gnupic: Re: [gnupic] upgrading gpasm/gplink warnings to errors
Subject:
Re: [gnupic] upgrading gpasm/gplink warnings to errors
From:
Craig Franklin ####@####.####
Date:
11 Feb 2005 05:07:38 +0000
Message-Id: <420C3E26.3030502@users.sourceforge.net>
David McNab wrote:
> Hi,
>
> (thanks Craig for earlier fix of 'undetected branch out of range' bug).
>
No problem.
> Is there any way to force gpasm/gplink to generate errors, as opposed
> to warnings for 'branch out of range' conditions?
>
I can't think of one.
I suppose for absolute code you could write a macro. That won't help
gplink.
> Or, would it be within reason to request a change to gpasm/gplink so
> this condition generates an error?
>
I could add a switch to error on all warnings. This would work well for
gplink. gpasm might be a problem. You would have to use the error
level directive to suppress the common warnings.
> If I've got an out-of-range branch, I'd really prefer for gpasm/gplink
> to die with an error, so as to crash the 'make'. Branches out of range
> are a very serious condition - IMHO, reacting with a 'warning' might
> be taking the issue a bit too lightly.
>
The severity is assigned by Microchip. I would like to keep the default
behavior that way. There is no reason I can't add a -Werror switch to
both tools, or something like it.
> Thoughts?
>