymore because the version control 
and build toolchain we use at work uses an older version of the cygwin1.dll, which means I get the infamous shared memory error if I 
try to run them both simultaneously.
> > $ case $dbg in
> > > " foo bar ") echo "1" ;;
> > > \ foo\ bar\ ) echo "2" ;;
> > > esac
> > 1
>
> It appears you have found a bug in the POSIX specification.  POSIX states:
> http://www.opengroup.org/onlinepubs/009695399/utilities/xcu_chap02.html#tag_02_09_04
>
> "In order from the beginning to the end of the case statement, each
> pattern that labels a compound-list shall be subjected to tilde expansion,
> parameter expansion, command substitution, and arithmetic expansion, and
> the result of these expansions shall be compared against the expansion of
> word"
>
> but mentions quote removal for the word immediately after for.  But
> traditional behavior has always performed quote removal on the patterns,
> and changing bash would break many existing scripts.

Understood. I actually prefer the existing behavior, since there are certainly times when you would wish to disable the glob special 
meanings of characters in a pattern. Also, it makes sense that case patterns should be processed the same as are pathname expansion 
patterns.

Thanks,
   Brett S.
> - --
> Life is short - so eat dessert first!
>
> Eric Blake             ebb9@byu.net
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.5 (Cygwin)
> Comment: Public key at home.comcast.net/~ericblake/eblake.gpg
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
>
> iD8DBQFFfH7q84KuGfSFAYARAoRsAJ4lCGpncu/8BWPbJHR98scdYRHc9wCguJpb
> yI5sOsI+iGQJD0XEXpPUSBw=
> =St9Q
> -----END PGP SIGNATURE-----