This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: [PATCH 2/2] x86: replace Intel syntax parser
- From: "Jan Beulich" <JBeulich at novell dot com>
- To: "H.J. Lu" <hjl dot tools at gmail dot com>
- Cc: <binutils at sourceware dot org>
- Date: Wed, 03 Jun 2009 11:05:00 +0100
- Subject: Re: [PATCH 2/2] x86: replace Intel syntax parser
- References: <49E708CE.76EA.0078.0@novell.com> <6dc9ffc80905300614p1c25412bw13bd6e4a2c56eaf6@mail.gmail.com>
Thanks for the quick fix for the problem. Jan
>>> "H.J. Lu" <hjl.tools@gmail.com> 30.05.09 15:14 >>>
On Thu, Apr 16, 2009 at 1:30 AM, Jan Beulich <jbeulich@novell.com> wrote:
> This patch removes the old, ad-hoc (and slightly incomplete) Intel
> syntax parser that was usable only for instruction operands, and
> replaces it with one hooked up to the normal expression parser, thus
> now usable also in expressions used in directives.
>
> There's one particular aspect where masm would reject a construct gas
> now accepts - this is recorded as two "#XXX?"-commented cases in the
> intelbad.[sl] test: masm apparently treats [] as having a higher
> precedence than the unary offset operator, something not easily doable
> in gas' expression handling, which assumes that all unary operators
> have a higher precedence than all binary operators. This is not viewed
> as a problem, however, as the meaning of these constructs is
> unambiguous and would never occur in ported over sources.
>
This patch caused:
http://sourceware.org/bugzilla/show_bug.cgi?id=10198
--
H.J.