James Bailey <derhindemith <at> googlemail.com> writes:
> Okay, in just scanning the file, I was a bit concerned about that.
> I think actually a smaller example would be
> better. If you can pare this down just a bit, I'll be happy to add
> it to the tracker.
Of course a smaller example would be nice, but I already spent hours getting it
down to the point it's at, and as I explained earlier in the thread I'm not
able
to get it any smaller and still have it exhibit the problem.
This focus on small examples is misbegotten, in my opinion. I am not a LilyPond
developer but I speak from a lot of compiler experience when I say it's more
important to get coverage from lots of problems than to insist that every bug
be
a three-line test case. I recommend making it easy to report bugs so you can
build a robust piece of software, rather than throwing up obstacles to people
who
would like to help by showing where something has gone wrong. Again speaking
from
my experience, the difference between a 40- or 50-line test case and a 3-line
test case does not have a big effect on the amount of effort required to
diagnose
the bug.
As I said before, I'm perfectly willing to follow up with another test case for
this bug that's a couple of lines longer but is perfectly valid input. I am
unfortunately not willing to spend any more time trying to make this smaller.
As
I said, I already spent hours. I'm afraid you'll need to take it or leave it.
I hope you'll take it. Thanks!
-- Robert
_______________________________________________
bug-lilypond mailing list
bug-lilypond@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-lilypond
