Fix test
43f1f886a92f — Chris Cannam 2 months ago
Use a plainer, more functional string-list format (i.e. don't divide it onto multiple lines, it's too often confusing). If you want something nicer, arrange it yourself
82b6a5a4fc8d — Chris Cannam 6 months ago
Add SV
0f0a1bbf1710 — Chris Cannam 10 months ago
Fix api usage
9973aaad085d — Chris Cannam 10 months ago
Experiment with using Ryu as an alternative
Add real32 ffi as well
Remove leftover debugging line
Deps files shouldn't be committed, I don't think
Correct strfromd import: it returns a C int, not the default ML int (which may be Int64.int)
Docs
We aren't going to use this
Warning should not be fatal (I only did that for test purposes); remove old implementation. New one might be slower than the way the old one did it, but I haven't been able to measure a difference
One more test
Overhaul to use Qt-style %1 instead of just %; this allows for %n. Incompatible in dangerous ways with the former behaviour, I'm afraid. Big review about to begin.
Add R32
Fix the ultimate corner case: printing -INT_MAX
Add required header (for linux)
Add a strfromd implementation that can be used on platforms that lack it (we can't reliably call varargs functions like snprintf through ffi, e.g. on ARM Mac it fails)
Factor out tests so they can be run from another program too if desired
Add option option
Next