 test
|
b2b48f8b0d
|
Add 89 new tree-sitter grammars (66 to 155 languages)
Vendor, wire up, and compile 89 new tree-sitter grammars, expanding
language support from 66 to 155 languages. All grammars pass security
audit (no dangerous patterns in scanner.c files).
New programming languages (31):
Solidity, Typst, GDScript, Gleam, PowerShell, Pascal, D, Nim, Scheme,
Fennel, Fish, AWK, Zsh, Tcl, Ada, Agda, Racket, Odin, ReScript,
PureScript, Nickel, Crystal, Teal, Hare, Pony, Luau, Janet, Sway,
NASM, Assembly, TLA+, Pkl, Cairo, Move, Squirrel, ISPC, FunC, Smali
New config/data/IDL formats (31):
Just, Astro, Blade, Go Template, Templ, Liquid, Jinja2, Prisma,
Hyprlang, DotEnv, Diff, WGSL, KDL, JSON5, Jsonnet, RON, Thrift,
Cap'n Proto, Properties, SSH Config, BibTeX, Starlark, Bicep, CSV,
Requirements, HLSL, VHDL, SystemVerilog, DeviceTree, Linker Script,
GN, Kconfig, BitBake, TableGen, Slang, LLVM IR, Smithy, WIT,
Go Mod, Mermaid, RST, Beancount, Puppet, PO, Regex, JSDoc,
gitattributes, gitignore, Apex, SOQL, SOSL
Infrastructure:
- scripts/new-languages.json: manifest for all new languages
- scripts/generate-lang-code.py: generates boilerplate from manifest
- scripts/audit-grammar-security.sh: pre-vendoring security scanner
- Fixed angle-bracket includes in 18 grammars
- Fixed PureScript scanner const mismatch
- Fixed VHDL scanner void* API signatures
- Fixed RST tree_sitter_rst/ subdirectory include paths
- Copied crystal unicode.c (extra scanner dependency)
All new languages start with minimal lang specs (module_types only).
Function/class/call extraction specs to be refined incrementally.
|
2026-04-15 21:28:51 +02:00 |
|
|
test
|
46319a53ff
|
Add grammar security audit and enhance vendor script
- vendor-grammar.sh: copy extra headers (.h, .inc) and common/ subdirs
from grammar src/ directories. Needed for Astro (tag.h), PureScript/
Typst (unicode.h), VHDL (.h/.inc files), F# (common/scanner.h).
- audit-grammar-security.sh: pre-vendoring scanner for dangerous patterns
in vendored grammar C files. Checks for dangerous includes (sys/*,
unistd.h, dlfcn.h), dangerous calls (system, exec, popen, fopen,
socket, getenv, fork, dlopen), and suspicious patterns (constructor
attributes, inline assembly, base64 blobs). PASS/WARN/BLOCK per grammar.
|
2026-04-15 20:08:35 +02:00 |
|