Closed mmanela closed 3 months ago
Your approach looks correct to me. However, I believe we want to ensure both sourcegraph and zoekt use the same logic for file type identification. In that case it may be fine for us to just import zoekt, but more than likely @varungandhi-src has a better idea. I know he was exploring consistency in language identification a few months ago.
There are languages which Linguist doesn't plan to handle but we want to still support. This change allows us to define our own list of these languages as a fallback for when linguist (and go-enry) do not know about a language.
Part of https://linear.app/sourcegraph/issue/GRAPH-622/magik-search-filter-support