This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: [Patch, RFC] avoid potential recursion, collect error and warning summaries.
- From: Alan Modra <amodra at gmail dot com>
- To: Iain Sandoe <developer at sandoe-acoustics dot co dot uk>
- Cc: binutils Development <binutils at sourceware dot org>
- Date: Tue, 21 Feb 2012 22:21:13 +1030
- Subject: Re: [Patch, RFC] avoid potential recursion, collect error and warning summaries.
- Authentication-results: mr.google.com; spf=pass (google.com: domain of amodra@gmail.com designates 10.68.211.202 as permitted sender) smtp.mail=amodra@gmail.com; dkim=pass header.i=amodra@gmail.com
- References: <6F81E34D-1E55-4450-BB03-03C733E61DDF@sandoe-acoustics.co.uk>
On Mon, Feb 20, 2012 at 09:41:03PM +0000, Iain Sandoe wrote:
> Unfortunately, it can't be lightly applied - since the reporting of
> error summary causes a lot of test-cases to need amendment. Thus
> I've guarded that section of the code, and ports can enable it when
> they have time to go through their test-cases and amend as
> necessary.
I don't like this. There are quite a lot of generic gas tests that
fail if you define TC_WANTS_ERROR_SUMMARY. So they will at least need
modifying, and likely duplicating, to suit two possible gas error
behaviours.
If you really think this summary output is desirable, then you will
think it worth spending your time fixing the gas testsuite. On all
targets. Anything less is just pushing the job onto Nick, me, or
some other global maintainer since most targets don't really have
active maintainers.
I'm more inclined to recognize we currently have some dead code in
write_object_file and just delete it.
--
Alan Modra
Australia Development Lab, IBM