Matthew Craven

Results 16 issues of Matthew Craven

As discussed at #535, `withForeignPtr` will be much slower in at least GHC-9.2.4 and GHC-9.4.1. So I made a quick pass at removing all internal uses of this function. To...

`Lazy.length s` must force the entirety of `s`, while (by design) `Lazy.compareLength s 10` stops after finding at least 11 words. And yet we try to rewrite `Lazy.length s >...

documentation

That is, good list consumer [in the sense of the GHC user guide](https://downloads.haskell.org/ghc/9.2.1/docs/html/users_guide/exts/rewrite_rules.html#list-fusion). If this can be done without a significant performance penalty when no fusion happens, I think this...

Looking at a test coverage report brought it to my attention. (Somehow this was missed in #590.)

GHC 8.4 was first released in March 2018, very nearly six years ago. I think that's certainly an adequate support window. I propose making our next release (0.12.1.0) the last...

In [this comment](https://github.com/haskell/bytestring/pull/631#discussion_r1451112873) it was noticed that we have a mismatch between the actual C prototype of `sbs_elem_index` and the foreign import in our Haskell code: ```c ptrdiff_t sbs_elem_index(const void...

bug

Since #580 we expect this to happen. Like #574 this is most naturally done with something like `tasty-inspection-testing`. Such a test was not added as part of #580 because of...

Although we've enabled `-fno-strict-aliasing` in #582, I would on principle rather we just didn't have any strict-aliasing violations in our code. But: * I expect the `bytestring` C code contains...

For example [this problem](https://github.com/haskell/bytestring/pull/582#issuecomment-1522476846) with a recent patch would have been trivial to diagnose if we enabled `-Wconversion`. But enabling a reasonable warning set would cause a lot of noise...

Problems with these have been noticed multiple times in the past. At least https://github.com/haskell/bytestring/pull/569#discussion_r1070686036 and https://github.com/haskell/bytestring/pull/526#issuecomment-1173933301 come to mind. Having tests would help us root out any remaining problems here,...