IRC channel logs
2025-04-24.log
back to list of logs
<lisbeths>how can I get wisp like python indentation in scheme <old>lisbeths: Wisp is new in 3.0.10 (the next version of Guile) <old>what distro are you on? <old>any idea why the profiler is pointing at ice-9/popen.scm:183:0:reap-pipes <old>Nobody is calling open-pipe in the program. I use spawn <civodul>old: it’s called at every GC, due to its guardian IIRC <old>right. But it seems that the profielr thinks this is the bottleneck <old>but it's actually another hook <old>I use fibers, and it seems that the real bottleneck after GC is blue/fibers/events.scm:165:2:pump-epoll-guardian <old>(I've removed the reap-pipes hook) <old>and I get the exact same usage but now in fibers hook <civodul>old: interesting; i’m not sure how reliable this is <civodul>i mean, it mostly shows that your program is creating garbage <old>civodul: right. The next most hot procedure is bytevector->pointer. I've ported fibers to pure Guile so I can embedded it in a pure Guile project. I wonder if this is a symptom of generating garbage in the FFI <old>anyway, does not seem like it's a bottleneck afterall <civodul>the FFI is often a source of garbage <civodul>ports.test: keyword arguments for file openers <old>civodul: I guess it's less performant and not compatible with Guile 2 <old>But I only want a async scheduler for calling external tasks for my build system <old>so I don't really care <civodul>rlb: does the build failure above ring a bell? ↑ <ieure>Still wondering about #78004, anyone have an explanation of what's going on with that? Am I doing something wrong? <old>ieure: what are you refering to? <rlb>sneek: later tell civodul, looks like there's also an rdelim.test failure a bit further down, and don't remember anything offhand.