On Jul 4, 2011, at 5:04 PM, Reinhold Kainhofer wrote: > Am Montag, 4. Juli 2011, 16:39:08 schrieb Neil Puttock: >> On 4 July 2011 15:31, Carl Sorensen <c_soren...@byu.edu> wrote: >>> Plus, I don't think it's really a redundant check; I think it's a real >>> check. Absent such a check, we're trusting on the *-init.ly files being >>> correct, which admits a potential programming error. >> >> The *-init.ly files are covered by regression testing since >> -dcheck-internal-types triggers an assertion error for incorrect >> context property settings. > > Is there any possibility to install those checks only after all internal init > files have been processed? >
Alternatively, there could be a lazy_internal_set_property method in context.cc that has a scheme binding accessed in the init files. To make sure regtests work, we could make it so that this is still responsive to the assert error for type checking. Cheers, MS _______________________________________________ lilypond-devel mailing list lilypond-devel@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-devel
