<paroneayea>mudsync works again under new 8sync refactor \\o/ <davexunit>heap-allocated floats will be the death of me. <daviid>guile-gnome (simple apps) triggers goops related bugs for 2.1.4, same code that pass using 2.0.13 :( it is going to be a nightmare to produce a snipset to allow wingo to reproduce <davexunit>wingo: I have a program that could benefit from f64 specialized br-if-* operations. I have some 2D vector math mostly optimized by using bytevectors, no allocation when adding/multiplying vectors, but inevitably one wants to do comparisons like "is the x coordinate less than 0?". <davexunit>this requires boxing the float taken from the bytevector before the comparison can be done. <davexunit>my test program animates N particles that bounce off of the window edges. the GC runs once per second with 1000 particles on my machine. if I remove the bounds checking, and thus the bouncing, the GC doesn't run at all after program initialization. <amz3`>ACTION trying to install guix :/ <amz3`>wiredtiger made a new release with all the fixes I need ***karswell` is now known as karswell
<amz3`>ohoh! some of my tests are broken <paroneayea>galex-713: might depend on what you're benchmarking <paroneayea>if you use large cons based lists and try doing random access on them frequently in your program, it won't be faster :) <paroneayea>galex-713: I'm not aware of any "general" benchmarks of Guile v Python v Ruby etc <amz3`>guile 2.2 is faster than guile 2.1 <amz3`>I've been stuck for no apparent reasons <amz3`>but I've been working on guile git bindings too <amz3`>and it's missing the primary operations like git pull/checkout/clone which are not as easy to bind as I thought <amz3`>right now I try to get guix up-to-date <amz3`>what are you up to galex-713 ? ***random-nickname is now known as random-nick