SubRip detected as SRecode Template

[MichaelHinrichs]

https://github.com/search?q=language%3A%22SRecode+Template%22+&type=repositories
The entire first page, and much of the following pages, are subtitles.

[lildude]

So there are a few things here:

1. you're looking at the Search results. Search is a completely independent library from Linguist though it feeds off Linguist
2. the results are showing the predominant language detected in the repo, not the repo's language
3. SubRip isn't likely to appear commonly in those results as it is considered "data" by Linguist so will only appear on the language stats if someone implements an override.

The fact you're seeing SubRip detected as SRecode will be because both languages share the .srt extension and all those repos do not contain files that predominantly match the heuristic:

linguist/lib/linguist/heuristics.yml

Lines 704 to 707 in 39fd5e9

	- extensions: ['.srt']
	rules:
	- language: SubRip Text
	pattern: '^(\d{2}:\d{2}:\d{2},\d{3})\s*(-->)\s*(\d{2}:\d{2}:\d{2},\d{3})$'

... so things are falling through to the classifier and it's "knowledge" is based entirely on the samples we have.

If you can improve the heuristic, we're open to a PR.