freenode/#sbcl - IRC Chatlog
Search
13:41:31
stassats`
i really want tests for optimizations, but no idea how, you can't really benchmark reliably microptimizations
13:45:37
flip214
well, _really_ small microbenchmarks should be reliably compiled to one assembler sequence, I guess.. and these should be stable enough
16:55:35
slyrus
I don't see it when I run run-tests.sh on it, but I wasn't sure if that is was because it was just getting muffled somewhere.
17:03:14
stassats`
and now i'm wasting time on sbcl-devel@ arguing with people who don't like the amount of unfixed bugs on launchpad
17:04:27
stassats`
the blunder was engaging in any sort of communication to begin with, snarky or not
17:04:51
scymtym
the problem seems to be a misunderstanding about what sbcl is ("team", "goals", "commercial support", etc.)
17:05:43
scymtym
stassats`: sorry to bother you with again but (COMPILE :NODE/LVAR :DERIVE-TYPE :MISC.2) still fails on x86: https://ci.cor-lab.org/job/sbcl-master/featureset=1,label=ubuntu_trusty_32bit/lastCompletedBuild/consoleFull
17:06:09
stassats`
scymtym: i guess it got missed, about that, should :allow-warnings t include style warnings?
17:06:42
stassats`
cause originally i had allow-style-warnings t, for both tests, but on x86-64 one of the tests failed with a warning
17:07:42
scymtym
it current doesn't so you can specific. you can specify both, though. but the fact that x86 and x86_64 signal different warnings is suspicious, isn't it?