Hi Urs,
I am the most ignorant of all when it comes to lilypond internal
processing, but I recall in the past when I have tried to put diagnostic
print statements in Scheme inside music functions that they come out of
order sometimes in unpredictable ways. I recall DK pointing out that the
function in question for that one operation gets called repeatedly as
the code is parsed and compiled and the layout engine does its passes
and print statements don't therefore work in a nice straightforward
linear order, and can be very misleading. I may have the details wrong
here, and I cant remember the specific problem I was using the function
for, but I mention this to say that debugging Scheme inside Frescobaldi
may be fraught with deep complexity. I don't know, but I would be
worried. I'd like for somebody enlightened in these matters to comment.
[I am happy for this to be dismissed as nonsense!]
Andrew
On 25/9/19 4:23 pm, Urs Liska wrote:
a question on the Frescobaldi mailing list got me thinking: is it possible (or
could it be made possible with reasonable effort) to attach a debugger to a
LilyPond compilation process?
Probably there would be quite some managing to be done because there's no
simple relation of code steps like in a programming language, but if the
LilyPond parser (or the Scheme parser) could be hooked into it would be
terrific if we could make Frescobaldi step through the input files (or probably
one would rather want to set a breakpoint, e.g. in a music function and then
step through that function to better understand what it's doing).
_______________________________________________
lilypond-user mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/lilypond-user
