pandoc-crossref filter

pandoc-crossref is a pandoc filter for numbering figures, equations, tables and cross-references to them.

The input file (like demo.md) can be converted into HTML, LaTeX, PDF, Markdown or other formats.

Optionally, you can use cleveref for LaTeX/PDF output, e.g. cleveref PDF, cleveref LaTeX, and listings package, e.g. listings PDF, listings LaTeX

You can also enable per-chapter numbering (as with --chapters for latex output). You need to specify -M chapters for non-LaTeX/PDF output however. Examples: HTML, Markdown, LaTeX, PDF.

This work is inspired by pandoc-fignos and pandoc-eqnos by @tomduck.

This package tries to use LaTeX labels and references if output type is LaTeX. It also tries to supplement rudimentary LaTeX configuration that should mimic metadata configuration by setting header-includes variable.

Installation

The easiest option to get pandoc-crossref on Windows, macOS, or Linux, is to download pre-built executables available at the releases page. Bear in mind that those are a product of automated build scripts, and as such, provided as-is, with zero guarantees. Feel free to open issues if those don't work though, I'll try to do what I can.

WARNING: When using pre-built executables, make sure that your pandoc version matches the version pandoc-crossref was built against, otherwise weird stuff will likely happen. Feel free to open issues if there's a new version of pandoc available, for which there are no pandoc-crossref builds.

NOTE: Linux and Windows binaries are packed with upx (not macOS though, since upx apparently has questionable interactions with Apple's x86 emulation on A1 processors). If you don't like the overhead, and don't mind 40-megabyte binaries, you can unpack those manually with upx -d pandoc-crossref. Also please notice that upx-packed binaries can break in some exotic environments, like empty chroot with no access to /proc, etc.

Also, for those feeling adventurous, the automatic builds for the latest commits are available. Latest builds can be found on the nightlies tag (despite the name, those aren't actually built nightly, but on each push instead)

If you don't trust random binaries downloaded off the Internet (which is completely reasonable), you're welcome to build from source. You have two preferred options for that: building from Hackage with cabal-install, or from repository with stack (you'll only need stack and maybe git). See below for build instructions.

If you're completely new to Haskell, the latter, i.e. building from repo with stack, is the easier option in most cases.

This repository is also a nix flake. You can use nix to get it installed.

Alternatively, you can use a version provided by a third party. At the time of writing, pandoc-crossref is provided on the following platforms (that I am aware of):

• Arch Linux
• NixOS or Nix package manager (via nixpkgs.haskellPackages attribute)
• MacOS (via Homebrew)
• FreeBSD official binary package textproc/hs-pandoc-crossref
• Any Linux distribution (via Linuxbrew)
• Gentoo Linux (via gentoo-haskell overlay)
• Windows (via scoop)

Building from Hackage with cabal-install

You'll need to get GHC and cabal-install installed first. By far the easiest way to get those is via ghcup.

Describing using ghcup is out of scope for this small guide, but TL;DR is this:

ghcup install ghc
ghcup install cabal

After you got cabal-install and ghc, run:

cabal v2-update
cabal v2-install --install-method=copy pandoc-cli pandoc-crossref

This will get pandoc-crossref and pandoc executables copied to $HOME/.cabal/bin (by default, if not, check your cabal config file installdir setting -- find out where your config file is by running cabal help user-config), which you can then add to PATH or copy/move the symlinks where you want them.

Refer to cabal documentation if you need to build a particular version (TL;DR: add --constraint pandoc-crossref==<version> to the installation command)

Note: if you're using cabal to build from a repo checkout, and not from Hackage as described above, you'll need to either match the compiler version specified in ghcver in .github/workflows/haskell.yml, or remove cabal.project.freeze from the root of the repository. Otherwise, cabal will complain about version mismatch of boot packages (like base, ghc-boot-th, etc)

Building from repo with stack

First of all, get stack if you don't have it already: see the official stack documentation. Note that stack can also be installed via ghcup, and on Linux it is usually available in your package manager.

If you have git, you can now clone the repository and build:

git clone https://github.com/lierdakil/pandoc-crossref.git
cd pandoc-crossref
git checkout <commit/tag/branch>
stack install

If you don't have git, just download the sources for your preferred commit/branch/tag via the GitHub interface, and run stack install in the directory that contains stack.yaml file.

This will install pandoc-crossef executable to $HOME/.local/bin. You might also want to separately run stack install pandoc-cli in the same directory (i.e. the root of the repository, the one containing stack.yaml file)

Installing as a nix flake

TL;DR:

nix profile install github:lierdakil/pandoc-crossref

will install the latest commit from the master branch. You can also specify a commit, branch or tag, e.g.:

nix profile install github:lierdakil/pandoc-crossref/71c8c8508c222bf4110794457fdf0391b05fb9a9

You can also get the corresponding pandoc version installed via

nix profile install github:lierdakil/pandoc-crossref#pandoc

Since you will generally want both, there's an option to install both at the same time, too:

nix profile install github:lierdakil/pandoc-crossref#pandoc-with-crossref

Aside from added convenience, this guarantees pandoc and pandoc-crossref versions to be consistent across updates.

Finally, you can start a nix shell with both pandoc and pandoc-crossref using

nix develop github:lierdakil/pandoc-crossref

Warning: this uses haskell.nix infrastructure for builds (because Haskell support in Nix is borked, and has been for a long time). This means that unless you use their substituters, you'll build multiple GHC versions from source. To avoid that, add https://cache.iog.io to substituters in nix.conf and hydra.iohk.io:f/Ea+s+dFdN+3Y/G+FDgSq+a5NEWhJGzdjvKNGv0/EQ= to trusted-public-keys.

You can also use pandoc-crossref's binary cache by adding https://pandoc-crossref.cachix.org and pandoc-crossref.cachix.org-1:LI9ABFTkGpPCTkUTzoopVSSpb1a26RSTJNMsqVbDtPM= to substituters and trusted-public-keys respectively.

The flake includes both by default, so if you're a nix trusted user and accept these configurations during flake evaluation those will be used automatically.

Notice Fedora users

cabal-install package is not enough to build pandoc-crossref (see #132). To get a sane Haskell build environment, you need to install the haskell-platform package (dnf install haskell-platform).

While on topic, if you don't want to rebuild pandoc itself from source, make sure you have pandoc and ghc-pandoc-devel dnf packages before attempting to build pandoc-crossref.

Usage

Usage information is available at https://lierdakil.github.io/pandoc-crossref/

Projects

The following projects use this filter:

• TechnicalMarkdown

License

This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.

See LICENSE for details.

© 2016 Nikolay Yakimov et al

Contributors (per GPL, holders of copyright on their respective contributions):

• Nikolay Yakimov
• gesh
• Rui Chen
• Raphael Das Gupta
• Masamichi Hosoda
• Felix Yan
• Wlad
• Wandmalfarbe
• Vekhir
• Silas Benson
• shutingrz
• Shaun Jackman
• scoavoux
• Salim B
• Matthew Salganik
• Jan-T. Brinkmann
• hseg
• Han Joosten
• Hadrien Mary
• Gleb Popov
• Gabriel Nützi
• David Arnold
• Chris Black
• Bart Mesuere
• Albert

This repository includes code from https://github.com/roelvandijk/roman-numerals, covered by a different license. See licenses/LICENSE.roman-numerals for details.

How to bump pandoc version

1. Change PANDOC_VERSION in .github/workflows/haskell.yml to the new Pandoc version.

2. Run make update. You need at least nix, stack and cabal (i.e. cabal-install) installed and in PATH.

   If it doesn't do anything, consider nuking cabal.project.freeze, flake.lock, stack.yaml and stack.yaml.lock and trying again.

3. Build and test.

4. Fix broken tests.

   Note that you can regenerate most golden tests with either make regen-test-fixtures if using Nix, or just running ./mkcheck.sh and ./mkinttest.sh with appropriate pandoc and pandoc-crossref binaries in scope (so e.g. via stack exec).

5. Repeat 3-4 until all tests pass.