Nicolás Ojeda Bär
Nicolás Ojeda Bär
Hello, thank you for your report. However, precise reproduction instructions are needed in order to investigate any issue: what exact system are you using, how exactly did you try to...
Is this a duplicate of #11799? Note that this mode already exists: you can add `(empty_module_interface_if_absent)` to your library, see the manual: https://dune.readthedocs.io/en/latest/reference/dune/library.html.
> I'd still like to be able to exempt certain files though where I deliberately want to expose everything (useful for modules that just contain type definitions). A PR would...
> The changes to the `.cmx` files have intentionally been as extensible as possible to let future changes not require changes to that format again. I haven't been following the...
> I've end-to-end tested this PR successfully with MSVC for bytecode with the modifications I posted. Thanks! Thanks for the fixes, and the confirmation!
Also, some tests are failing because this addition modifies some backtraces. They will need to be adapted to the new results.
> Not quite sure what's happening with the failing hygiene test. The output advises to run `tools/sync_stdlib_doc` but that just erases the new functions from hashtbl.mli, which is very much...
Thanks @giltho. Since it touches the stdlib, this PR needs a second official approval before it can be merged. Besides @dra27 or @gasche maybe @yallop could be also interested in...
Thanks for the benchmark @giltho! As far as I'm concerned, they confirm there is no performance regression in this PR.
Merged, thanks!