split panther
An effort to simplify building Caxton.
Currently, Caxton relies on the C++ msdfgen library, which requires using cross and adding hacky patches to one crate to get it to build for Windows. Once this project is complete to a satisfactory level, then I can use just two languages for Caxton instead of three.
Need to figure out what’s going wrong with my curve-to-point calculations now.
Code so far: https://gitlab.com/Kyarei/fdsm
barren sedge
how fast is it?
split panther
I haven’t measured that yet, but I plan to benchmark it against msdfgen when I fix the outstanding bugs and implement some other features from msdfgen
I hope I can make it faster than msdfgen, though
doh forgot to read this carefully when I was first writing the code
barren sedge
also, question: do you generate all characters at once?
split panther
I’m only generating the [A-Z] characters in a preliminary test
static hull
how is this fabric related im confused
left spindle
I'm just confused in general.
Therefore, good work!
split panther
how is this fabric related im confused
Caxton is a Fabric mod
static hull
ah ok
split panther
Running benches/msdfgen.rs (target/release/deps/msdfgen-03eb1f4200915223)
Benchmarking fdsm/fdsm noto 100: Warming up for 3.0000 s
Warning: Unable to complete 10 samples in 5.0s. You may wish to increase target time to 9.8s.
fdsm/fdsm noto 100 time: [984.23 ms 988.31 ms 993.38 ms]
thrpt: [100.67 elem/s 101.18 elem/s 101.60 elem/s]
Found 1 outliers among 10 measurements (10.00%)
1 (10.00%) high mild
Benchmarking fdsm/msdfgen noto 100: Warming up for 3.0000 s
Warning: Unable to complete 10 samples in 5.0s. You may wish to increase target time to 13.1s.
fdsm/msdfgen noto 100 time: [1.3066 s 1.3092 s 1.3115 s]
thrpt: [76.246 elem/s 76.385 elem/s 76.536 elem/s]
honest flower
big fan of caxton! thanks for making such a nice mod
split panther
Spent some time moving things to use “prepared shapes”. Also tracked down a few artifacts in the resulting distance fields.