This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: [Patch]: Fix a windres (rcparse.y) parsing bug for controls without text fields
- From: Christopher Faylor <cgf at redhat dot com>
- To: Danny Smith <danny_r_smith_2001 at yahoo dot co dot nz>
- Cc: binutils <binutils at sources dot redhat dot com>
- Date: Fri, 27 Jun 2003 11:27:01 -0400
- Subject: Re: [Patch]: Fix a windres (rcparse.y) parsing bug for controls without text fields
- References: <20030627060749.6931.qmail@web21405.mail.yahoo.com>
On Fri, Jun 27, 2003 at 04:07:49PM +1000, Danny Smith wrote:
>Okay, here is a revised patch that avoids code duplication (and, in fact,
>make it easier to separate out some more inconsistencies between rc and
>windres independently -- later. Thanks Chris).
Wow, it's much more invasive than the last patch but I guess that's to
be expected. My yacc grammar parsing skills are slightly rusty but
this looks fine to me. Please check in.
>On further testing with MS rc.exe I am finding that, in some cases, the
>leading text field in control definition statements is not optional with
>MS rc.exe but obligatory. That is, gnu windres (before or after my proposed
>patch) accepts syntax for which rc.exe reports syntax errors. Should I
>incorporate these additional changes (to make make windres fail or warn) now
>or do that separately since those changes are orthogonal to what *this patch
>does?
I agree with Ian on this one. I see no reason why windres can't be more flexible
than rc. So, I don't think that we need to be error compliant if there is a
reasonable default behavior.
cgf