Herwin
Herwin
Added a few extra lines to `m.c` ```c if (z) { // allocate from a chain of free blocks jt->mfree[-PMINL + 1 + blockx].pool = AFCHAIN(z); // remove & use...
That makes more sense than the `epsilon` how I read it
That first one looks very much as what I've experienced/written down in #183, I've tried writing down some ideas on how to fix this, but I haven't gotten around to...
Just to save the next person some time: this code is currently in `jsrc/array.hpp`. To add a bit of documentation: the current implementation (without the typecase) is: ```c++ Bnum +...
After reading some more code: the definition of `num` and `refactorme_num` is: ```c++ #define num(n) ((A)(Bnum + 2 + (n)-NUMMIN)) [[nodiscard]] inline auto refactorme_num(int64_t n) { return reinterpret_cast(Bnum + 2...
Refactors I would suggest: * In `make_scalar_integer`, replace the check `xor_replicate_sign(k)
Yeah, I messed up `AD` and `A` in that one. I think an issue here is the values are generated by a macro: ```c #define CBAIVAL(t, v) \ { 7...
I guess this case is done, that function no longer exists. The case does include some scribbles of documentation that may be worth saving, but I guess the wiki would...
Well, the more you learn, the more you realize the thing you're trying to do won't work. The `AD` struct has a number of fields, this refactor relates mainly to...
There is some analysis of the current code in #87, which describes `num` as well.