19:57:11Krystofin full worse-is-better spirit, though, I'll be happy with something that helps us break less of the world every month, even if it's not the perfect interface
19:59:01Krystof_death: I don't think we know what a supportable assembler interface looks like
19:59:12Krystofif we did, we would probably have implemented it internally already
19:59:44Krystofthe fact that we change assembler internals to make things clearer strongly suggests that we haven't got it right yet
20:05:38_deathyeah, and if it's not for a particular platform, it can be a major undertaking that may not succeed anyway.. but if nibbles needs some way to remove bounds checking and basic fetch/store that may be useful for other things
20:09:02_deathjust thinking aloud (sorry if it's noise, coming from someone who is not an sbcl dev)
21:11:12stassatsnot breaking existing third party tools would take probably as much, or less, effort as creating a new contrib
23:43:14|3b|Krystof: possibly, though i don't know sbcl internals so it would mostly just be "copy the VOPs out of nibbles". it also might take a while since i have a few other projects ahead of it :/