detect: add common extension language coverage

[justrach]

Summary

Adds language detection coverage for the high-volume extensions found during corpus inspection:

• .cc remains covered by C++ detection; .mm now maps into the C++ family path
• .java, .kt
• .svelte, .vue, .astro
• .sh
• .css, .scss
• .sql, .proto
• .f90
• .ll, .mlir, .td

This is detection/tree/snapshot coverage, not full native parser coverage for every language. The goal is to stop these files showing up as unknown and make repo trees more accurate before deeper parser work lands.

Notes

• .mm is routed through the existing C/C++ family classification so the current C/C++ outline path can still extract C-like symbols where applicable.
• Added compact-comment handling for the newly detected language tags.
• Updated telemetry language name output to include the new tags.
• README now separates parser support from detection-only coverage so we do not overclaim.

Validation

• zig build test
• zig build

[github-actions]

Benchmark Regression Report

Threshold: 10.00%

Tool	Base (ns)	Head (ns)	Delta	Status
codedb_bundle	488017	477144	-2.23%	OK
codedb_changes	55354	54365	-1.79%	OK
codedb_deps	9305	9041	-2.84%	OK
codedb_edit	6102	6115	+0.21%	OK
codedb_find	60095	60833	+1.23%	OK
codedb_hot	95206	98571	+3.53%	OK
codedb_outline	238240	237478	-0.32%	OK
codedb_read	86101	84592	-1.75%	OK
codedb_search	176802	171104	-3.22%	OK
codedb_snapshot	2633728	2606327	-1.04%	OK
codedb_status	219268	206318	-5.91%	OK
codedb_symbol	67880	57037	-15.97%	OK
codedb_tree	89001	65627	-26.26%	OK
codedb_word	71142	67536	-5.07%	OK