Le 2012-03-01 à 14:33:00, Jonathan Wilkes a écrit :
Oh interesting-- there's an entire thread on namespaces for send/receive. Never noticed that before.
From 2006 all I see is some stuff about deallocatable symbols-- is that what you're referring to?
That's an approximate year. There are probably several threads on the topic(s). I remember writing about deallocatable symbols in more recent years, but the thing that I think was in 2006, is about splitting the receive-table away from the symbol-table so that receive-symbols could become local : have a global t_symbol * but have a local s_thing. Then receive-symbols wouldn't necessarily be symbols anymore, they'd be pairs of one $0 and one t_symbol *... I'm reinventing this in my head as I write it, maybe.
It's possible to fit a very large $0 in a_type because most values of t_atomtype aren't taken. For example, all negative values of t_atomtype could be reserved to mean the local-symbol where $0 = -a_type.
| Mathieu BOUCHARD ----- téléphone : +1.514.383.3801 ----- Montréal, QC