[codex] Add native C outline parser

[justrach]

Summary

Adds native C/C++ outline parsing to codedb's existing line parser path.

This covers:

• #include imports
• #define macros
• named struct, enum, and union declarations
• simple typedef aliases
• likely function definitions, including pointer returns and brace-on-next-line style
• additional C++ extension detection for .cc, .hh, .cxx, and .hxx

Why

wiki.codes production storage currently has a large C/C++ corpus with tree/search coverage but no structural symbols:

language=c    files=177123  files_with_symbols=0  symbols=0
language=cpp  files=50596   files_with_symbols=0  symbols=0

The root cause is that detectLanguage recognizes .c and .h, and the brace-language helpers already include C/C++, but parseOutlineWithParser never dispatches C/C++ files to a parser.

Safety

The parser is intentionally conservative. It skips prototypes, function pointer typedefs, control-flow calls, comments, strings, and macro calls instead of trying to classify ambiguous C syntax.

Validation

zig build test
zig build

Closes #319.

[github-actions]

Benchmark Regression Report

Threshold: 10.00%

Tool	Base (ns)	Head (ns)	Delta	Status
codedb_bundle	322877	308076	-4.58%	OK
codedb_changes	37349	38376	+2.75%	OK
codedb_deps	5300	5125	-3.30%	OK
codedb_edit	4195	4303	+2.57%	OK
codedb_find	43447	47102	+8.41%	OK
codedb_hot	55665	54727	-1.69%	OK
codedb_outline	156150	154026	-1.36%	OK
codedb_read	57547	50385	-12.45%	OK
codedb_search	142098	145000	+2.04%	OK
codedb_snapshot	2058274	2088012	+1.44%	OK
codedb_status	237790	248683	+4.58%	OK
codedb_symbol	39692	39616	-0.19%	OK
codedb_tree	49180	51527	+4.77%	OK
codedb_word	39721	39503	-0.55%	OK