Better language state detect rule

[yp05327]

Description

gitea/modules/git/repo_language_stats_nogogit.go

Line 73 in ff96873

// by default we will only count the sizes of programming languages or markup languages

gitea/modules/git/repo_language_stats_nogogit.go

Line 76 in ff96873

// or if there's only one language in the repository

When there's many non programming language or non markup language files in a repo, but if there's only one language, it will show this language no matter what type it is. This logic may confuse user in some cases.
e.g. there's only one python file, and many SQL file in the repo, then you will see that the language is 100% Python.

But if you remove the python file, you will see the language is 100% SQL.

It seems like a bug to the end user.

Gitea Version

latest

Can you reproduce the bug on the Gitea demo site?

Yes

Log Gist

No response

Screenshots

No response

Git Version

No response

Operating System

No response

How are you running Gitea?

build

Database

None

[henrygoodman]

It seems like Github by default does not show any languages that are not configured by linguist (SQL is classified a data language according to them), so a normal GH repo with only SQL files shows no language stats.

github/markup#348

Seems like it might be a deliberate choice here to show some fallback if no languages were detected (my speculation), although I agree it could be muddying the waters (i.e. if you want to show stats for data languages as a fallback, why not just include them by default too)