Re: [PATCH 1/2] gdb: improve error reporting from expression parser

[John Baldwin <jhb@FreeBSD.org>]

On 12/23/23 11:56 AM, Andrew Burgess wrote:
> This commits changes how errors are reported from the expression
> parser.  Previously, parser errors were reported like this:
> 
>    (gdb) p a1 +}= 432
>    A syntax error in expression, near `}= 432'.
>    (gdb) p a1 +
>    A syntax error in expression, near `'.
> 
> The first case is fine, a user can figure out what's going wrong, but
> the second case is a little confusing; as the error occurred at the
> end of the expression GDB just reports the empty string to the user.
> 
> After this commit the errors are now reported like this:
> 
>    (gdb) p a1 +}= 432
>    A syntax error in expression near '<HERE>' marker in: `a1 +<HERE>}= 432'.
>    (gdb) p a1 +
>    A syntax error in expression near '<HERE>' marker in: `a1 +<HERE>'.
> 
> Now GDB reports the entire expression along with a <HERE> marker to
> indicate where the error occurred.
> 
> I did consider trying to have multi-line errors here, in the style
> that gcc produces, with some kind of '~~~~~^' marker on the second
> line to indicate where the error occurred; but I rejected this due to
> the places in GDB where we catch an error and repackage the message
> within some longer string, I don't think multi-line error messages
> would work well in that case.
> 
> Adding the '<HERE>' marker was actually a later extension I made.  My
> first implementation simply tried to address the empty string
> case (where GDB reports "..., near `'").  Originally I just changed
> this to be:
> 
>    (gdb) p a1 +
>    A syntax error in expression, at the end of `a1 +'.
> 
> But then I thought adding the '<HERE>' markers was pretty neat, so I
> did that instead.  However, if folk feel the '<HERE>' marker is more
> confusing than helpful I can always fall back to my original plan, and
> just fix the empty string case.
> 
> I originally wanted to try and style the '<HERE>' marker, I thought
> that would help make it clear that it wasn't part of the expression.
> However, GDB's error function doesn't support styling right now.  We
> could possibly add this in the future, in which case the <HERE> marker
> could be given maybe the metadata style to help it stand out.
> 
> I've updated the small number of tests that check for a syntax error,
> and add a couple of extra tests in gdb.base/exprs.exp.

Hmm, I do kind of find the <HERE> approach somewhat verbose (or at least
I feel like it's a bit shouty).  I might prefer either your simple fix
for end of string (which I think is a perfectly fine fix), or the more
complicated multi-line error similar to what GCC/clang do for compile
errors.  It does sound like the multi-line case might be too complicated
to implement.

-- 
John Baldwin