Alex Shinn

Sorry, I don't follow your argument that a `syntax-rules` transformer is "dumb" and therefore can't work with unhygienic macros, much less how syntactic closures differs from syntax-case here. Regarding `define-record-type`....

> > Sorry, I don't follow your argument that a `syntax-rules` transformer is "dumb" and therefore can't work with unhygienic macros, [...] > > What I meant by "dumb" is...

Sorry, you guys seem to be having fun but I'm afraid I'm too busy join in just now :sweat_smile:

Source: https://arxiv.org/abs/2112.08926 It was actually a place holder until I could derive something myself but I may just leave it as is.

There is also some interesting discussion in: https://stackoverflow.com/questions/1659440/32-bit-to-16-bit-floating-point-conversion. Note one answer is just to convert the 3 float components separately, capping the exponent and rounding/truncating the mantissa depending on direction,...

The rounding I'm not too concerned about, but I'll fix the nan handling.

Sorry, you were asking how to test this and I didn't reply. The code is exposed in the C API, but from Scheme you have to go indirectly via uniform...

Note we could consider some rewriting options to remove bos, eos, bol, eol, bow, eow, nwb in cases. As a contrived example, consider a regex to extract "key: value" headers,...

Sorry, lost track of this. Yes, actually the bounds are currently only happening at the storage level, so can fail for shared arrays and non-specialized arrays. I can add these...

Sorry, the top link should be http://synthcode.com/scheme/chibi/ with a trailing slash "/".