Re: [PATCH]: Fix issue 1096

[Carl Sorensen]

On 5/27/10 8:43 AM, "Graham Percival" <address@hidden> wrote:

> On Thu, May 27, 2010 at 3:27 PM, Carl Sorensen <address@hidden> wrote:
>> 
>> On 5/27/10 8:15 AM, "David Kastrup" <address@hidden> wrote:
>> 
>>> I think one should try to feed erroneous strings to the includable lexer
>>> and look at the resulting error messages.  That's probably hard to do
>>> with a regtest (since the regtest is not supposed to produce errors),
>>> but it would provide some sanity check with regard to the error
>>> mechanisms tracking the right input.
>> 
>> Would you be willing to do so?
> 
> As David said, it would be hard to do so with a normal regtest.  Now,
> it's theoretically possible to add a separate directory which produced
> regtest output (including error messages)... in some ways, I saw
> something similar to this in the lilypond-book regtest patch.
> 
> However, this involves fairly advanced screwing around with the build
> *and* GUB, which frankly is not something I would wish on my worst
> enemy.
> 
> 
> IMO, this falls into the "ridiculously too much work for the benefit"
> category.  If somebody wants to add the idea to the issue tracker,
> sure, but it gets priority-Postponed.

I saw the comment from David not as a suggestion to build an automated
process for feeding the erroneous strings, but as a one-time testing
suggestion.  Hence my request for help.

Thanks,

Carl