line numbers in error messages

<20021213140143.GA25983@kiwi.intellektik.informatik.tu-darmstadt.de>
<3DFDC921.4EAF899D@icparc.ic.ac.uk>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.4i
X-MailScanner: Found to be clean
Sender: owner-eclipse-users@icparc.ic.ac.uk
Precedence: bulk

First of all: Thank you, Joachim, for your help with external C predicates
under Windows and Linux.  A first test seems to work.

Now for something completely different:  

I write ECLiPSe programs with emacs and test/run them in a terminal (I don't
like constant clicking).  Although all works well, I'd like to report some
small trifles.  Please don't get me wrong: I'm not complaining!

Sometimes, the error messages are not very helpful.  Eg. the messages
 
WARNING: Unrecognized field name(s) in 'literal with [spolarity : pos, id : clear, parameters :
[[1]]]' in module eclipse.

and 

WARNING: predicate used but not declared or defined in xxx / 0 in module eclipse

would be improved by giving the line number (and in the second case, the file)
of the first reference to the unknown field name/predicate.  Sometimes, it is
very hard to find the mistake, eg if you have predicates with same name but
different arity.


Furthermore

*** Warning: Singleton variables in clause 3 of sh_get_mapped_variable/3: X

would be improved by a line number, too, maybe the line number of the head of
the clause: In emacs, it is much easier to jump to a line number than to the
n-th clause.


emacs.el has a bug, too: Sometimes, the command eclipse-comment-region
comments more than the region between the mark and the current point.  If this
happens at the end of a file, this problem can even freeze emacs - it tries to
comment out an unlimited number of lines.  In general, it would be better to
comment a marked region instead of commenting up to the point.  An example is
the comment command of auctex.


As I said, just some small remarks,

Thanks, Uli
-- 
Ulrich Scholz

scholz@informatik.tu-darmstadt.de
http://www.intellektik.informatik.tu-darmstadt.de/~scholz