Re: too many gotos

[Akim Demaille]

>>> "Paul" == Paul Eggert <address@hidden> writes:

 > Akim Demaille <address@hidden> writes:
 >> Unless someone knows a means to generate a grammar with arbitrarily
 >> many gotos.  That would allow us to test the output.

 > Don't know offhand, but that patch seems quite reasonable as far as it
 > goes.  A couple of things, though.  It should use size_t rather than
 > unsigned int (since the quantities in question in general might
 > overrun 32-bit counters on a 64-bit host).

Well, are you serious?  Can you possibly imagine a (real life) grammar
exceeding our new boundary?  Anyway, I agree your change doesn't cost
much, so it's fine with me.

 > Also, GOTO_NUMBER_MAXIMUM should be revamped and/or removed
 > accordingly.

Thanks for this part.  I completely missed it.

 > I installed the following.  Of course this is not at all a complete
 > fix for the "grammar is too large for 16 (or 32) bits" problem, but at
 > least it's a step in the right direction.

I'm not sure there remain many arbitrary limits.

Could you also update the NEWS file?