restructure: move el compiler content into lang/
This commit is contained in:
Will Anderson
File diff suppressed because it is too large
Load Diff
File diff suppressed because it is too large
Load Diff
@@ -0,0 +1,542 @@
|
||||
// compiler.el — el self-hosting compiler pipeline
|
||||
//
|
||||
// Wires lexer -> parser -> codegen into a single compile() function.
|
||||
// This is the bootstrap entry point: compiled once by the Rust el-compiler,
|
||||
// then self-hosted from that point forward.
|
||||
//
|
||||
// Two backends:
|
||||
// - C (default) compile() -> emits C source linked against el_runtime.c
|
||||
// - JS (--target=js) compile_js() -> emits JS source linked against el_runtime.js
|
||||
//
|
||||
// Compile the C output with:
|
||||
// cc -o <prog> <prog>.c el_runtime.c
|
||||
//
|
||||
// Run the JS output with:
|
||||
// node <prog>.js (after copying el_runtime.js next to it)
|
||||
|
||||
import "lexer.el"
|
||||
import "parser.el"
|
||||
import "codegen.el"
|
||||
import "codegen-js.el"
|
||||
|
||||
// compile — full pipeline (C target): source string -> C source string
|
||||
fn compile(source: String) -> String {
|
||||
let tokens: [Map<String, Any>] = lex(source)
|
||||
let stmts: [Map<String, Any>] = parse(tokens)
|
||||
// Token list is no longer needed after parsing — release it to free memory
|
||||
// before codegen allocates its own working data on large source files.
|
||||
el_release(tokens)
|
||||
codegen(stmts, source)
|
||||
}
|
||||
|
||||
// compile_js — full pipeline (JS target, module mode): source string -> JS source string
|
||||
fn compile_js(source: String) -> String {
|
||||
let tokens: [Map<String, Any>] = lex(source)
|
||||
let stmts: [Map<String, Any>] = parse(tokens)
|
||||
// Token list is no longer needed after parsing — release it to free memory.
|
||||
el_release(tokens)
|
||||
codegen_js(stmts, source)
|
||||
}
|
||||
|
||||
// compile_js_with_bundle — JS target in bundle mode.
|
||||
// Reads el_runtime.js from runtime_path and inlines it inside an IIFE.
|
||||
fn compile_js_with_bundle(source: String, runtime_path: String) -> String {
|
||||
let tokens: [Map<String, Any>] = lex(source)
|
||||
let stmts: [Map<String, Any>] = parse(tokens)
|
||||
el_release(tokens)
|
||||
let runtime_content: String = fs_read(runtime_path)
|
||||
if str_eq(runtime_content, "") {
|
||||
println("el-compiler: warning: --bundle: could not read runtime at " + runtime_path)
|
||||
println("el-compiler: warning: bundle output will be incomplete")
|
||||
}
|
||||
codegen_js_bundle(stmts, source, runtime_content)
|
||||
}
|
||||
|
||||
// compile_dispatch — pick a backend based on the requested target.
|
||||
// tgt = "c" | "js"
|
||||
// (The parameter is named `tgt` because `target` is a reserved keyword
|
||||
// in El's lexer — it would be tokenised as `Target`, breaking the
|
||||
// parser's identifier resolution.)
|
||||
fn compile_dispatch(tgt: String, source: String) -> String {
|
||||
if str_eq(tgt, "js") { return compile_js(source) }
|
||||
compile(source)
|
||||
}
|
||||
|
||||
// compile_dispatch_bundle — like compile_dispatch but bundle mode for JS.
|
||||
fn compile_dispatch_bundle(tgt: String, source: String, runtime_path: String) -> String {
|
||||
if str_eq(tgt, "js") { return compile_js_with_bundle(source, runtime_path) }
|
||||
compile(source)
|
||||
}
|
||||
|
||||
// Detect a `--target=<lang>` flag in argv and return the target.
|
||||
// Returns "c" if none specified or unrecognized.
|
||||
fn detect_target(argv: [String]) -> String {
|
||||
let n: Int = native_list_len(argv)
|
||||
let i = 0
|
||||
while i < n {
|
||||
let a: String = native_list_get(argv, i)
|
||||
if str_starts_with(a, "--target=") {
|
||||
let v: String = str_slice(a, 9, str_len(a))
|
||||
return v
|
||||
}
|
||||
let i = i + 1
|
||||
}
|
||||
return "c"
|
||||
}
|
||||
|
||||
// Strip flags from argv, leaving only positional arguments.
|
||||
fn strip_flags(argv: [String]) -> [String] {
|
||||
let out: [String] = native_list_empty()
|
||||
let n: Int = native_list_len(argv)
|
||||
let i = 0
|
||||
while i < n {
|
||||
let a: String = native_list_get(argv, i)
|
||||
if !str_starts_with(a, "--") {
|
||||
let out = native_list_append(out, a)
|
||||
}
|
||||
let i = i + 1
|
||||
}
|
||||
return out
|
||||
}
|
||||
|
||||
// Detect --emit-header flag in argv.
|
||||
fn detect_emit_header(argv: [String]) -> Bool {
|
||||
let n: Int = native_list_len(argv)
|
||||
let i = 0
|
||||
while i < n {
|
||||
let a: String = native_list_get(argv, i)
|
||||
if str_eq(a, "--emit-header") { return true }
|
||||
let i = i + 1
|
||||
}
|
||||
return false
|
||||
}
|
||||
|
||||
// Detect --bundle flag in argv.
|
||||
fn detect_bundle(argv: [String]) -> Bool {
|
||||
let n: Int = native_list_len(argv)
|
||||
let i = 0
|
||||
while i < n {
|
||||
let a: String = native_list_get(argv, i)
|
||||
if str_eq(a, "--bundle") { return true }
|
||||
let i = i + 1
|
||||
}
|
||||
return false
|
||||
}
|
||||
|
||||
// Detect --minify flag in argv.
|
||||
fn detect_minify(argv: [String]) -> Bool {
|
||||
let n: Int = native_list_len(argv)
|
||||
let i = 0
|
||||
while i < n {
|
||||
let a: String = native_list_get(argv, i)
|
||||
if str_eq(a, "--minify") { return true }
|
||||
let i = i + 1
|
||||
}
|
||||
return false
|
||||
}
|
||||
|
||||
// Detect --obfuscate flag in argv.
|
||||
fn detect_obfuscate(argv: [String]) -> Bool {
|
||||
let n: Int = native_list_len(argv)
|
||||
let i = 0
|
||||
while i < n {
|
||||
let a: String = native_list_get(argv, i)
|
||||
if str_eq(a, "--obfuscate") { return true }
|
||||
let i = i + 1
|
||||
}
|
||||
return false
|
||||
}
|
||||
|
||||
// Build a unique temp file path: /tmp/elc-<pid>-<timestamp>.<suffix>
|
||||
fn make_temp_path(suffix: String) -> String {
|
||||
let pid: Int = getpid_now()
|
||||
let ts: Int = time_now()
|
||||
"/tmp/elc-" + native_int_to_str(pid) + "-" + native_int_to_str(ts) + "." + suffix
|
||||
}
|
||||
|
||||
// Reserved globals that terser and javascript-obfuscator must not mangle.
|
||||
// These are referenced from HTML onclick= attributes and other direct window usage.
|
||||
fn js_reserved_names() -> String {
|
||||
"neuronDemoToggle,neuronDemoSend,neuronDemoReset,signInWith,signInWithEmail,signUpWithEmail,sendMagicLink,signOut,resetPassword,sendResetEmail,updatePassword,showSignIn,showSignUp,hideReset,setSort,addFamilyMember,removeFamilyMember,copyForPlatform,entHeadcountChange,NEURON_CFG"
|
||||
}
|
||||
|
||||
// Find a CLI tool by checking node_modules paths first, then falling back to npx.
|
||||
// src_dir is the directory of the source file being compiled.
|
||||
// Returns the command string to invoke the tool, or "" if not found.
|
||||
fn find_node_tool(tool_name: String, src_dir: String) -> String {
|
||||
// 1. Check ./node_modules/.bin/<tool> relative to source file
|
||||
let cand1: String = src_dir + "/node_modules/.bin/" + tool_name
|
||||
let check1: String = str_trim(exec_capture("test -x " + cand1 + " && echo yes 2>/dev/null"))
|
||||
if str_eq(check1, "yes") { return cand1 }
|
||||
// 2. Check ../node_modules/.bin/<tool> (monorepo layout)
|
||||
let parent_dir: String = dirname_of(src_dir)
|
||||
let cand2: String = parent_dir + "/node_modules/.bin/" + tool_name
|
||||
let check2: String = str_trim(exec_capture("test -x " + cand2 + " && echo yes 2>/dev/null"))
|
||||
if str_eq(check2, "yes") { return cand2 }
|
||||
// 3. Fall back to npx if it is on PATH. npx will use the globally cached
|
||||
// package or download on first use. Use --no to avoid auto-install if
|
||||
// the package is not already cached; if that fails, try with --yes.
|
||||
let npx_path: String = str_trim(exec_capture("which npx 2>/dev/null"))
|
||||
if !str_eq(npx_path, "") { return "npx --yes " + tool_name }
|
||||
return ""
|
||||
}
|
||||
|
||||
// apply_minify — run terser on js_path, write result to out_path.
|
||||
// Returns true on success, false on failure.
|
||||
fn apply_minify(js_path: String, out_path: String, src_dir: String) -> Bool {
|
||||
let terser: String = find_node_tool("terser", src_dir)
|
||||
if str_eq(terser, "") {
|
||||
println("el-compiler: error: terser not found. Run 'npm install terser' in your project directory.")
|
||||
return false
|
||||
}
|
||||
let names: String = js_reserved_names()
|
||||
// Single-quote the mangle reserved list so the shell does not glob-expand
|
||||
// the bracket expression. The compress options are safe without quoting.
|
||||
let compress_opts: String = "passes=2,drop_console=false,drop_debugger=true"
|
||||
let mangle_reserved: String = "'reserved=[" + names + "]'"
|
||||
let cmd: String = terser + " " + js_path + " --compress " + compress_opts + " --mangle " + mangle_reserved + " --output " + out_path
|
||||
let ret: Int = exec_command(cmd)
|
||||
if ret == 0 { return true }
|
||||
println("el-compiler: error: terser failed (exit " + native_int_to_str(ret) + ")")
|
||||
return false
|
||||
}
|
||||
|
||||
// apply_obfuscate — run javascript-obfuscator on js_path, write result to out_path.
|
||||
// Returns true on success, false on failure.
|
||||
fn apply_obfuscate(js_path: String, out_path: String, src_dir: String) -> Bool {
|
||||
let obfuscator: String = find_node_tool("javascript-obfuscator", src_dir)
|
||||
if str_eq(obfuscator, "") {
|
||||
println("el-compiler: error: javascript-obfuscator not found. Run 'npm install javascript-obfuscator' in your project directory.")
|
||||
return false
|
||||
}
|
||||
let names: String = js_reserved_names()
|
||||
let cmd: String = obfuscator + " " + js_path + " --output " + out_path + " --compact true --simplify true --string-array true --string-array-encoding base64 --string-array-threshold 0.75 --identifier-names-generator hexadecimal --rename-globals false --self-defending false --reserved-names " + names
|
||||
let ret: Int = exec_command(cmd)
|
||||
if ret == 0 { return true }
|
||||
println("el-compiler: error: javascript-obfuscator failed (exit " + native_int_to_str(ret) + ")")
|
||||
return false
|
||||
}
|
||||
|
||||
// Resolve the runtime path for --bundle mode.
|
||||
// Looks for el_runtime.js next to the source file first;
|
||||
// if not found there, looks next to the elc binary itself.
|
||||
// Returns "" if not found anywhere (caller emits a warning).
|
||||
fn resolve_runtime_path(src_path: String) -> String {
|
||||
let src_dir: String = dirname_of(src_path)
|
||||
let candidate: String = src_dir + "/el_runtime.js"
|
||||
let existing: String = fs_read(candidate)
|
||||
if !str_eq(existing, "") {
|
||||
return candidate
|
||||
}
|
||||
return ""
|
||||
}
|
||||
|
||||
// Reconstruct an El type annotation string from a parsed type node.
|
||||
fn type_node_to_el(t: Map<String, Any>) -> String {
|
||||
let k: String = t["kind"]
|
||||
if str_eq(k, "Simple") { return t["name"] }
|
||||
if str_eq(k, "List") {
|
||||
let inner: String = type_node_to_el(t["inner"])
|
||||
return "[" + inner + "]"
|
||||
}
|
||||
if str_eq(k, "Map") {
|
||||
let kt: String = type_node_to_el(t["key"])
|
||||
let vt: String = type_node_to_el(t["val"])
|
||||
return "Map<" + kt + ", " + vt + ">"
|
||||
}
|
||||
"Any"
|
||||
}
|
||||
|
||||
// emit_header — write a .elh file from parsed statements.
|
||||
// Scans for FnDef nodes and emits 'extern fn' declarations.
|
||||
fn emit_header(stmts: [Map<String, Any>], hdr_path: String) -> Void {
|
||||
let n: Int = native_list_len(stmts)
|
||||
let i = 0
|
||||
let parts: [String] = native_list_empty()
|
||||
let parts = native_list_append(parts, "// auto-generated by elc --emit-header — do not edit\n")
|
||||
while i < n {
|
||||
let stmt = native_list_get(stmts, i)
|
||||
let kind: String = stmt["stmt"]
|
||||
if str_eq(kind, "FnDef") {
|
||||
let name: String = stmt["name"]
|
||||
if !str_eq(name, "main") {
|
||||
let params = stmt["params"]
|
||||
let ret_type: String = stmt["ret_type"]
|
||||
// build param list
|
||||
let np: Int = native_list_len(params)
|
||||
let pi = 0
|
||||
let param_parts: [String] = native_list_empty()
|
||||
while pi < np {
|
||||
let param = native_list_get(params, pi)
|
||||
let pname: String = param["name"]
|
||||
let ptype: String = param["type"]
|
||||
if str_eq(ptype, "") { let ptype = "Any" }
|
||||
let param_parts = native_list_append(param_parts, pname + ": " + ptype)
|
||||
let pi = pi + 1
|
||||
}
|
||||
let params_str: String = str_join(param_parts, ", ")
|
||||
let ret_str: String = ret_type
|
||||
if str_eq(ret_str, "") { let ret_str = "Any" }
|
||||
let sig: String = "extern fn " + name + "(" + params_str + ") -> " + ret_str
|
||||
let parts = native_list_append(parts, sig + "\n")
|
||||
}
|
||||
}
|
||||
let i = i + 1
|
||||
}
|
||||
let content: String = str_join(parts, "")
|
||||
let ok: Bool = fs_write(hdr_path, content)
|
||||
}
|
||||
|
||||
// ── Import resolution ────────────────────────────────────────────────────────
|
||||
//
|
||||
// elc supports two forms of import:
|
||||
// import "path/to/file.el" — quoted relative path
|
||||
// from module import { Name } — bare module name resolves to module.el
|
||||
// in the entry source's directory
|
||||
//
|
||||
// Codegen treats Import statements as no-ops (declarations only), so to
|
||||
// actually link bodies across files we textually concatenate every imported
|
||||
// source ahead of the entry source before lex/parse. resolve_imports does a
|
||||
// depth-first traversal with deduplication so any module that gets pulled in
|
||||
// transitively is included exactly once.
|
||||
|
||||
fn dirname_of(path: String) -> String {
|
||||
let n: Int = str_len(path)
|
||||
let i: Int = n - 1
|
||||
while i >= 0 {
|
||||
let c: String = str_slice(path, i, i + 1)
|
||||
if str_eq(c, "/") {
|
||||
return str_slice(path, 0, i)
|
||||
}
|
||||
let i = i - 1
|
||||
}
|
||||
return "."
|
||||
}
|
||||
|
||||
// Extract the resolved file path from a single trimmed source line. Returns
|
||||
// "" if the line is not an import.
|
||||
fn parse_import_line(trimmed: String, dir: String) -> String {
|
||||
if str_starts_with(trimmed, "import \"") {
|
||||
let after: String = str_slice(trimmed, 8, str_len(trimmed))
|
||||
let q: Int = str_index_of(after, "\"")
|
||||
if q > 0 {
|
||||
let mod: String = str_slice(after, 0, q)
|
||||
return dir + "/" + mod
|
||||
}
|
||||
}
|
||||
if str_starts_with(trimmed, "from ") {
|
||||
let after: String = str_slice(trimmed, 5, str_len(trimmed))
|
||||
// module name is the first whitespace-delimited token
|
||||
let sp: Int = str_index_of(after, " ")
|
||||
if sp > 0 {
|
||||
let mod_raw: String = str_slice(after, 0, sp)
|
||||
let mod: String = str_trim(mod_raw)
|
||||
if !str_eq(mod, "") {
|
||||
return dir + "/" + mod + ".el"
|
||||
}
|
||||
}
|
||||
}
|
||||
return ""
|
||||
}
|
||||
|
||||
// Recursively resolve imports starting from src_path. Returns the combined
|
||||
// source text with every imported module's body inlined ahead of the entry
|
||||
// source, deduplicated by absolute path. Uses state_set to track which paths
|
||||
// have already been pulled in for this run.
|
||||
//
|
||||
// Accumulates chunks into lists and joins once at the end to avoid the O(n²)
|
||||
// memory growth caused by repeated `prefix = prefix + chunk` concatenation.
|
||||
fn resolve_imports(src_path: String) -> String {
|
||||
let seen_key: String = "__elc_imp__:" + src_path
|
||||
let already: String = state_get(seen_key)
|
||||
if !str_eq(already, "") { return "" }
|
||||
state_set(seen_key, "1")
|
||||
|
||||
let source: String = fs_read(src_path)
|
||||
let dir: String = dirname_of(src_path)
|
||||
let lines: [String] = str_split(source, "\n")
|
||||
let n: Int = native_list_len(lines)
|
||||
|
||||
// Collect chunks into lists — O(1) amortized per append.
|
||||
// Join once at the end — O(n) single pass.
|
||||
let prefix_chunks: [String] = native_list_empty()
|
||||
let body_chunks: [String] = native_list_empty()
|
||||
let i: Int = 0
|
||||
while i < n {
|
||||
let line: String = native_list_get(lines, i)
|
||||
let trimmed: String = str_trim(line)
|
||||
let imp_path: String = parse_import_line(trimmed, dir)
|
||||
if !str_eq(imp_path, "") {
|
||||
// Use pre-compiled header if available (separate compilation).
|
||||
// Only check .elh for imported files — never for the entry file itself.
|
||||
let imp_elh_path: String = str_slice(imp_path, 0, str_len(imp_path) - 3) + ".elh"
|
||||
let imp_elh: String = fs_read(imp_elh_path)
|
||||
if !str_eq(imp_elh, "") {
|
||||
// Header exists: mark the .el as seen (so it won't be re-inlined
|
||||
// if something else also imports it) and use the header text.
|
||||
let seen_imp_key: String = "__elc_imp__:" + imp_path
|
||||
state_set(seen_imp_key, "1")
|
||||
let prefix_chunks = native_list_append(prefix_chunks, imp_elh)
|
||||
} else {
|
||||
let imp_body: String = resolve_imports(imp_path)
|
||||
let prefix_chunks = native_list_append(prefix_chunks, imp_body)
|
||||
}
|
||||
} else {
|
||||
let body_chunks = native_list_append(body_chunks, line + "\n")
|
||||
}
|
||||
let i = i + 1
|
||||
}
|
||||
return str_join(prefix_chunks, "") + str_join(body_chunks, "")
|
||||
}
|
||||
|
||||
// run_with_postprocess — codegen + minify + optional obfuscate pipeline.
|
||||
//
|
||||
// Called from main() when --minify or --obfuscate is active. Redirects stdout
|
||||
// to a temp file during codegen so the output can be passed through the
|
||||
// external tools (terser, javascript-obfuscator) before final emission.
|
||||
//
|
||||
// Pipeline: codegen -> terser -> (javascript-obfuscator) -> stdout or file
|
||||
fn run_with_postprocess(tgt: String, source: String, src_path: String, do_bundle: Bool, do_obfuscate: Bool, argc: Int, positional: [String]) -> Void {
|
||||
let src_dir: String = dirname_of(src_path)
|
||||
let tmp_gen: String = make_temp_path("js")
|
||||
let tmp_min: String = make_temp_path("min.js")
|
||||
|
||||
// Redirect stdout to tmp_gen so codegen println output is captured.
|
||||
stdout_to_file(tmp_gen)
|
||||
if do_bundle {
|
||||
let runtime_path: String = resolve_runtime_path(src_path)
|
||||
compile_dispatch_bundle(tgt, source, runtime_path)
|
||||
} else {
|
||||
compile_dispatch(tgt, source)
|
||||
}
|
||||
stdout_restore()
|
||||
|
||||
// Run terser: tmp_gen -> tmp_min
|
||||
let ok_min: Bool = apply_minify(tmp_gen, tmp_min, src_dir)
|
||||
if !ok_min {
|
||||
exec_command("rm -f " + tmp_gen + " " + tmp_min)
|
||||
exit(1)
|
||||
}
|
||||
|
||||
// Determine final result path (either tmp_min or post-obfuscation file).
|
||||
// Use state to pass the final path out of the optional obfuscation branch.
|
||||
state_set("__elc_final_js", tmp_min)
|
||||
|
||||
if do_obfuscate {
|
||||
let tmp_obf: String = make_temp_path("obf.js")
|
||||
let ok_obf: Bool = apply_obfuscate(tmp_min, tmp_obf, src_dir)
|
||||
if !ok_obf {
|
||||
exec_command("rm -f " + tmp_gen + " " + tmp_min + " " + tmp_obf)
|
||||
exit(1)
|
||||
}
|
||||
state_set("__elc_final_js", tmp_obf)
|
||||
}
|
||||
|
||||
let final_path: String = state_get("__elc_final_js")
|
||||
let final_js: String = fs_read(final_path)
|
||||
|
||||
// Clean up all temp files.
|
||||
exec_command("rm -f " + tmp_gen + " " + tmp_min)
|
||||
if do_obfuscate {
|
||||
exec_command("rm -f " + final_path)
|
||||
}
|
||||
|
||||
if argc >= 2 {
|
||||
let out_path: String = native_list_get(positional, 1)
|
||||
let ok: Bool = fs_write(out_path, final_js)
|
||||
if ok {
|
||||
return
|
||||
} else {
|
||||
println("el-compiler: failed to write output")
|
||||
exit(1)
|
||||
}
|
||||
}
|
||||
// No output file: print final JS to stdout.
|
||||
print(final_js)
|
||||
}
|
||||
|
||||
// main — CLI entry point.
|
||||
//
|
||||
// elc <source.el> # emit C to stdout
|
||||
// elc --target=js <source.el> # emit JS (module) to stdout
|
||||
// elc --target=js --bundle <source.el> # emit self-contained JS (IIFE) to stdout
|
||||
// elc --target=js --bundle --minify <source.el> # emit minified IIFE to stdout
|
||||
// elc --target=js --bundle --obfuscate <source.el> # emit minified+obfuscated IIFE to stdout
|
||||
// elc --target=c <source.el> <out.c> # write C to file
|
||||
// elc --target=js <source.el> <out.js> # write JS to file
|
||||
// elc --target=js --bundle <source.el> <out.js> # write bundled JS to file
|
||||
// elc --target=js --bundle --minify <source.el> <out.min.js> # write minified JS to file
|
||||
fn main() -> Void {
|
||||
let argv: [String] = args()
|
||||
// Use `tgt` not `target`: `target` is a reserved keyword in the lexer
|
||||
// (Section 1.5 of the language spec). detect_target itself is fine
|
||||
// because the function-name position has no token-class restriction.
|
||||
let tgt: String = detect_target(argv)
|
||||
let do_emit_header: Bool = detect_emit_header(argv)
|
||||
let do_bundle: Bool = detect_bundle(argv)
|
||||
let do_minify: Bool = detect_minify(argv)
|
||||
let do_obfuscate: Bool = detect_obfuscate(argv)
|
||||
// --obfuscate implies --minify: obfuscating unminified code is pointless.
|
||||
if do_obfuscate {
|
||||
let do_minify = true
|
||||
}
|
||||
let positional: [String] = strip_flags(argv)
|
||||
let argc: Int = native_list_len(positional)
|
||||
if argc < 1 {
|
||||
println("el-compiler: usage: elc [--target=c|js] [--bundle] [--minify] [--obfuscate] [--emit-header] <source.el> [<output>]")
|
||||
exit(1)
|
||||
}
|
||||
|
||||
// --minify and --obfuscate require --target=js
|
||||
if do_minify {
|
||||
if !str_eq(tgt, "js") {
|
||||
println("el-compiler: error: --minify and --obfuscate require --target=js")
|
||||
exit(1)
|
||||
}
|
||||
}
|
||||
|
||||
let src_path: String = native_list_get(positional, 0)
|
||||
|
||||
// When --emit-header is requested, parse the source file directly
|
||||
// (without inlining imports) and write out a .elh file alongside the .c.
|
||||
if do_emit_header {
|
||||
let raw_source: String = fs_read(src_path)
|
||||
let hdr_tokens: [Map<String, Any>] = lex(raw_source)
|
||||
let hdr_stmts: [Map<String, Any>] = parse(hdr_tokens)
|
||||
el_release(hdr_tokens)
|
||||
let hdr_path: String = str_slice(src_path, 0, str_len(src_path) - 3) + ".elh"
|
||||
emit_header(hdr_stmts, hdr_path)
|
||||
el_release(hdr_stmts)
|
||||
}
|
||||
|
||||
let source: String = resolve_imports(src_path)
|
||||
|
||||
// When post-processing (--minify or --obfuscate) is requested, redirect
|
||||
// stdout to a temp file so codegen output can be captured and piped through
|
||||
// the external tools. After codegen, restore stdout before emitting the
|
||||
// final result.
|
||||
if do_minify {
|
||||
run_with_postprocess(tgt, source, src_path, do_bundle, do_obfuscate, argc, positional)
|
||||
exit(0)
|
||||
}
|
||||
|
||||
// Standard path (no post-processing).
|
||||
let out: String = ""
|
||||
if do_bundle {
|
||||
let runtime_path: String = resolve_runtime_path(src_path)
|
||||
let out = compile_dispatch_bundle(tgt, source, runtime_path)
|
||||
} else {
|
||||
let out = compile_dispatch(tgt, source)
|
||||
}
|
||||
if argc >= 2 {
|
||||
let out_path: String = native_list_get(positional, 1)
|
||||
let ok: Bool = fs_write(out_path, out)
|
||||
if ok {
|
||||
exit(0)
|
||||
} else {
|
||||
println("el-compiler: failed to write output")
|
||||
exit(1)
|
||||
}
|
||||
}
|
||||
// No output path: codegen streamed to stdout already; out is "".
|
||||
}
|
||||
@@ -0,0 +1,993 @@
|
||||
// lexer.el - el self-hosting lexer
|
||||
//
|
||||
// Tokenises an el source string into a list of token maps.
|
||||
// Each token is a Map<String, Any> with keys:
|
||||
// "kind" -> String (e.g. "Int", "Ident", "Plus")
|
||||
// "value" -> String (the raw text of the token)
|
||||
//
|
||||
// Entry point: fn lex(source: String) -> [Map<String, Any>]
|
||||
//
|
||||
// Uses native_string_chars to split the source into a chars list,
|
||||
// then indexes it with native_list_get - avoids O(N-) string cloning.
|
||||
|
||||
// -- Character helpers ---------------------------------------------------------
|
||||
|
||||
fn lex_is_digit(ch: String) -> Bool {
|
||||
if ch == "0" { return true }
|
||||
if ch == "1" { return true }
|
||||
if ch == "2" { return true }
|
||||
if ch == "3" { return true }
|
||||
if ch == "4" { return true }
|
||||
if ch == "5" { return true }
|
||||
if ch == "6" { return true }
|
||||
if ch == "7" { return true }
|
||||
if ch == "8" { return true }
|
||||
if ch == "9" { return true }
|
||||
false
|
||||
}
|
||||
|
||||
fn lex_is_alpha(ch: String) -> Bool {
|
||||
if ch == "a" { return true }
|
||||
if ch == "b" { return true }
|
||||
if ch == "c" { return true }
|
||||
if ch == "d" { return true }
|
||||
if ch == "e" { return true }
|
||||
if ch == "f" { return true }
|
||||
if ch == "g" { return true }
|
||||
if ch == "h" { return true }
|
||||
if ch == "i" { return true }
|
||||
if ch == "j" { return true }
|
||||
if ch == "k" { return true }
|
||||
if ch == "l" { return true }
|
||||
if ch == "m" { return true }
|
||||
if ch == "n" { return true }
|
||||
if ch == "o" { return true }
|
||||
if ch == "p" { return true }
|
||||
if ch == "q" { return true }
|
||||
if ch == "r" { return true }
|
||||
if ch == "s" { return true }
|
||||
if ch == "t" { return true }
|
||||
if ch == "u" { return true }
|
||||
if ch == "v" { return true }
|
||||
if ch == "w" { return true }
|
||||
if ch == "x" { return true }
|
||||
if ch == "y" { return true }
|
||||
if ch == "z" { return true }
|
||||
if ch == "A" { return true }
|
||||
if ch == "B" { return true }
|
||||
if ch == "C" { return true }
|
||||
if ch == "D" { return true }
|
||||
if ch == "E" { return true }
|
||||
if ch == "F" { return true }
|
||||
if ch == "G" { return true }
|
||||
if ch == "H" { return true }
|
||||
if ch == "I" { return true }
|
||||
if ch == "J" { return true }
|
||||
if ch == "K" { return true }
|
||||
if ch == "L" { return true }
|
||||
if ch == "M" { return true }
|
||||
if ch == "N" { return true }
|
||||
if ch == "O" { return true }
|
||||
if ch == "P" { return true }
|
||||
if ch == "Q" { return true }
|
||||
if ch == "R" { return true }
|
||||
if ch == "S" { return true }
|
||||
if ch == "T" { return true }
|
||||
if ch == "U" { return true }
|
||||
if ch == "V" { return true }
|
||||
if ch == "W" { return true }
|
||||
if ch == "X" { return true }
|
||||
if ch == "Y" { return true }
|
||||
if ch == "Z" { return true }
|
||||
false
|
||||
}
|
||||
|
||||
fn is_alnum_or_underscore(ch: String) -> Bool {
|
||||
if lex_is_digit(ch) { return true }
|
||||
if lex_is_alpha(ch) { return true }
|
||||
if ch == "_" { return true }
|
||||
false
|
||||
}
|
||||
|
||||
fn lex_is_whitespace(ch: String) -> Bool {
|
||||
if ch == " " { return true }
|
||||
if ch == "\t" { return true }
|
||||
if ch == "\n" { return true }
|
||||
if ch == "\r" { return true }
|
||||
false
|
||||
}
|
||||
|
||||
fn make_tok(kind: String, value: String) -> Map<String, Any> {
|
||||
{ "kind": kind, "value": value }
|
||||
}
|
||||
|
||||
// -- Keyword lookup ------------------------------------------------------------
|
||||
|
||||
fn keyword_kind(word: String) -> String {
|
||||
if word == "let" { return "Let" }
|
||||
if word == "fn" { return "Fn" }
|
||||
if word == "type" { return "Type" }
|
||||
if word == "enum" { return "Enum" }
|
||||
if word == "match" { return "Match" }
|
||||
if word == "return" { return "Return" }
|
||||
if word == "if" { return "If" }
|
||||
if word == "else" { return "Else" }
|
||||
if word == "for" { return "For" }
|
||||
if word == "in" { return "In" }
|
||||
if word == "while" { return "While" }
|
||||
if word == "import" { return "Import" }
|
||||
if word == "from" { return "From" }
|
||||
if word == "as" { return "As" }
|
||||
if word == "with" { return "With" }
|
||||
if word == "sealed" { return "Sealed" }
|
||||
if word == "activate" { return "Activate" }
|
||||
if word == "where" { return "Where" }
|
||||
if word == "test" { return "Test" }
|
||||
if word == "seed" { return "Seed" }
|
||||
if word == "assert" { return "Assert" }
|
||||
if word == "protocol" { return "Protocol" }
|
||||
if word == "impl" { return "Impl" }
|
||||
if word == "retry" { return "Retry" }
|
||||
if word == "times" { return "Times" }
|
||||
if word == "fallback" { return "Fallback" }
|
||||
if word == "reason" { return "Reason" }
|
||||
if word == "parallel" { return "Parallel" }
|
||||
if word == "trace" { return "Trace" }
|
||||
if word == "requires" { return "Requires" }
|
||||
if word == "deploy" { return "Deploy" }
|
||||
if word == "to" { return "To" }
|
||||
if word == "via" { return "Via" }
|
||||
if word == "target" { return "Target" }
|
||||
if word == "true" { return "Bool" }
|
||||
if word == "false" { return "Bool" }
|
||||
if word == "cgi" { return "Cgi" }
|
||||
if word == "service" { return "Service" }
|
||||
if word == "manager" { return "Manager" }
|
||||
if word == "engine" { return "Engine" }
|
||||
if word == "accessor" { return "Accessor" }
|
||||
if word == "vessel" { return "Vessel" }
|
||||
if word == "extern" { return "Extern" }
|
||||
if word == "break" { return "Break" }
|
||||
if word == "continue" { return "Continue" }
|
||||
""
|
||||
}
|
||||
|
||||
// -- Scan helpers --------------------------------------------------------------
|
||||
// All scan helpers receive the chars list and total length.
|
||||
|
||||
// scan_digits - advance i while chars[i] is a digit
|
||||
// Returns { "text": ..., "pos": i }
|
||||
fn scan_digits(chars: [String], start: Int, total: Int) -> Map<String, Any> {
|
||||
let i = start
|
||||
let parts: [String] = native_list_empty()
|
||||
let running = true
|
||||
while running {
|
||||
if i >= total {
|
||||
let running = false
|
||||
} else {
|
||||
let ch: String = native_list_get(chars, i)
|
||||
if lex_is_digit(ch) {
|
||||
let parts = native_list_append(parts, ch)
|
||||
let i = i + 1
|
||||
} else {
|
||||
let running = false
|
||||
}
|
||||
}
|
||||
}
|
||||
{ "text": str_join(parts, ""), "pos": i }
|
||||
}
|
||||
|
||||
// scan_ident - advance i while chars[i] is alphanumeric or underscore
|
||||
fn scan_ident(chars: [String], start: Int, total: Int) -> Map<String, Any> {
|
||||
let i = start
|
||||
let parts: [String] = native_list_empty()
|
||||
let running = true
|
||||
while running {
|
||||
if i >= total {
|
||||
let running = false
|
||||
} else {
|
||||
let ch: String = native_list_get(chars, i)
|
||||
if is_alnum_or_underscore(ch) {
|
||||
let parts = native_list_append(parts, ch)
|
||||
let i = i + 1
|
||||
} else {
|
||||
let running = false
|
||||
}
|
||||
}
|
||||
}
|
||||
{ "text": str_join(parts, ""), "pos": i }
|
||||
}
|
||||
|
||||
// -- Code-bearing string detection + comment strip ----------------------------
|
||||
// Inline JS/CSS literals embedded in El source (e.g. <script>-</script> blobs
|
||||
// or stylesheet payloads inside string literals) carry their own line and
|
||||
// block comments. Those comments leak into the served HTML and reveal build
|
||||
// notes the visitor should never see. We strip them at the lexer so every
|
||||
// downstream consumer (codegen-c, codegen-js, parser) gets the cleaned form.
|
||||
//
|
||||
// looks_like_code - heuristic gate so we only strip strings that actually
|
||||
// embed JS or CSS. Plain prose, hex blobs, JSON, etc. pass through verbatim.
|
||||
|
||||
fn substr_at(chars: [String], start: Int, total: Int, needle: String) -> Bool {
|
||||
let nchars: [String] = native_string_chars(needle)
|
||||
let nlen: Int = native_list_len(nchars)
|
||||
if start + nlen > total { return false }
|
||||
let i = 0
|
||||
let matched = true
|
||||
while i < nlen {
|
||||
let a: String = native_list_get(chars, start + i)
|
||||
let b: String = native_list_get(nchars, i)
|
||||
if a == b { let i = i + 1 } else { let matched = false; let i = nlen }
|
||||
}
|
||||
matched
|
||||
}
|
||||
|
||||
fn str_has(s: String, needle: String) -> Bool {
|
||||
let chars: [String] = native_string_chars(s)
|
||||
let total: Int = native_list_len(chars)
|
||||
let i = 0
|
||||
let found = false
|
||||
while i < total {
|
||||
if substr_at(chars, i, total, needle) {
|
||||
let found = true
|
||||
let i = total
|
||||
} else {
|
||||
let i = i + 1
|
||||
}
|
||||
}
|
||||
found
|
||||
}
|
||||
|
||||
fn looks_like_code(s: String) -> Bool {
|
||||
if str_has(s, "<script") { return true }
|
||||
if str_has(s, "<style") { return true }
|
||||
if str_has(s, "function") {
|
||||
if str_has(s, ";") { return true }
|
||||
}
|
||||
false
|
||||
}
|
||||
|
||||
// strip_code_comments - character-by-character walk. Tracks JS string state
|
||||
// (single, double, backtick) and never strips inside one. Backslash escapes
|
||||
// inside JS strings consume the next char verbatim. URLs like https:// are
|
||||
// preserved by checking the previous char before treating // as a line
|
||||
// comment opener: if the char immediately before '/' is ':', emit the '/'
|
||||
// literally and advance one position.
|
||||
fn strip_code_comments(s: String) -> String {
|
||||
let chars: [String] = native_string_chars(s)
|
||||
let total: Int = native_list_len(chars)
|
||||
let out_parts: [String] = native_list_empty()
|
||||
let i = 0
|
||||
let in_squote = false
|
||||
let in_dquote = false
|
||||
let in_btick = false
|
||||
let prev = ""
|
||||
while i < total {
|
||||
let ch: String = native_list_get(chars, i)
|
||||
let in_js_string = false
|
||||
if in_squote { let in_js_string = true }
|
||||
if in_dquote { let in_js_string = true }
|
||||
if in_btick { let in_js_string = true }
|
||||
|
||||
if in_js_string {
|
||||
// Backslash escape: consume next char verbatim regardless of which.
|
||||
if ch == "\\" {
|
||||
let out_parts = native_list_append(out_parts, ch)
|
||||
let next_i = i + 1
|
||||
if next_i < total {
|
||||
let nc: String = native_list_get(chars, next_i)
|
||||
let out_parts = native_list_append(out_parts, nc)
|
||||
let prev = nc
|
||||
let i = next_i + 1
|
||||
} else {
|
||||
let prev = ch
|
||||
let i = next_i
|
||||
}
|
||||
} else {
|
||||
if in_squote {
|
||||
if ch == "'" { let in_squote = false }
|
||||
} else {
|
||||
if in_dquote {
|
||||
if ch == "\"" { let in_dquote = false }
|
||||
} else {
|
||||
if in_btick {
|
||||
if ch == "`" { let in_btick = false }
|
||||
}
|
||||
}
|
||||
}
|
||||
let out_parts = native_list_append(out_parts, ch)
|
||||
let prev = ch
|
||||
let i = i + 1
|
||||
}
|
||||
} else {
|
||||
// Not in a JS string. Check for comment openers.
|
||||
let next_i = i + 1
|
||||
let next_ch = ""
|
||||
if next_i < total {
|
||||
let next_ch: String = native_list_get(chars, next_i)
|
||||
}
|
||||
|
||||
if ch == "/" {
|
||||
if next_ch == "/" {
|
||||
// URL guard: prev char ':' means this is "://", not a comment.
|
||||
if prev == ":" {
|
||||
let out_parts = native_list_append(out_parts, ch)
|
||||
let prev = ch
|
||||
let i = i + 1
|
||||
} else {
|
||||
// Skip until newline (newline itself is preserved so
|
||||
// surrounding line counts/structure stay sane).
|
||||
let i = i + 2
|
||||
let scanning = true
|
||||
while scanning {
|
||||
if i >= total {
|
||||
let scanning = false
|
||||
} else {
|
||||
let lc: String = native_list_get(chars, i)
|
||||
if lc == "\n" {
|
||||
let scanning = false
|
||||
} else {
|
||||
let i = i + 1
|
||||
}
|
||||
}
|
||||
}
|
||||
let prev = ""
|
||||
}
|
||||
} else {
|
||||
if next_ch == "*" {
|
||||
// Skip until matching "*/".
|
||||
let i = i + 2
|
||||
let scanning2 = true
|
||||
while scanning2 {
|
||||
if i >= total {
|
||||
let scanning2 = false
|
||||
} else {
|
||||
let bc: String = native_list_get(chars, i)
|
||||
if bc == "*" {
|
||||
let after = i + 1
|
||||
if after < total {
|
||||
let nc2: String = native_list_get(chars, after)
|
||||
if nc2 == "/" {
|
||||
let i = after + 1
|
||||
let scanning2 = false
|
||||
} else {
|
||||
let i = i + 1
|
||||
}
|
||||
} else {
|
||||
let i = i + 1
|
||||
}
|
||||
} else {
|
||||
let i = i + 1
|
||||
}
|
||||
}
|
||||
}
|
||||
let prev = ""
|
||||
} else {
|
||||
let out_parts = native_list_append(out_parts, ch)
|
||||
let prev = ch
|
||||
let i = i + 1
|
||||
}
|
||||
}
|
||||
} else {
|
||||
// Open a JS string?
|
||||
if ch == "'" {
|
||||
let in_squote = true
|
||||
let out_parts = native_list_append(out_parts, ch)
|
||||
let prev = ch
|
||||
let i = i + 1
|
||||
} else {
|
||||
if ch == "\"" {
|
||||
let in_dquote = true
|
||||
let out_parts = native_list_append(out_parts, ch)
|
||||
let prev = ch
|
||||
let i = i + 1
|
||||
} else {
|
||||
if ch == "`" {
|
||||
let in_btick = true
|
||||
let out_parts = native_list_append(out_parts, ch)
|
||||
let prev = ch
|
||||
let i = i + 1
|
||||
} else {
|
||||
let out_parts = native_list_append(out_parts, ch)
|
||||
let prev = ch
|
||||
let i = i + 1
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
str_join(out_parts, "")
|
||||
}
|
||||
|
||||
// scan_string - scan a quoted string literal, handling \" escapes.
|
||||
// Starts AFTER the opening quote. Returns { "text": content, "pos": i_after_close }
|
||||
fn scan_string(chars: [String], start: Int, total: Int) -> Map<String, Any> {
|
||||
let i = start
|
||||
let parts: [String] = native_list_empty()
|
||||
let running = true
|
||||
while running {
|
||||
if i >= total {
|
||||
let running = false
|
||||
} else {
|
||||
let ch: String = native_list_get(chars, i)
|
||||
if ch == "\\" {
|
||||
// escape: peek next char
|
||||
let next_i = i + 1
|
||||
if next_i < total {
|
||||
let next_ch: String = native_list_get(chars, next_i)
|
||||
if next_ch == "\"" {
|
||||
let parts = native_list_append(parts, "\"")
|
||||
let i = next_i + 1
|
||||
} else {
|
||||
if next_ch == "n" {
|
||||
let parts = native_list_append(parts, "\n")
|
||||
let i = next_i + 1
|
||||
} else {
|
||||
if next_ch == "t" {
|
||||
let parts = native_list_append(parts, "\t")
|
||||
let i = next_i + 1
|
||||
} else {
|
||||
if next_ch == "r" {
|
||||
let parts = native_list_append(parts, "\r")
|
||||
let i = next_i + 1
|
||||
} else {
|
||||
if next_ch == "\\" {
|
||||
let parts = native_list_append(parts, "\\")
|
||||
let i = next_i + 1
|
||||
} else {
|
||||
let parts = native_list_append(parts, next_ch)
|
||||
let i = next_i + 1
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
} else {
|
||||
let i = i + 1
|
||||
}
|
||||
} else {
|
||||
if ch == "\"" {
|
||||
let i = i + 1
|
||||
let running = false
|
||||
} else {
|
||||
let parts = native_list_append(parts, ch)
|
||||
let i = i + 1
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
{ "text": str_join(parts, ""), "pos": i }
|
||||
}
|
||||
|
||||
// -- String interpolation ------------------------------------------------------
|
||||
//
|
||||
// scan_interp_brace - scan from `start` (the char after `${`) to the matching
|
||||
// `}`, tracking brace depth so inner braces (e.g. fn calls, map literals) are
|
||||
// handled correctly. Returns { "text": inner_source, "pos": i_after_close }.
|
||||
fn scan_interp_brace(chars: [String], start: Int, total: Int) -> Map<String, Any> {
|
||||
let i = start
|
||||
let parts: [String] = native_list_empty()
|
||||
let depth = 1
|
||||
let running = true
|
||||
while running {
|
||||
if i >= total {
|
||||
let running = false
|
||||
} else {
|
||||
let ch: String = native_list_get(chars, i)
|
||||
if ch == "{" {
|
||||
let depth = depth + 1
|
||||
let parts = native_list_append(parts, ch)
|
||||
let i = i + 1
|
||||
} else {
|
||||
if ch == "}" {
|
||||
let depth = depth - 1
|
||||
if depth <= 0 {
|
||||
// Closing brace of the interpolation - stop, do not include it
|
||||
let i = i + 1
|
||||
let running = false
|
||||
} else {
|
||||
let parts = native_list_append(parts, ch)
|
||||
let i = i + 1
|
||||
}
|
||||
} else {
|
||||
let parts = native_list_append(parts, ch)
|
||||
let i = i + 1
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
{ "text": str_join(parts, ""), "pos": i }
|
||||
}
|
||||
|
||||
// interp_tokens_append_all - copy every token from src into dst, skipping the
|
||||
// trailing Eof sentinel that lex() always appends. Returns the updated dst list.
|
||||
fn interp_tokens_append_all(dst: [Map<String, Any>], src: [Map<String, Any>]) -> [Map<String, Any>] {
|
||||
let src_len: Int = native_list_len(src)
|
||||
let j = 0
|
||||
let result = dst
|
||||
while j < src_len {
|
||||
let tok: Map<String, Any> = native_list_get(src, j)
|
||||
let tk: String = tok["kind"]
|
||||
if tk == "Eof" {
|
||||
let j = src_len
|
||||
} else {
|
||||
let result = native_list_append(result, tok)
|
||||
let j = j + 1
|
||||
}
|
||||
}
|
||||
result
|
||||
}
|
||||
|
||||
// scan_interp_string - scan a string literal that may contain ${expr}
|
||||
// interpolations. Starts AFTER the opening `"`.
|
||||
// Returns { "tokens": [token list to inject], "pos": i_after_close_quote }.
|
||||
//
|
||||
// For a plain string (no ${}) this emits a single Str token, identical to the
|
||||
// old scan_string path. For an interpolated string it emits a flat sequence
|
||||
// of tokens equivalent to the string-concat expression, for example:
|
||||
//
|
||||
// "hello ${name}!"
|
||||
// => Str("hello ") Plus <tokens for name> Plus Str("!")
|
||||
//
|
||||
// Empty literal segments between adjacent ${ } blocks are omitted. The
|
||||
// resulting token stream is consumed by the existing parse_binop / parse_primary
|
||||
// path in the parser with zero parser changes required.
|
||||
//
|
||||
// Supported escape sequences: \" \n \t \r \\ \$ (literal dollar sign).
|
||||
// Nested quotes inside ${} are not supported; use a variable instead.
|
||||
fn scan_interp_string(chars: [String], start: Int, total: Int) -> Map<String, Any> {
|
||||
let i = start
|
||||
let out_tokens: [Map<String, Any>] = native_list_empty()
|
||||
let cur_part: [String] = native_list_empty()
|
||||
let has_interp = false
|
||||
let need_plus = false
|
||||
let running = true
|
||||
|
||||
while running {
|
||||
if i >= total {
|
||||
let running = false
|
||||
} else {
|
||||
let ch: String = native_list_get(chars, i)
|
||||
|
||||
if ch == "\\" {
|
||||
// Escape sequence
|
||||
let next_i = i + 1
|
||||
if next_i < total {
|
||||
let next_ch: String = native_list_get(chars, next_i)
|
||||
if next_ch == "$" {
|
||||
// \$ => literal '$' (escape for interpolation syntax)
|
||||
let cur_part = native_list_append(cur_part, "$")
|
||||
let i = next_i + 1
|
||||
} else {
|
||||
if next_ch == "\"" {
|
||||
let cur_part = native_list_append(cur_part, "\"")
|
||||
let i = next_i + 1
|
||||
} else {
|
||||
if next_ch == "n" {
|
||||
let cur_part = native_list_append(cur_part, "\n")
|
||||
let i = next_i + 1
|
||||
} else {
|
||||
if next_ch == "t" {
|
||||
let cur_part = native_list_append(cur_part, "\t")
|
||||
let i = next_i + 1
|
||||
} else {
|
||||
if next_ch == "r" {
|
||||
let cur_part = native_list_append(cur_part, "\r")
|
||||
let i = next_i + 1
|
||||
} else {
|
||||
if next_ch == "\\" {
|
||||
let cur_part = native_list_append(cur_part, "\\")
|
||||
let i = next_i + 1
|
||||
} else {
|
||||
let cur_part = native_list_append(cur_part, next_ch)
|
||||
let i = next_i + 1
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
} else {
|
||||
let i = i + 1
|
||||
}
|
||||
} else {
|
||||
if ch == "\"" {
|
||||
// Closing quote - stop scanning
|
||||
let i = i + 1
|
||||
let running = false
|
||||
} else {
|
||||
if ch == "$" {
|
||||
// Check for ${ (start of interpolation)
|
||||
let next_i = i + 1
|
||||
let is_interp = false
|
||||
if next_i < total {
|
||||
let next_ch: String = native_list_get(chars, next_i)
|
||||
if next_ch == "{" {
|
||||
let is_interp = true
|
||||
}
|
||||
}
|
||||
if is_interp {
|
||||
// Flush the accumulated literal part (if non-empty)
|
||||
let part_len: Int = native_list_len(cur_part)
|
||||
if part_len > 0 {
|
||||
let part_text = str_join(cur_part, "")
|
||||
if need_plus {
|
||||
let out_tokens = native_list_append(out_tokens, make_tok("Plus", "+"))
|
||||
}
|
||||
let clean_part = part_text
|
||||
if looks_like_code(part_text) {
|
||||
let clean_part = strip_code_comments(part_text)
|
||||
}
|
||||
let out_tokens = native_list_append(out_tokens, make_tok("Str", clean_part))
|
||||
let need_plus = true
|
||||
}
|
||||
let cur_part = native_list_empty()
|
||||
let has_interp = true
|
||||
|
||||
// Scan brace-balanced expression source
|
||||
let brace_result = scan_interp_brace(chars, next_i + 1, total)
|
||||
let expr_src: String = brace_result["text"]
|
||||
let new_i: Int = brace_result["pos"]
|
||||
let i = new_i
|
||||
|
||||
// Re-lex the expression and inline the tokens.
|
||||
// Wrap in ( ) so that operators inside ${} (e.g.
|
||||
// age + 1) are parsed as a grouped sub-expression
|
||||
// rather than merging with the surrounding concat
|
||||
// Plus tokens at the wrong precedence level.
|
||||
let inner_toks: [Map<String, Any>] = lex(expr_src)
|
||||
let inner_len: Int = native_list_len(inner_toks)
|
||||
|
||||
if need_plus {
|
||||
let out_tokens = native_list_append(out_tokens, make_tok("Plus", "+"))
|
||||
}
|
||||
// Empty interpolation ${} => empty string segment
|
||||
if inner_len <= 1 {
|
||||
let out_tokens = native_list_append(out_tokens, make_tok("Str", ""))
|
||||
} else {
|
||||
let out_tokens = native_list_append(out_tokens, make_tok("LParen", "("))
|
||||
let out_tokens = interp_tokens_append_all(out_tokens, inner_toks)
|
||||
let out_tokens = native_list_append(out_tokens, make_tok("RParen", ")"))
|
||||
}
|
||||
let need_plus = true
|
||||
} else {
|
||||
// Plain '$' not followed by '{' - treat as literal
|
||||
let cur_part = native_list_append(cur_part, "$")
|
||||
let i = i + 1
|
||||
}
|
||||
} else {
|
||||
let cur_part = native_list_append(cur_part, ch)
|
||||
let i = i + 1
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// Flush remaining literal segment and build final token list
|
||||
let part_text = str_join(cur_part, "")
|
||||
let part_len: Int = native_list_len(cur_part)
|
||||
if has_interp {
|
||||
// Interpolated string: only emit trailing segment if non-empty
|
||||
if part_len > 0 {
|
||||
let clean_part = part_text
|
||||
if looks_like_code(part_text) {
|
||||
let clean_part = strip_code_comments(part_text)
|
||||
}
|
||||
if need_plus {
|
||||
let out_tokens = native_list_append(out_tokens, make_tok("Plus", "+"))
|
||||
}
|
||||
let out_tokens = native_list_append(out_tokens, make_tok("Str", clean_part))
|
||||
}
|
||||
} else {
|
||||
// Plain string with no interpolation - same behaviour as old scan_string
|
||||
let clean_text = part_text
|
||||
if looks_like_code(part_text) {
|
||||
let clean_text = strip_code_comments(part_text)
|
||||
}
|
||||
let out_tokens = native_list_append(out_tokens, make_tok("Str", clean_text))
|
||||
}
|
||||
|
||||
{ "tokens": out_tokens, "pos": i }
|
||||
}
|
||||
|
||||
// -- Main lexer ----------------------------------------------------------------
|
||||
|
||||
fn lex(source: String) -> [Map<String, Any>] {
|
||||
let chars: [String] = native_string_chars(source)
|
||||
let total: Int = native_list_len(chars)
|
||||
let tokens: [Map<String, Any>] = native_list_empty()
|
||||
let i: Int = 0
|
||||
|
||||
while i < total {
|
||||
let ch: String = native_list_get(chars, i)
|
||||
|
||||
// Skip whitespace
|
||||
if lex_is_whitespace(ch) {
|
||||
let i = i + 1
|
||||
} else {
|
||||
// Line comments: //
|
||||
if ch == "/" {
|
||||
let next_i = i + 1
|
||||
if next_i < total {
|
||||
let next_ch: String = native_list_get(chars, next_i)
|
||||
if next_ch == "/" {
|
||||
// skip to end of line
|
||||
let i = i + 2
|
||||
let running2 = true
|
||||
while running2 {
|
||||
if i >= total {
|
||||
let running2 = false
|
||||
} else {
|
||||
let lch: String = native_list_get(chars, i)
|
||||
if lch == "\n" {
|
||||
let running2 = false
|
||||
} else {
|
||||
let i = i + 1
|
||||
}
|
||||
}
|
||||
}
|
||||
} else {
|
||||
let tokens = native_list_append(tokens, make_tok("Slash", "/"))
|
||||
let i = i + 1
|
||||
}
|
||||
} else {
|
||||
let tokens = native_list_append(tokens, make_tok("Slash", "/"))
|
||||
let i = i + 1
|
||||
}
|
||||
} else {
|
||||
// String literal (plain or interpolated with ${expr} syntax).
|
||||
// scan_interp_string handles both cases: plain strings emit a
|
||||
// single Str token; interpolated strings emit a flat token
|
||||
// sequence (Str Plus expr-tokens Plus Str ...) that the parser
|
||||
// naturally assembles into a BinOp concat tree.
|
||||
if ch == "\"" {
|
||||
let interp_result = scan_interp_string(chars, i + 1, total)
|
||||
let interp_toks: [Map<String, Any>] = interp_result["tokens"]
|
||||
let new_pos: Int = interp_result["pos"]
|
||||
let tokens = interp_tokens_append_all(tokens, interp_toks)
|
||||
let i = new_pos
|
||||
} else {
|
||||
// Number literal
|
||||
if lex_is_digit(ch) {
|
||||
let result = scan_digits(chars, i, total)
|
||||
let num_text: String = result["text"]
|
||||
let new_pos: Int = result["pos"]
|
||||
// check for float (dot followed by digit)
|
||||
if new_pos < total {
|
||||
let dot_ch: String = native_list_get(chars, new_pos)
|
||||
if dot_ch == "." {
|
||||
let after_dot = new_pos + 1
|
||||
if after_dot < total {
|
||||
let after_dot_ch: String = native_list_get(chars, after_dot)
|
||||
if lex_is_digit(after_dot_ch) {
|
||||
let frac_result = scan_digits(chars, after_dot, total)
|
||||
let frac_text: String = frac_result["text"]
|
||||
let frac_pos: Int = frac_result["pos"]
|
||||
let tokens = native_list_append(tokens, make_tok("Float", num_text + "." + frac_text))
|
||||
let i = frac_pos
|
||||
} else {
|
||||
let tokens = native_list_append(tokens, make_tok("Int", num_text))
|
||||
let i = new_pos
|
||||
}
|
||||
} else {
|
||||
let tokens = native_list_append(tokens, make_tok("Int", num_text))
|
||||
let i = new_pos
|
||||
}
|
||||
} else {
|
||||
let tokens = native_list_append(tokens, make_tok("Int", num_text))
|
||||
let i = new_pos
|
||||
}
|
||||
} else {
|
||||
let tokens = native_list_append(tokens, make_tok("Int", num_text))
|
||||
let i = new_pos
|
||||
}
|
||||
} else {
|
||||
// Identifier or keyword
|
||||
if lex_is_alpha(ch) || ch == "_" {
|
||||
let result = scan_ident(chars, i, total)
|
||||
let word: String = result["text"]
|
||||
let new_pos: Int = result["pos"]
|
||||
let kw = keyword_kind(word)
|
||||
if kw == "" {
|
||||
let tokens = native_list_append(tokens, make_tok("Ident", word))
|
||||
} else {
|
||||
let tokens = native_list_append(tokens, make_tok(kw, word))
|
||||
}
|
||||
let i = new_pos
|
||||
} else {
|
||||
// Multi-char and single-char operators/delimiters
|
||||
let peek_i = i + 1
|
||||
let peek_ch = ""
|
||||
if peek_i < total {
|
||||
let peek_ch: String = native_list_get(chars, peek_i)
|
||||
}
|
||||
|
||||
if ch == "=" {
|
||||
if peek_ch == "=" {
|
||||
let tokens = native_list_append(tokens, make_tok("EqEq", "=="))
|
||||
let i = i + 2
|
||||
} else {
|
||||
if peek_ch == ">" {
|
||||
let tokens = native_list_append(tokens, make_tok("FatArrow", "=>"))
|
||||
let i = i + 2
|
||||
} else {
|
||||
let tokens = native_list_append(tokens, make_tok("Eq", "="))
|
||||
let i = i + 1
|
||||
}
|
||||
}
|
||||
} else {
|
||||
if ch == "!" {
|
||||
if peek_ch == "=" {
|
||||
let tokens = native_list_append(tokens, make_tok("NotEq", "!="))
|
||||
let i = i + 2
|
||||
} else {
|
||||
let tokens = native_list_append(tokens, make_tok("Not", "!"))
|
||||
let i = i + 1
|
||||
}
|
||||
} else {
|
||||
if ch == "<" {
|
||||
if peek_ch == "=" {
|
||||
let tokens = native_list_append(tokens, make_tok("LtEq", "<="))
|
||||
let i = i + 2
|
||||
} else {
|
||||
let tokens = native_list_append(tokens, make_tok("Lt", "<"))
|
||||
let i = i + 1
|
||||
}
|
||||
} else {
|
||||
if ch == ">" {
|
||||
if peek_ch == "=" {
|
||||
let tokens = native_list_append(tokens, make_tok("GtEq", ">="))
|
||||
let i = i + 2
|
||||
} else {
|
||||
let tokens = native_list_append(tokens, make_tok("Gt", ">"))
|
||||
let i = i + 1
|
||||
}
|
||||
} else {
|
||||
if ch == "&" {
|
||||
if peek_ch == "&" {
|
||||
let tokens = native_list_append(tokens, make_tok("And", "&&"))
|
||||
let i = i + 2
|
||||
} else {
|
||||
let i = i + 1
|
||||
}
|
||||
} else {
|
||||
if ch == "|" {
|
||||
if peek_ch == "|" {
|
||||
let tokens = native_list_append(tokens, make_tok("Or", "||"))
|
||||
let i = i + 2
|
||||
} else {
|
||||
if peek_ch == ">" {
|
||||
let tokens = native_list_append(tokens, make_tok("PipeOp", "|>"))
|
||||
let i = i + 2
|
||||
} else {
|
||||
let tokens = native_list_append(tokens, make_tok("Pipe", "|"))
|
||||
let i = i + 1
|
||||
}
|
||||
}
|
||||
} else {
|
||||
if ch == "-" {
|
||||
if peek_ch == ">" {
|
||||
let tokens = native_list_append(tokens, make_tok("Arrow", "->"))
|
||||
let i = i + 2
|
||||
} else {
|
||||
let tokens = native_list_append(tokens, make_tok("Minus", "-"))
|
||||
let i = i + 1
|
||||
}
|
||||
} else {
|
||||
if ch == ":" {
|
||||
if peek_ch == ":" {
|
||||
let tokens = native_list_append(tokens, make_tok("ColonColon", "::"))
|
||||
let i = i + 2
|
||||
} else {
|
||||
let tokens = native_list_append(tokens, make_tok("Colon", ":"))
|
||||
let i = i + 1
|
||||
}
|
||||
} else {
|
||||
if ch == "+" {
|
||||
let tokens = native_list_append(tokens, make_tok("Plus", "+"))
|
||||
let i = i + 1
|
||||
} else {
|
||||
if ch == "*" {
|
||||
let tokens = native_list_append(tokens, make_tok("Star", "*"))
|
||||
let i = i + 1
|
||||
} else {
|
||||
if ch == "%" {
|
||||
let tokens = native_list_append(tokens, make_tok("Percent", "%"))
|
||||
let i = i + 1
|
||||
} else {
|
||||
if ch == "(" {
|
||||
let tokens = native_list_append(tokens, make_tok("LParen", "("))
|
||||
let i = i + 1
|
||||
} else {
|
||||
if ch == ")" {
|
||||
let tokens = native_list_append(tokens, make_tok("RParen", ")"))
|
||||
let i = i + 1
|
||||
} else {
|
||||
if ch == "{" {
|
||||
let tokens = native_list_append(tokens, make_tok("LBrace", "{"))
|
||||
let i = i + 1
|
||||
} else {
|
||||
if ch == "}" {
|
||||
let tokens = native_list_append(tokens, make_tok("RBrace", "}"))
|
||||
let i = i + 1
|
||||
} else {
|
||||
if ch == "[" {
|
||||
let tokens = native_list_append(tokens, make_tok("LBracket", "["))
|
||||
let i = i + 1
|
||||
} else {
|
||||
if ch == "]" {
|
||||
let tokens = native_list_append(tokens, make_tok("RBracket", "]"))
|
||||
let i = i + 1
|
||||
} else {
|
||||
if ch == "," {
|
||||
let tokens = native_list_append(tokens, make_tok("Comma", ","))
|
||||
let i = i + 1
|
||||
} else {
|
||||
if ch == "." {
|
||||
// Check for ..= (inclusive range) before .. (exclusive range) before single .
|
||||
let peek2_i = i + 2
|
||||
let peek2_ch = ""
|
||||
if peek2_i < total {
|
||||
let peek2_ch: String = native_list_get(chars, peek2_i)
|
||||
}
|
||||
if peek_ch == "." {
|
||||
if peek2_ch == "=" {
|
||||
let tokens = native_list_append(tokens, make_tok("DotDotEq", "..="))
|
||||
let i = i + 3
|
||||
} else {
|
||||
let tokens = native_list_append(tokens, make_tok("DotDot", ".."))
|
||||
let i = i + 2
|
||||
}
|
||||
} else {
|
||||
let tokens = native_list_append(tokens, make_tok("Dot", "."))
|
||||
let i = i + 1
|
||||
}
|
||||
} else {
|
||||
if ch == ";" {
|
||||
let tokens = native_list_append(tokens, make_tok("Semicolon", ";"))
|
||||
let i = i + 1
|
||||
} else {
|
||||
if ch == "@" {
|
||||
let tokens = native_list_append(tokens, make_tok("At", "@"))
|
||||
let i = i + 1
|
||||
} else {
|
||||
if ch == "?" {
|
||||
let tokens = native_list_append(tokens, make_tok("QuestionMark", "?"))
|
||||
let i = i + 1
|
||||
} else {
|
||||
// unknown char - skip
|
||||
let i = i + 1
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
let tokens = native_list_append(tokens, make_tok("Eof", ""))
|
||||
tokens
|
||||
}
|
||||
File diff suppressed because it is too large
Load Diff
Reference in New Issue
Block a user