[Date Prev][Date Next][Thread Prev][Thread Next]
- Subject: Increasing MAXSTACK implications
- From: liam mail <liam.list@...>
- Date: Sat, 31 Dec 2011 15:13:07 +0100
On Saturday, 31 December 2011, Andreas Falkenhahn <email@example.com> wrote:
> Well, never mind, thanks for the explanation. Next time I should investigate
> more closely into such issues to see what's really going wrong. But it certainly
> wouldn't have happened if there had been a word of warning above the MAXSTACK
> definition that mentioned the implications of increasing this value :-) There are
> at least two:
> 1) You can't pass more than 255 arguments to a function even if you raise
With a little work I do not see why this is not possible, for starters I would look at increasing the instruction size.
> 2) Fiddling with MAXSTACK will break compatibility with bytecode compiled
> with different MAXSTACK settings.
I think anything which changes generated bytecode should also change the official flag in the header, which would break loading official byte code without more changes. IIRC this is recommend in the source.