On Friday 13 January 2006 11:21, Erik de Castro Lopo wrote:

> My argument was:
> 
>    1) Either do the tests

This requires extensive changes that make the opcodes slower,
and potentially introduce new bugs just when we are about to
have a release.

> or
> 
>    2) Make the tests unnecesary by disallowing inputs which
>       trigger this segfault.

This would break some existing orchestra code that relies on
replacing tables while there are active instruments. Moving the
table is not a problem as long as none of the active instruments
use that particular table that is being replaced.