Serge S. Gulin

Actual state of tests from CI: ``` 838/871 tests successful ``` - [ ] allbackends/evaluator004 - [ ] allbackends/popen2 - [ ] chez/chez007 - [ ] chez/constfold2 - [ ]...

``` 845/872 tests successful ``` - [ ] allbackends/evaluator004 - [ ] allbackends/popen2 - [ ] chez/constfold2 - [ ] idris2/evaluator/spec001 - [ ] idris2/interactive/interactive004 - [ ] idris2/interactive/interactive005 -...

``` 861/874 tests successful Failing tests: idris2/interactive/interactive004 idris2/interactive/interactive005 idris2/literate/literate003 idris2/reg/reg049 idris2/evaluator/spec001 idris2/schemeeval/schemeeval003 idris2/schemeeval/schemeeval004 idris2/schemeeval/schemeeval001 idris2/schemeeval/schemeeval005 idris2/schemeeval/schemeeval002 idris2/schemeeval/schemeeval006 idris2/reflection/reflection006 chez/constfold2 ```

Closed in favor * https://github.com/idris-lang/Idris2/pull/3512 * https://github.com/idris-lang/Idris2/pull/3513

[Here](https://gitlab.haskell.org/ghc/ghc/-/merge_requests/13856#note_620056) I suspect that the issue with `c_openat` is due of fact we do not use system-defined `O_*`. So, they are just hardcoded. When `O_CREAT` check happened at ghc js-backend...

Unix package provides primitives which are used by cabal. Cabal makes assumptions about "atomicity" which really have no strong reasons for file movements. So, if cabal wants to continue using...

> A "low-levelish" (= without retries and fallbacks) atomicWriteFile should live in file-io Yeah, it makes lot of sense

Working on https://github.com/idris-lang/Idris2/pull/3513 to move forward with coming updates

@gallais @wizard7377 https://github.com/idris-lang/Idris2/pull/3513 is ready for review and merge.