This issue is scoped to syntax highlighting and Matlab only. Subpart of https://github.com/sourcegraph/sourcegraph/issues/55554

For reference:

• Place where the language filetype is optionally set in the SG backend: https://sourcegraph.com/github.com/sourcegraph/sourcegraph@64014d315541081a078bfd63473bd2fcca025005/-/blob/internal/highlight/highlight.go?L413-430
• Use of the filetype in the syntax highlighter: https://sourcegraph.com/github.com/sourcegraph/sourcegraph@64014d3/-/blob/docker-images/syntax-highlighter/crates/sg-syntax/src/lib.rs?L142:32-142:40#tab=def

AFAICT, no frontend code is involved in determining the language as passed to the syntax highlighter.

Recommended fix: For the .m file extension, go through go-enry to detect the language.

• Potential tweak: Always determine the language through go-enry and use that to set the filetype for syntax highlighting, regardless of the file extension.
• Caution: We should probably cut the file to an arbitrary fixed size to bound the amount of time spent by go-enry doing language detection, in case the file in question is very large.

Related: https://github.com/sourcegraph/sourcegraph/issues/56379