freenode/#sbcl - IRC Chatlog
Search
16:48:26
flip214
the M-V-B in https://paste.debian.net/hidden/c0610c25/ seems to break the type derivation logic in some way...
16:50:33
flip214
Fact is that multiple branches that all have a (MAKE-INSTANCE 'class) (with identical class) are not combined for the return type...
16:52:15
flip214
OTOH, even without the COND I get no useful type here?! Is *DERIVE-FUNCTION-TYPES* broken since 1.5.0?
16:57:06
scymtym
even with a constant class name, the class associated to that name can change at runtime
16:58:55
flip214
and I'm thinking about _type_ derivation, not static class-slots-lifecycle-analysis or so?!
17:04:17
scymtym
i'm having trouble thinking of useful things the compiler could do given the possibility of redefinition
23:15:18
scymtym
i would like to push this bit of cleanup: https://github.com/scymtym/sbcl/commit/0cf6ae83c9372fb3c5904470f8bdae44986788d3 any objections?
23:17:59
scymtym
the whole tlf is guarded. the commit only removes (or is supposed to at least) redundant or unsatisfiable read-time conditionals
23:21:09
scymtym
this found it: https://techfak.de/~jmoringe/eclector-inconsistent-feature-expressions.png
23:22:37
Xach
a long time ago xof wrote me a reader tweak that issued style-warnings for unnecessary ::s.
23:24:51
scymtym
ACTION has to update that since it now knows about s-expression-level syntax and doesn't get the '(if …) wrong
23:24:53
stassats
i have abandoned my work on better source locations, maybe eclector can just be used instead?
23:27:19
scymtym
also, i didn't do any benchmarks, but i suspect eclector to be at least an order of magnitude slower than the current reader
23:28:20
Xach
scymtym: i would love to see something like that integrated into emacs via slime or anything really.
23:28:54
stassats
scymtym: ok, scratch the reader then, my main problem was recovering source locations
23:30:34
scymtym
stassats: maybe a bit of coordination is needed. the "trick" is something like scheme's syntax objects or beach's csts to get source locations for atoms
23:32:45
scymtym
Xach: i use the language server protocol for most of my tests. maybe slime could work as a transport as well. this shows some of the things i have been trying with eclector: https://techfak.de/~jmoringe/eclector-context-completion.ogv