Le dimanche 09 juillet 2023 à 12:57 +0200, David Kastrup a écrit : > Ugh. Looks like our test pipeline setup would warrant including testing > about this, then.
IIRC, when "make bytecode" and "make install-bytecode" were introduced, we still considered them a bit experimental and didn't want to force everyone to use bytecode. It definitely makes sense to keep the option to build without bytecode, since that is quite faster for incremental rebuilds (especially in Guile 2.2, as opposed to Guile 3.0), but maybe we should let "make install" install bytecode by default at this point, or even mandatorily, and maybeb "make doc" should always use bytecode (I haven't measured how much time it costs in the doc build vs. how much running a slower LilyPond costs).
signature.asc
Description: This is a digitally signed message part
