Support for multiple languages with the same file extension

[remyblok]

cloc already looks at the shebangs for files. But it would be very useful if there could be other methods for defining types that have the same file extensions. I'm thinking specifically of json files, which contain so many different types information. In my scenario I have Microsoft Azure Logic App code and Microsoft Azure Resource Management (ARM) templates both with json-extensions. But these are very different types and I would like to count them separately.

Would it be possible to add an optional regex to a type definition for 'type identification'. So I can define a type Logic App with extension json where the files should contain .*https://schema.management.azure.com/providers/Microsoft.Logic/schemas/2016-06-01/workflowdefinition.json.* for example.

Ordering is then important, if a file is matched it will count towards that defined type. All more specific types, i.e. with 'type identification' regex should be before any types with the same extension but without the 'type identification'.

This may also solve #837.

[AlDanial]

Re: "where the file should contain..." does the position of the line to match matter? It would be nice to only have to examine the first, say, 10 lines instead of every line in the file. I've seen some large JSON files.

[remyblok]

Hmm, good question. I've check for my scenario, and then the identification would be between the 1st and 120th line. Not the most restraining... I would have hoped that the file needs to be read and analyzed for comments etc. in full anyway, so that the position would not matter. That's why I suggested regex, since that is already used. Maybe limit the function to files smaller than, say, 250kb?

[AlDanial]

Can you point me to, or attach, a couple of JSON files that would map to different languages?