freenode/#sbcl - IRC Chatlog
Search
19:07:37
stassats
is it possible for coerce on a complex non-displaced vector just return the underlying data vector?
20:14:06
stassats
i guess checking for displaced, no fill-pointer is tedious, especially if no one is going to use it
20:19:46
stassats
hmm, the div->mul doesn't appear to be good at cutting down the constant on narrow types
20:23:41
|3b|
ACTION is wondering if it is automated enough to add for example a curl command or similar, that could poke a web service and tell it to start some CI tests on github or whatever
20:24:20
|3b|
ACTION has no idea if such services exist, or how long it would take me to get around to writing one if not though
20:25:30
|3b|
possibly. if nothing else that would be easier to check externally than scraping the mailing list :)
21:32:00
|3b|
yeah, looks like tagging RCs would probably be best option, since github can be configured to poke a web service if it ever exists, and until then it is easy to scrape
21:52:52
stassats
gcc recodes power-o-two+1 multiplication as a shift and an add, but it doesn't seem to be faster here
22:51:00
fiddlerwoaroof
Krystof: that sounds interesting, are you saying that the spaces in src/runtime/validate.c:103-105 don't need to have fixed addresses?