Simon Albrecht <simon.albre...@mail.de> writes: > Hi Joel, > > thanks for the report. > > On the user list, I may have exaggerated my zeal for minimal > examples. Here on the bug list, however, they are more strictly > required, so the bass clef and bar checks (which I believe are > certainly unrelated with the bug – can’t test on Linux) could’ve been > left out.
There is some motivation for this: some of the harder problems need to run the debugger in "recording" mode in order to be able to "step backwards" from the occurence of a problem. The recording mode runs several orders of magnitude slower than normal runs, and if too much code occurs between the reason of a problem and the error condition, it will not be possible to "step backwards" to the place of the problem. Now to be honest: for a problem of that kind the code sleuth is likely to apply manual reduction/changes anyway in order to arrive at the required minimalism, and that work is dwarved by the actually necessary effort for debugging when you need to crank out the heavy-duty tools. So while reports with several pages of code are rarely all that helpful, a few lines are more often than not quite sufficient as a work example even when there are still a few single elements that can be thrown out. -- David Kastrup _______________________________________________ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
