James <[email protected]> writes: > On 23/10/13 18:12, David Kastrup wrote: >> >> Should we create a release of 2.16.3 which only differs from 2.16.2 in >> cherry-picking the autoconf change making it refuse to compile with >> buggy versions of Metapost? >> > Well personally I think we should 'warn' than refuse to compile. Those > that know how to 'go back' to an earlier version of metapost will, > those that don't will not then be stuck with something that doesn't > work. > > If you see what I mean.
I see what you mean, but we are not talking about cosmetic differences here. We are talking about fatally broken output. It's not just every treble clef that looks like something the cat dragged in, but also all the flags are missing pieces of their spine. This is not something that a warning is appropriate for. People who want to continue compiling with that setup _really_ need to hand-edit the configuration test in order to _insist_ getting a broken version of LilyPond. -- David Kastrup _______________________________________________ lilypond-devel mailing list [email protected] https://lists.gnu.org/mailman/listinfo/lilypond-devel
