Add World of Warcraft TOC file lexer

[t-mart]

This PR adds lexer support for World of Warcraft TOC files, which are metadata files that describe addons that augment the game. See 1, 2 for examples of TOC files.

This is a widely developed and distributed file format: On Curseforge, an addon index, there are roughly 10,000 addons. The top addons have 100s of millions of downloads (see 1, 2, 3).

This PR has support for all constructs in a TOC file:

• Tags, including determination of tag type: official (including locale), user-defined, or non-conforming tags
• Addon files
• Comments

I based my work off on nebularg/language-toc-wow, which serves as GitHub's linguist reference, and the Wowpedia TOC Format article.

Example files have been included, and their --golden output looks good to me. All tests pass.

[t-mart]

Oh, also, the standard file suffix for TOCs (.toc) conflicts with an existing Tex suffix. I'm not sure what to do about that.

[jeanas]

Some nits, generally looks OK.

For the conflict with TeX on *.toc, you need to write an analyse_text method that returns a score indicating how probable it is that the file is a Warcraft TOC file. Are there any recurring patterns occurring in those files?

[t-mart]

Are there any recurring patterns occurring in those files?

There are definitely tells, yes. I'll have to see look into how to implement this method and look into how this toc differs from Tex's.

Working on this today.

[t-mart]

@jean-abou-samra, thank you for your feedback earlier. Since then, I have

• added a analyse_text method to this lexer
• incorporated your suggestions to my regex patterns
• Made some minor reformats/refactors

I also fixed some other small things I found that are unrelated to this lexer, but that impeded my development workflow:

• The Makefile specifies the python3 executable, but this was unavailable on my Windows machine. I only have python, and since no other versions of Python except 3 are supported, it's probably a safe move. I have removed the 3.
• Speaking of supported versions, the URL in Contributing.md for supported versions was broken. I have replaced it with a working one.
• Another Windows quirk: the default encoding for file IO is, unfortunately, not UTF-8 like most everyone else, but instead cp1252. Therefore, I was unable to decode some of the lexer source files in the scripts/count_token_references.py. To fix, I added an explicit UTF-8 encoding to the read_text call. (Note: this does not affect importing source files or anything -- it only manifests if you're trying to read arbitrary files. See this for more info.)

Are these other (minor?) changes acceptable in this PR, or should I make new ones?

[jeanas]

encoding="utf-8" is good (I can't wait for Python 3.15 where it will be the default, it will fix more bugs than it will create…).

The URL fix is good.

The python3 change is more debatable. There are still systems where python refers to Python 2. You can just run PYTHON=python make (or whatever the Windows shell syntax is) to use python.

[t-mart]

The python3 change is more debatable.

Yeah, I know what you mean. I don't want to get stuck on it though, so I will revert.

[t-mart]

Reversion to python3 committed and or "" removed. Awaiting your feedback.

[jeanas]

LGTM.

[t-mart]

Thanks @jean-abou-samra! This was a very pleasant PR. I really appreciate your help.