False Classification with unknown language M

[Kiyotoko]

Hello everyone,

I use Matlab files in my repository which end with .m. However, some files are not recognized as Matlab, but as M. When I browse the repository, it does not show me which files they are.

Thank you all for your time and support.

[lildude]

There is only one file that has been incorrectly identified:

➜ github-linguist vault/ --breakdown
46.96%  6113       Python
38.88%  5062       MATLAB
10.58%  1377       M
3.58%   466        Ruby

[...]

M:
Klasse 11/Physik/Projects/Octave/task_11.m

[...]
➜ 

It hasn't been correctly identified because the content didn't match the Matlab regex used to differentiate all the users of the .m extension:

linguist/lib/linguist/heuristics.yml

Lines 400 to 417 in 916bd8f

	- extensions: ['.m']
	rules:
	- language: Objective-C
	named_pattern: objectivec
	- language: Mercury
	pattern: ':- module'
	- language: MUF
	pattern: '^: '
	- language: M
	pattern: '^\s*;'
	- language: Mathematica
	and:
	- pattern: '\(\*'
	- pattern: '\*\)$'
	- language: MATLAB
	pattern: '^\s*%'
	- language: Limbo
	pattern: '^\w+\s*:\s*module\s*\{'

... so it's fallen through to the classifier which can only guess based on the samples in this repo.

Now things get interesting... the reason the Search that happens when you click the M in the sidebar can't find it appears to be down to a caching issue somewhere. At some point the file was identified as Objective-C (for the same reason as before) and this has been cached as we can see if we search for the filename instead of the language:

Unfortunately, Linguist can't do anything about the caching, but you can implement an override to force the language. You'll probably have to make a change to another file after adding the override to force a full analysis again which should then correct things.

[Kiyotoko]

Thank you very much, this has helped a lot.