Re: (tuplet . around) causes regtests to fail to compile

[Carl Sorensen]

On 10/28/10 4:52 PM, "Carl Sorensen" <address@hidden> wrote:
> On 10/28/10 4:44 PM, "Carl Sorensen" <address@hidden> wrote:
>> On 10/28/10 4:39 PM, "Werner LEMBERG" <address@hidden> wrote:
>>>> I don't think it's a Heisenburg.  The make test fails for me, too.
>>> 
>>> Me too...  Very, very strange.  Anybody out here who knows enough
>>> about the Scheme internals to debug it?
>> 
>> I'm currently seeing if I can reproduce it  by a single command-line
>> command.
>> 
> 
> OK, I've got a command that appears to reproduce the problem
> 
> lilypond -dbackend=eps --formats=ps,png,pdf  -dinclude-eps-fonts \
> -dgs-load-fonts --header=doctitle --header=doctitlede \
> --header=doctitlees --header=doctitlefr --header=doctitlehu \
> --header=doctitleit --header=doctitleja --header=doctitlenl \
> --header=texidoc --header=texidocde --header=texidoces \
> --header=texidocfr --header=texidochu --header=texidocit \
> --header=texidocja --header=texidocnl -dcheck-internal-types \
> -ddump-signatures -danti-alias-factor=2 \
> --formats=eps  --verbose  -deps-box-padding=3.000000 \
> lily-4bb72c87.ly

OK, I've narrowed the error down to the presence of the switch

-dcheck-internal-types

If this switch is omitted, the snippet compiles cleanly.

Also, I can generate the error by

lilypond -dcheck-internal-types input/regression/slur-tuplet.ly

It appears that there is some internal error in LilyPond that is causing
-dcheck-internal-types not to work properly.

I'm afraid this is over my head, although I'll look around a bit more at it.

Thanks,

Carl