pypandoc: 1.2.0 -> 1.3.3

[kristoff3r]

Motivation for this change

pypandoc 1.2.0 is broken with the newest pandoc, as it relied on scraping "pandoc -h" output which was removed and replaced with a flag (--list-output-formats).

Two new tests fail under sandbox. One of them requires network access, the other one I'm not sure about, but both work with sandboxing disabled. Maybe the tests should be patched instead?

Things done

• Tested using sandboxing
  (nix.useSandbox on NixOS,
  or option build-use-sandbox in nix.conf
  on non-NixOS)
• Built on platform(s)
  • NixOS
  • macOS
  • Linux
• Tested compilation of all pkgs that depend on this change using nix-shell -p nox --run "nox-review wip"
• Tested execution of all binary files (usually in ./result/bin/)
• Fits CONTRIBUTING.md.

---

[mention-bot]

@kristoff3r, thanks for your PR! By analyzing the history of the files in this pull request, we identified @zraexy, @zimbatm, @peti and @FRidh to be potential reviewers.

[FRidh]

Yes, please fix or disable the specific tests instead of disabling the whole test suite.

[kristoff3r]

I have patched out the two tests, the second one was a real bug triggered by having dashes in the tmp dir, reported here: JessicaTegner/pypandoc#130

[kristoff3r]

The test failure in Travis is weird. When I run

nix-shell -p pandoc haskellPackages.pandoc-citeproc texlive.combined.scheme-small python35 -I nixpkgs=/home/kris/git/nixpkgs --run "./tests.py"

or

nix-shell -p python35Packages.pypandoc -I nixpkgs=https://github.com/kristoff3r/nixpkgs/tree/fix/pypandoc.tar.gz

it works, but not when I run

nix-shell -p python35Packages.pypandoc -I nixpkgs=/home/kris/git/nixpkgs

The only difference I have found is that a variable has type <class 'bytes'> when it fails, but type <class 'str'> when it succeeds. Any ideas?

[FRidh]

I don't know what failure you refer to. What is ./tests.py? You enter a shell with the specified packages on your PATH. Is tests.py a copy of the test suite that you have locally?

The only issue that is left that I see is fixing the tests for Python 3. You likely need to add pkgs.glibcLocales and set LC_ALL = "en_US.UTF-8";

[kristoff3r]

Ah thanks, it was the python 3 test failure I meant.

Why is that necessary, and/or is it documented somewhere?

[kristoff3r]

Is there more I can do for this to go through? The darwin error seems unrelated to this change, and I can't debug it myself.

[zimbatm]

Just tested and compiled fine on Darwin