Closed jeltsch closed 4 months ago
Hi, thank you for this ticket, but Haddock now lives full-time in the GHC repository! Let's talk about this on https://gitlab.haskell.org/ghc/ghc/ :)
The haddock
package on Cabal still mentions this here as the bug tracker, and the Haddock documentation even says that Haddock’s code is maintained in a darcs repository (and doesn’t mention where to report bugs).
I am fully aware, thank you very much. This migration has happened after the latest release.
Issue filed in the GHC GitLab space: https://gitlab.haskell.org/ghc/ghc/-/issues/24917.
Fantastic.
Currently, [the subsection on SGML-style character references][character-references] in the documentation still says the following:
However, it seems that Haddock 2.30.0 mostly supports Haskell code and documentation with non-ASCII characters encoded in UTF-8. I have only spotted the following problems when running
cabal haddock
under GNU/Linux with the locale set toC
:A non-ASCII character in the
synopsis
field of the Cabal file causes an error message of the following form:haddock: internal error: ⟨path⟩/haddock-response⟨number⟩-⟨number⟩.txt: hGetContents: invalid argument (cannot decode byte sequence starting from ⟨number⟩)
A non-ASCII character in the identifier of an undocumented entity causes, during the attempt to output a corresponding warning, an error message of the following form:
haddock: internal error:: commitBuffer: invalid argument (cannot encode character ⟨character⟩)
I think that it would be good to fix these and any other such deficiencies in UTF-8 support and then update the above-mentioned section of the documentation accordingly. Does this sound sensible?
[character-references]: https://haskell-haddock.readthedocs.io/v2.20/markup.html#character-references "Character References"