IRC channel logs

2026-06-06.log

back to list of logs

<ArneBab>old: the context was that I am hitting on fibers sleep 0.01 not returning.
<ArneBab>mwette: missing events could be the problem. I have a ton of sleeps in the code and if one specific sleep doesn’t return, it locks up.
<mwette>Can you generate a minimal test case?
<mwette>I have not dug into that code in years.
<ArneBab>I’ll try -- may take a while, because I don’t know why it happens unreliably with strace.
<ArneBab>s/with/except with/
<mwette>ArneBab: if you are interested in timer wheel: https://blog.acolyer.org/2015/11/23/hashed-and-hierarchical-timing-wheels/
<ArneBab>that sounds interesting -- is it what epoll uses?
<mwette>fibers. I saw timer-wheel.scm in fibers, with link to https://github.com/snabbco/snabb/blob/master/src/lib/fibers/timer.lua, with link to https://www.snellman.net/blog/archive/2016-07-27-ratas-hierarchical-timer-wheel/ with link to above.
<old>ArneBab: strace is slowing down your system too much
<old>for root-cause analysis of deadlock, you want a low latency tracer
<optimal>is it possible for classes of a certain metaclass to use another slot definition class by default
<janneke>old: ty for merging #115 -- woot, we have srfi-235 combinators! \o/
<janneke>=> https://codeberg.org/guile/guile/pulls/115
<ArneBab>mwette: thank you!
<optimal>if anyone's interested, i have a bunch of hand-crafted and somewhat optimized SRFI implementation at https://codeberg.org/jgg/various/src/branch/main/srfi that i haven't found the time to upstream, if anybody's interested
<JohnCowan>optimal: Just tell Arthur Gleckler about them and he'll add them to a contrib directory on each repo on srfi.schemers.org
<old>janneke: pleasure! I haven't tried it to I trust that it works as intended :-)
<old>(other than running the tests, I have not tested srfi-235)
<old>that was quite a rush yesterday for PRs!
<rlb>...the ./check-guile srfi-235.test output is a bit unusual, i.e. a bunch of default output, including "* PASS:\n" lines with no message.