Add 'gel' extension and alias to EdgeQL

[1st1]

Description

EdgeDB got recently renamed to Gel [1]. The schema language extension changed from .esdl to .gel. This PR adds .gel as a recognized file extension and gel as a recognized code-block language selector.

[1] https://www.geldata.com/blog/edgedb-is-now-gel-and-postgres-is-the-future

Checklist:

• I am adding a new extension to a language.
  • The new extension is used in hundreds of repositories on GitHub.com
    • Search results for each extension:
      • https://github.com/search?type=code&q=NOT+is%3Afork+path%3A*.gel+%2Fmodule+%5Cw%2B+%5C%7B%2F
      • This is a very recent change in the defaults regarding EdgeDB being renamed to Gel, so realistically we expect similar adoption of this format to what we have for older extension .esdl (https://github.com/search?type=code&q=NOT+is%3Afork+path%3A*.esdl). The new extension (taking heuristics into account) is currently ~95 hits in the search, though.
  • I have included a real-world usage sample for all extensions added in this PR:
    • Sample source(s):
      • Self-written
    • Sample license(s):
      • MIT
  • I have included a change to the heuristics to distinguish my language from others using the same extension.

[lildude]

Please follow the instructions in CONTRIBUTING.md for adding a new extension. Please also use and correctly complete the PR template… it's there for a reason.

[1st1]

@lildude, sincere apologies for not following the process. As an open source maintainer myself I should have had better judgement.

We've updated the commit / PR message, please take another look.

[lildude]

I recommend replacing your one-liner samples with more representative examples as they're used to train the classifier and very small samples are known to cause incorrect results.

[lildude]

This is unnecessary as this is the only language associated with the .gel extension so we'll never get to the heuristics.

Suggested change

	- extensions: ['.gel']
	rules:
	- language: EdgeQL
	pattern:
	# We definitely want the match unicode characters as part of the \w
	- '(?i:^\s*using\s+(extension|future)[^\n]*?;)'
	- '(?iu:^\s*module\s+(\w+|`.+?`)\s*\{)'
	- '(?iu:^\s*abstract\s+(inheritable\s+)?annotation\s+(\w+|`.+?`)::(\w+|`.+?`))'
	- '(?iu:^\s*abstract\s+(constraint|index|link|property)\s+(\w+|`.+?`)::(\w+|`.+?`))'
	- '(?iu:^\s*(alias|function)\s+(\w+|`.+?`)::(\w+|`.+?`))'
	- '(?iu:^\s*(optional|required|single|multi)?\s*global\s+(\w+|`.+?`)::(\w+|`.+?`))'
	- '(?iu:^\s*(abstract\s+)?(scalar\s+)?type\s+(\w+|`.+?`)::(\w+|`.+?`))'
	negative_pattern:
	# Check for common comments that are invalid in EdgeQL.
	#
	# The caveat is that they technically would be legal inside string
	# literals, but it is extremely unlikely to have such string literals in
	# the Gel schema (i.e. appearing as defaults or annotations). However, to
	# reduce the possibility of false positives here we also restrict illegal
	# comment detection to the beginning of a line.
	- '^\/\/.+$'
	- '^\/\*.*?\*\/'

[vpetrovykh]

We've included the heuristics largely because we've found a lot of .gel files that have nothing to do with our database schema (https://github.com/search?type=code&q=NOT+is%3Afork+path%3A*.gel has ~1.8K code hits). So we didn't think it would be appropriate to target all .gel files indiscriminately.

If you're saying that regardless of that intent, this is not how the heuristics work and would not help there, we can remove them, of course.

[vpetrovykh]

I've added a more representative example. It's still synthetic, but it's based much more closely on actual real schema files.