Re: 2.15.8 Regtests

[Neil Puttock]

On 6 August 2011 15:31, David Kastrup <address@hidden> wrote:

> I have a hard time counting the removal of a band aid for an artificial
> test case with undefined behavior (try finding a place in the user
> documentation that declares this kind of code as producing predictable
> results) as a regression because the original code did not fix the
> underlying problem, but merely masked it.

So how would you expect the following code to behave?  It's the
snippet from the original bug report, which segfaulted in stem.cc.

\relative c' {
  \time 2/4
  \voiceOne
  s16 [g s g ] s16 [g s g ] |
  s16 [g s g ] \override Stem #'(details beamed-lengths) = #'(15 15)
  s16 [g s g ] |
  s16 [g s g ] s16 [g s g ] |
  s16 [g s g ] \revert Stem #'(details beamed-lengths) s16 [g s g ] |
  s16 [g s g ] s16 [g s g ] |
}

The regression test is deliberately artificial since it gives a clear
indication of failure, which this code doesn't (the segfault no longer
occurs due to checking the nested property is a pair before using
robust_list_ref).  I don't think it's unreasonable to expect this code
to return 'beamed-lengths to the default value defined in
define-grobs.scm.

Cheers,
Neil