Dan Eble <d...@faithful.be> writes: > On Aug 27, 2015, at 23:59 , Dan Eble <d...@faithful.be> wrote: >> In your opinion, does scm_gc_mark (smob->self_scm ()) deserve to be >> implemented as member function in Smob? > > Never mind. That seems like a slippery slope.
I remember it being a complication with the mark hook that it could be called on already swept objects in GUILE 2.0. So it mustn't go through any virtual table mechanism as it must not rely on objects not already being destructed. And of course there must be no allocations etc. That's of course also doable with member functions but the headaches might not be worth the gain. -- David Kastrup _______________________________________________ lilypond-devel mailing list lilypond-devel@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-devel