Compare commits

..

9 Commits

Author SHA1 Message Date
will.anderson f5dcca0386 build: update dist/platform/elc with OOM fix and memory guard
El SDK CI - dev / build-and-test (pull_request) Successful in 4m16s
Rebuilt from fix/elc-oom-checkout: scan_fn_sigs_el() --emit-header path
+ el_mem_check() guard. Verified on checkout.el: all 3 sigs in .elh,
clean exit under normal load, exit(1) on memory limit exceeded.
2026-05-08 08:23:07 -05:00
will.anderson 53e0b99d5f fix(elc): add el_mem_check() memory guard — abort before OS OOM-kill
Add el_mem_check() to el_runtime.c: reads ELC_MAX_MEM_MB (default 512),
checks RSS via getrusage (macOS bytes / Linux KB normalised to MB), prints
a clear diagnostic to stderr and exits(1) if exceeded.

Wire it into two places:
- compiler.el: upfront check at --emit-header entry point
- codegen.el: per-function check in the streaming loop after each
  el_arena_pop, so runaway growth is caught at the earliest function
  boundary rather than after the machine is already dying.
2026-05-08 08:21:38 -05:00
will.anderson 5f9cad5908 fix(elc): eliminate OOM in --emit-header by using token-level signature scan
The --emit-header path previously called parse() which builds the entire
program AST in memory before writing the .elh file. For checkout.el (~491
lines with HTML template trees and deep BinOp string-concat chains), this
exhausted memory before the header could be written.

Fix: replace parse() + emit_header() with scan_fn_sigs_el() +
emit_header_from_sigs(). The new path tokenises the source once, then
walks the flat token list skipping over function bodies entirely — peak
memory is O(tokens) instead of O(whole-program AST).

New functions in parser.el:
- scan_type_el: reads a type annotation and returns its El source string
- scan_params_el: reads (name: Type, ...) and returns El params string
- scan_fn_sigs_el: token-level scan that collects El-style fn signatures
  without building any expression AST nodes

New function in compiler.el:
- emit_header_from_sigs: writes .elh from scan_fn_sigs_el output

Self-hosting check: elc compiled with new elc, diff of outputs is
identical (zero difference).

Smoke test: elc --emit-header checkout.el produces correct three-entry
.elh (previously truncated at two entries due to mid-parse OOM).
2026-05-08 08:20:13 -05:00
will.anderson 00629b39c4 Merge pull request 'fix(parser): str_join separator '' not ' ' — CSS selectors were emitting spaces' (#45) from fix/css-str-join-separator into dev
El SDK CI - dev / build-and-test (push) Failing after 12m6s
2026-05-07 23:00:19 +00:00
will.anderson ca1e4d57b8 Merge pull request 'ci: add three-tier ci-base rebuild (dev/stage)' (#44) from fix/html-template-if-style-script into dev
El SDK CI - dev / build-and-test (push) Has been cancelled
2026-05-07 23:00:13 +00:00
will.anderson f971e96dd5 fix(parser): str_join separator '' not ' ' — CSS selectors were emitting spaces between tokens
El SDK CI - dev / build-and-test (pull_request) Successful in 3m45s
2026-05-07 15:53:19 -05:00
will.anderson 81a1a624f1 add three-tier ci-base rebuild (dev/stage) to CI workflows
El SDK CI - dev / build-and-test (pull_request) Successful in 3m49s
2026-05-07 15:51:24 -05:00
will.anderson 7b7f9f353b Merge pull request 'fix(parser): add {#if}/{#else}/{/if} and raw-text <style>/<script> in HTML templates' (#43) from fix/html-template-if-style-script into dev
El SDK CI - dev / build-and-test (push) Successful in 4m28s
fix(parser): add {#if}/{#else}/{/if} and raw-text <style>/<script> in HTML templates
2026-05-07 18:44:26 +00:00
will.anderson a3732a1e9a fix(parser): add {#if}/{#else}/{/if} support and raw-text <style>/<script> in HTML templates
El SDK CI - dev / build-and-test (pull_request) Failing after 18m3s
The El lexer silently skips '#', so {#each} lexes as LBrace Ident:"each"
and {#if} lexes as LBrace If ... (using the If keyword token, not Hash).
The existing {#each} check used k2=="Hash" which was dead code.

Parser changes (parser.el):
- Add parse_raw_text_content(): collects all tokens as raw text until
  </tag_name>, bypassing El expression parsing. Used for <style> and
  <script> elements so CSS/JS content isn't parsed as El expressions.
- parse_html_element(): use raw-text mode for <style> and <script> tags.
- parse_html_children(): fix {#each} detection (k2=="Ident", k3=="each"
  instead of dead k2=="Hash" check). Add {#if cond}...{#else}...{/if}
  support generating HtmlIf AST nodes.

Codegen changes (codegen.el):
- Add cg_html_if(): generates if (cond_c) { then_c } else { else_c }
  for HtmlIf nodes.
- cg_html_parts(): dispatch HtmlIf to cg_html_if.
2026-05-07 13:39:12 -05:00
10 changed files with 713 additions and 287 deletions
+47
View File
@@ -252,4 +252,51 @@ jobs:
--source=el-compiler/runtime/el_runtime.js
echo "Published El SDK version=${VERSION} to foundation-dev"
# Keep key alive for the ci-base rebuild step below
# (deleted in that step after docker push)
- name: Rebuild ci-base with fresh El SDK (dev)
# Patches ci-base:dev in-place: pulls the existing image (which has all
# system deps — Node, Go, gcloud, Docker CLI, etc.) and overlays the freshly
# built El SDK on top. Keeps the full ci-base rebuild fast and incremental.
if: github.event_name == 'push'
env:
GCP_SA_KEY: ${{ secrets.GCP_SA_KEY }}
run: |
set -euo pipefail
CI_BASE="us-central1-docker.pkg.dev/neuron-785695/neuron-ci/ci-base"
SHA="${GITHUB_SHA:0:8}"
echo "${GCP_SA_KEY}" > /tmp/gcp-key.json
gcloud auth activate-service-account --key-file=/tmp/gcp-key.json
gcloud config set project neuron-785695
gcloud auth configure-docker us-central1-docker.pkg.dev --quiet
# Pull existing ci-base:dev (system deps stay cached in the base layer)
docker pull "${CI_BASE}:dev" || docker pull "${CI_BASE}:latest"
# Inline Dockerfile — only replaces the El SDK layer
cat > /tmp/Dockerfile.ci-base-patch << 'EOF'
ARG BASE
FROM ${BASE}
COPY dist/platform/elc /opt/el/dist/platform/elc
COPY dist/bin/elb /opt/el/dist/bin/elb
COPY el-compiler/runtime/el_runtime.c /opt/el/el-compiler/runtime/el_runtime.c
COPY el-compiler/runtime/el_runtime.h /opt/el/el-compiler/runtime/el_runtime.h
COPY el-compiler/runtime/el_runtime.js /opt/el/el-compiler/runtime/el_runtime.js
RUN chmod +x /opt/el/dist/platform/elc /opt/el/dist/bin/elb
EOF
docker build \
--build-arg BASE="${CI_BASE}:dev" \
--build-arg BUILDKIT_INLINE_CACHE=1 \
-f /tmp/Dockerfile.ci-base-patch \
-t "${CI_BASE}:dev" \
-t "${CI_BASE}:dev-${SHA}" \
.
docker push "${CI_BASE}:dev"
docker push "${CI_BASE}:dev-${SHA}"
echo "ci-base rebuilt: ${CI_BASE}:dev (${SHA})"
rm -f /tmp/gcp-key.json
+47
View File
@@ -246,4 +246,51 @@ jobs:
--source=el-compiler/runtime/el_runtime.h
echo "Published El SDK version=${VERSION} to foundation-stage"
# Keep key alive for the ci-base rebuild step below
# (deleted in that step after docker push)
- name: Rebuild ci-base with fresh El SDK (stage)
# Patches ci-base:stage in-place: pulls the existing image (which has all
# system deps — Node, Go, gcloud, Docker CLI, etc.) and overlays the freshly
# built El SDK on top. Keeps the full ci-base rebuild fast and incremental.
if: github.event_name == 'push'
env:
GCP_SA_KEY: ${{ secrets.GCP_SA_KEY }}
run: |
set -euo pipefail
CI_BASE="us-central1-docker.pkg.dev/neuron-785695/neuron-ci/ci-base"
SHA="${GITHUB_SHA:0:8}"
echo "${GCP_SA_KEY}" > /tmp/gcp-key.json
gcloud auth activate-service-account --key-file=/tmp/gcp-key.json
gcloud config set project neuron-785695
gcloud auth configure-docker us-central1-docker.pkg.dev --quiet
# Pull existing ci-base:stage (system deps stay cached in the base layer)
docker pull "${CI_BASE}:stage" || docker pull "${CI_BASE}:latest"
# Inline Dockerfile — only replaces the El SDK layer
cat > /tmp/Dockerfile.ci-base-patch << 'EOF'
ARG BASE
FROM ${BASE}
COPY dist/platform/elc /opt/el/dist/platform/elc
COPY dist/bin/elb /opt/el/dist/bin/elb
COPY el-compiler/runtime/el_runtime.c /opt/el/el-compiler/runtime/el_runtime.c
COPY el-compiler/runtime/el_runtime.h /opt/el/el-compiler/runtime/el_runtime.h
COPY el-compiler/runtime/el_runtime.js /opt/el/el-compiler/runtime/el_runtime.js
RUN chmod +x /opt/el/dist/platform/elc /opt/el/dist/bin/elb
EOF
docker build \
--build-arg BASE="${CI_BASE}:stage" \
--build-arg BUILDKIT_INLINE_CACHE=1 \
-f /tmp/Dockerfile.ci-base-patch \
-t "${CI_BASE}:stage" \
-t "${CI_BASE}:stage-${SHA}" \
.
docker push "${CI_BASE}:stage"
docker push "${CI_BASE}:stage-${SHA}"
echo "ci-base rebuilt: ${CI_BASE}:stage (${SHA})"
rm -f /tmp/gcp-key.json
BIN
View File
Binary file not shown.
@@ -1,94 +0,0 @@
#ifndef EL_PLATFORM_WIN_H
#define EL_PLATFORM_WIN_H
/*
* el_platform_win.h — Windows OS-boundary shim for el_runtime.c.
*
* Branch: feat/windows-el-runtime. Included ONLY when _WIN32 is defined; the POSIX build is
* untouched. Goal: let el_runtime.c (a BSD-sockets / dlfcn / fork host) compile and link with
* mingw-w64 into a native neuron.exe, with no behavioural change to the Linux/macOS build.
*
* What it maps:
* - sockets : winsock2 (same call names: socket/bind/listen/accept/recv/send/setsockopt).
* Sockets close with closesocket() (see el_closesocket), and the stack must be
* started once with WSAStartup — done automatically via a load-time constructor.
* - dlsym : el_runtime.c uses dlsym(RTLD_DEFAULT, name) to resolve callback/tool symbols
* exported by the main module. Windows equivalent: GetProcAddress on the process
* module. Link the soul with -Wl,--export-all-symbols so the symbols are findable.
* - popen : mapped to _popen/_pclose.
* - threads : UNCHANGED. mingw-w64 ships winpthreads, so <pthread.h> + -lpthread just work.
*/
#ifndef WIN32_LEAN_AND_MEAN
#define WIN32_LEAN_AND_MEAN
#endif
#include <winsock2.h>
#include <ws2tcpip.h>
#include <windows.h>
#include <io.h>
#include <process.h>
/* Portable headers mingw-w64 provides (verified present). */
#include <stdarg.h>
#include <stdint.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <strings.h> /* strcasecmp */
#include <ctype.h>
#include <math.h>
#include <time.h>
#include <sys/time.h> /* mingw-w64 provides gettimeofday here */
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <dirent.h>
#include <errno.h>
#include <pthread.h>
/* ── socket close ─────────────────────────────────────────────────────────── */
/* Winsock closes sockets with closesocket(), not close() (close() is for file fds). The POSIX
build defines the same helper as close() so the call sites are identical across platforms. */
static inline int el_closesocket(int s) { return closesocket((SOCKET)s); }
/* ── winsock init (once, at load) ─────────────────────────────────────────── */
static void el__win_net_init(void) {
static int inited = 0;
if (!inited) { WSADATA w; WSAStartup(MAKEWORD(2, 2), &w); inited = 1; }
}
__attribute__((constructor)) static void el__win_ctor(void) { el__win_net_init(); }
/* ── dlsym → GetProcAddress ───────────────────────────────────────────────── */
#ifndef RTLD_DEFAULT
#define RTLD_DEFAULT ((void*)0)
#endif
static inline void* el_win_dlsym(void* handle, const char* name) {
(void)handle;
return (void*)(uintptr_t)GetProcAddress(GetModuleHandleA(NULL), name);
}
#define dlsym(h, n) el_win_dlsym((h), (n))
/* ── popen / pclose ───────────────────────────────────────────────────────── */
#define popen _popen
#define pclose _pclose
/* ── misc POSIX → Win32 shims ─────────────────────────────────────────────── */
#include <direct.h> /* _mkdir */
#define mkdir(path, mode) _mkdir(path) /* POSIX mkdir(path,mode) → _mkdir(path) */
#define timegm _mkgmtime /* UTC tm → time_t */
/* setenv/unsetenv: not in the Windows CRT; map to _putenv_s. */
static inline int setenv(const char* name, const char* value, int overwrite) {
(void)overwrite;
return _putenv_s(name, value ? value : "");
}
static inline int unsetenv(const char* name) { return _putenv_s(name, ""); }
/* localtime_r/gmtime_r: Windows offers localtime_s/gmtime_s with reversed arg order. */
static inline struct tm* localtime_r(const time_t* t, struct tm* out) {
return localtime_s(out, t) == 0 ? out : (struct tm*)0;
}
static inline struct tm* gmtime_r(const time_t* t, struct tm* out) {
return gmtime_s(out, t) == 0 ? out : (struct tm*)0;
}
#endif /* EL_PLATFORM_WIN_H */
+59 -45
View File
@@ -21,10 +21,6 @@
#include "el_runtime.h"
#ifdef _WIN32
/* Windows OS-boundary shim (winsock/dlsym/popen). Threading stays on <pthread.h> (winpthreads). */
#include "el_platform_win.h"
#else
#include <stdarg.h>
#include <strings.h> /* strcasecmp */
#include <stdint.h>
@@ -46,10 +42,7 @@
#include <dirent.h>
#include <errno.h>
#include <pthread.h>
/* On POSIX, sockets close with the same close() as files; el_platform_win.h supplies the Windows
variant. Defined here so the socket call sites are identical across platforms. */
static inline int el_closesocket(int s) { return close(s); }
#endif
#include <sys/resource.h> /* getrusage — memory guard */
#ifdef HAVE_CURL
#include <curl/curl.h>
#endif
@@ -1595,17 +1588,17 @@ el_val_t http_serve(el_val_t port, el_val_t handler) {
int sock = socket(AF_INET6, SOCK_STREAM, 0);
if (sock < 0) { perror("socket"); return 0; }
int yes = 1; int no = 0;
setsockopt(sock, SOL_SOCKET, SO_REUSEADDR, (const char*)&yes, sizeof(yes));
setsockopt(sock, IPPROTO_IPV6, IPV6_V6ONLY, (const char*)&no, sizeof(no));
setsockopt(sock, SOL_SOCKET, SO_REUSEADDR, &yes, sizeof(yes));
setsockopt(sock, IPPROTO_IPV6, IPV6_V6ONLY, &no, sizeof(no));
struct sockaddr_in6 addr;
memset(&addr, 0, sizeof(addr));
addr.sin6_family = AF_INET6;
addr.sin6_addr = in6addr_any;
addr.sin6_port = htons((uint16_t)p);
if (bind(sock, (struct sockaddr*)&addr, sizeof(addr)) < 0) {
perror("bind"); el_closesocket(sock); return 0;
perror("bind"); close(sock); return 0;
}
if (listen(sock, 64) < 0) { perror("listen"); el_closesocket(sock); return 0; }
if (listen(sock, 64) < 0) { perror("listen"); close(sock); return 0; }
fprintf(stderr, "[http] listening on [::]:%d (dual-stack)\n", p);
while (1) {
struct sockaddr_in6 cli;
@@ -1622,11 +1615,11 @@ el_val_t http_serve(el_val_t port, el_val_t handler) {
_http_conn_active++;
pthread_mutex_unlock(&_http_conn_mu);
HttpWorkerArg* arg = malloc(sizeof(HttpWorkerArg));
if (!arg) { el_closesocket(cfd); continue; }
if (!arg) { close(cfd); continue; }
arg->fd = cfd;
pthread_t tid;
if (pthread_create(&tid, NULL, http_worker, arg) != 0) {
el_closesocket(cfd); free(arg);
close(cfd); free(arg);
pthread_mutex_lock(&_http_conn_mu);
_http_conn_active--;
pthread_cond_signal(&_http_conn_cv);
@@ -1635,7 +1628,7 @@ el_val_t http_serve(el_val_t port, el_val_t handler) {
}
pthread_detach(tid);
}
el_closesocket(sock);
close(sock);
return 0;
}
@@ -1845,17 +1838,17 @@ el_val_t http_serve_v2(el_val_t port, el_val_t handler) {
int sock = socket(AF_INET6, SOCK_STREAM, 0);
if (sock < 0) { perror("socket"); return 0; }
int yes = 1; int no = 0;
setsockopt(sock, SOL_SOCKET, SO_REUSEADDR, (const char*)&yes, sizeof(yes));
setsockopt(sock, IPPROTO_IPV6, IPV6_V6ONLY, (const char*)&no, sizeof(no));
setsockopt(sock, SOL_SOCKET, SO_REUSEADDR, &yes, sizeof(yes));
setsockopt(sock, IPPROTO_IPV6, IPV6_V6ONLY, &no, sizeof(no));
struct sockaddr_in6 addr;
memset(&addr, 0, sizeof(addr));
addr.sin6_family = AF_INET6;
addr.sin6_addr = in6addr_any;
addr.sin6_port = htons((uint16_t)p);
if (bind(sock, (struct sockaddr*)&addr, sizeof(addr)) < 0) {
perror("bind"); el_closesocket(sock); return 0;
perror("bind"); close(sock); return 0;
}
if (listen(sock, 64) < 0) { perror("listen"); el_closesocket(sock); return 0; }
if (listen(sock, 64) < 0) { perror("listen"); close(sock); return 0; }
fprintf(stderr, "[http v2] listening on [::]:%d (dual-stack)\n", p);
while (1) {
struct sockaddr_in6 cli;
@@ -1872,11 +1865,11 @@ el_val_t http_serve_v2(el_val_t port, el_val_t handler) {
_http_conn_active++;
pthread_mutex_unlock(&_http_conn_mu);
HttpWorkerArg* arg = malloc(sizeof(HttpWorkerArg));
if (!arg) { el_closesocket(cfd); continue; }
if (!arg) { close(cfd); continue; }
arg->fd = cfd;
pthread_t tid;
if (pthread_create(&tid, NULL, http_worker_v2, arg) != 0) {
el_closesocket(cfd); free(arg);
close(cfd); free(arg);
pthread_mutex_lock(&_http_conn_mu);
_http_conn_active--;
pthread_cond_signal(&_http_conn_cv);
@@ -1885,7 +1878,7 @@ el_val_t http_serve_v2(el_val_t port, el_val_t handler) {
}
pthread_detach(tid);
}
el_closesocket(sock);
close(sock);
return 0;
}
@@ -2059,23 +2052,6 @@ el_val_t exec(el_val_t cmdv) {
el_val_t exec_bg(el_val_t cmdv) {
const char* cmd = EL_CSTR(cmdv);
if (!cmd || !*cmd) return el_wrap_str(el_strdup(""));
#ifdef _WIN32
/* Windows: no fork/exec. Launch a detached `cmd /c <command>` with no console window via
CreateProcess (DETACHED_PROCESS | CREATE_NO_WINDOW). Returns the PID as a string, "" on fail.
Mirrors the POSIX branch: child runs independently, caller is not blocked. */
char cmdline[8192];
snprintf(cmdline, sizeof(cmdline), "cmd.exe /c %s", cmd);
STARTUPINFOA si; ZeroMemory(&si, sizeof(si)); si.cb = sizeof(si);
PROCESS_INFORMATION pi; ZeroMemory(&pi, sizeof(pi));
BOOL ok = CreateProcessA(NULL, cmdline, NULL, NULL, FALSE,
DETACHED_PROCESS | CREATE_NO_WINDOW, NULL, NULL, &si, &pi);
if (!ok) return el_wrap_str(el_strdup(""));
char pidbuf[32];
snprintf(pidbuf, sizeof(pidbuf), "%lu", (unsigned long)pi.dwProcessId);
CloseHandle(pi.hProcess);
CloseHandle(pi.hThread);
return el_wrap_str(el_strdup(pidbuf));
#else
pid_t pid = fork();
if (pid < 0) {
/* fork failed */
@@ -2098,7 +2074,6 @@ el_val_t exec_bg(el_val_t cmdv) {
char pidbuf[32];
snprintf(pidbuf, sizeof(pidbuf), "%d", (int)pid);
return el_wrap_str(el_strdup(pidbuf));
#endif
}
el_val_t fs_list(el_val_t pathv) {
@@ -4363,12 +4338,7 @@ static int _el_decompose_earth(el_caltime_t* ct, struct tm* tm_out, int* abbr_le
localtime_r(&s, &tm);
*tm_out = tm;
if (abbr_buf && abbr_cap > 0) {
/* mingw's struct tm has no tm_zone (BSD/glibc extension); no abbrev available there. */
#ifdef _WIN32
const char* z_str = "";
#else
const char* z_str = tm.tm_zone ? tm.tm_zone : "";
#endif
size_t n = strlen(z_str);
if (n >= abbr_cap) n = abbr_cap - 1;
memcpy(abbr_buf, z_str, n);
@@ -5705,6 +5675,50 @@ el_val_t getpid_now(void) {
return (el_val_t)getpid();
}
/* el_mem_check — self-terminating memory guard for long-running compiler runs.
*
* Call this periodically (e.g. after each function compiled) to detect runaway
* memory growth before the OS OOM-killer fires. Reads the limit from the env
* var ELC_MAX_MEM_MB (default 512 MB). If resident set size exceeds the limit,
* prints a diagnostic to stderr and exits with code 1 so the caller (elb or a
* CI script) can handle the failure gracefully instead of having the whole
* machine go down.
*
* Platform notes:
* macOS ru_maxrss is in bytes.
* Linux ru_maxrss is in kilobytes.
* We normalise to MB before comparing.
*
* Returns 0 always (the only non-return path is the exit() branch).
*/
el_val_t el_mem_check(void) {
/* Read limit from env; default 512 MB. */
long limit_mb = 512;
const char *env_val = getenv("ELC_MAX_MEM_MB");
if (env_val && *env_val) {
long v = atol(env_val);
if (v > 0) limit_mb = v;
}
struct rusage ru;
if (getrusage(RUSAGE_SELF, &ru) != 0) return 0; /* can't read — skip check */
long rss_mb;
#if defined(__APPLE__) || defined(__MACH__)
/* macOS: ru_maxrss is bytes */
rss_mb = (long)(ru.ru_maxrss / (1024L * 1024L));
#else
/* Linux: ru_maxrss is kilobytes */
rss_mb = (long)(ru.ru_maxrss / 1024L);
#endif
if (rss_mb >= limit_mb) {
fprintf(stderr, "elc: memory limit exceeded (%ldMB), aborting\n", limit_mb);
exit(1);
}
return 0;
}
/* ── args() — command-line argument access ──────────────────────────────────
* Compiled El programs call args() to get a list of CLI arguments.
* Call el_runtime_init_args(argc, argv) at the start of C main() to populate.
+6 -6
View File
@@ -52,12 +52,6 @@
typedef int64_t el_val_t;
/* HTTP request-handler function-pointer types. Public because soul modules (routes/chat/etc.)
* register handlers across translation units; previously defined only inside el_runtime.c, which
* made cross-module references (and the Windows build) fail. Home in the shared header. */
typedef el_val_t (*http_handler_fn)(el_val_t method, el_val_t path, el_val_t body);
typedef el_val_t (*http_handler4_fn)(el_val_t method, el_val_t path, el_val_t body, el_val_t headers);
#define EL_STR(s) ((el_val_t)(uintptr_t)(s))
#define EL_CSTR(v) ((const char*)(uintptr_t)(v))
#define EL_INT(v) (v)
@@ -537,6 +531,12 @@ el_val_t parse_int(el_val_t s, el_val_t default_val);
el_val_t exit_program(el_val_t code);
el_val_t getpid_now(void);
/* Self-terminating memory guard. Reads ELC_MAX_MEM_MB (default 512) and
* exits with code 1 if resident memory exceeds the limit. Call periodically
* during long compilation loops (e.g. after each function is compiled).
* Returns 0 when memory is within bounds. */
el_val_t el_mem_check(void);
/* ── CGI identity ─────────────────────────────────────────────────────────────
* Called at the start of main() in CGI programs (those with a `cgi {}` block).
* Records the program's DHARMA identity before any other code executes. */
+16
View File
@@ -324,6 +324,10 @@ fn cg_html_parts(children: [Map<String, Any>], acc_var: String) -> String {
let each_c: String = cg_html_each(child, acc_var)
let parts = native_list_append(parts, each_c)
}
if str_eq(html_kind, "HtmlIf") {
let if_c: String = cg_html_if(child, acc_var)
let parts = native_list_append(parts, if_c)
}
let i = i + 1
}
str_join(parts, "")
@@ -413,6 +417,17 @@ fn cg_html_each(node: Map<String, Any>, acc_var: String) -> String {
"{ el_val_t " + list_var + " = (" + list_c + "); el_val_t " + len_var + " = el_list_len(" + list_var + "); for (el_val_t " + idx_var + " = 0; " + idx_var + " < " + len_var + "; " + idx_var + "++) { el_val_t " + item_name + " = el_list_get(" + list_var + ", " + idx_var + "); " + inner_c + "} } "
}
// Generate code for {#if cond} ... {/if} (with optional {#else}).
fn cg_html_if(node: Map<String, Any>, acc_var: String) -> String {
let cond_expr = node["cond"]
let then_children: [Map<String, Any>] = node["then"]
let else_children: [Map<String, Any>] = node["else"]
let cond_c: String = cg_expr(cond_expr)
let then_c: String = cg_html_parts(then_children, acc_var)
let else_c: String = cg_html_parts(else_children, acc_var)
"if (" + cond_c + ") { " + then_c + " } else { " + else_c + " } "
}
// Top-level HTML template codegen returns a C statement-expression string.
fn cg_html_template(expr: Map<String, Any>) -> String {
let root = expr["root"]
@@ -3715,6 +3730,7 @@ fn codegen_streaming(tokens: [Any], sigs: [Map<String, Any>], source: String) ->
cg_fn(stmt)
el_release(stmt)
el_arena_pop(fn_arena_mark)
el_mem_check()
}
} else {
if is_top_level_decl(stmt) {
+38 -5
View File
@@ -287,6 +287,9 @@ fn type_node_to_el(t: Map<String, Any>) -> String {
// emit_header write a .elh file from parsed statements.
// Scans for FnDef nodes and emits 'extern fn' declarations.
// NOTE: This function requires the full AST. Prefer emit_header_from_sigs
// for the --emit-header path it works from a token-level scan without
// building expression ASTs, avoiding OOM on large files.
fn emit_header(stmts: [Map<String, Any>], hdr_path: String) -> Void {
let n: Int = native_list_len(stmts)
let i = 0
@@ -325,6 +328,32 @@ fn emit_header(stmts: [Map<String, Any>], hdr_path: String) -> Void {
let ok: Bool = fs_write(hdr_path, content)
}
// emit_header_from_sigs write a .elh file from pre-scanned El signatures.
// Uses the output of scan_fn_sigs_el() no full AST required.
// Peak memory is O(tokens) rather than O(whole-program AST), which prevents
// OOM on large files with HTML template bodies or deep BinOp chains.
fn emit_header_from_sigs(sigs: [Map<String, Any>], hdr_path: String) -> Void {
let n: Int = native_list_len(sigs)
let i: Int = 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 sig = native_list_get(sigs, i)
let kind: String = sig["kind"]
if str_eq(kind, "fn") {
let name: String = sig["name"]
let params_el: String = sig["params_el"]
let ret_el: String = sig["ret_el"]
if str_eq(ret_el, "") { let ret_el = "Any" }
let line: String = "extern fn " + name + "(" + params_el + ") -> " + ret_el
let parts = native_list_append(parts, line + "\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:
@@ -536,16 +565,20 @@ fn main() -> Void {
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.
// When --emit-header is requested, lex the source file and do a
// token-level signature scan (no full AST) to write a .elh file.
// This avoids OOM on large files with HTML template bodies or deep
// BinOp chains (e.g. checkout.el) parse() builds O(whole-program AST)
// while scan_fn_sigs_el keeps peak memory at O(tokens).
if do_emit_header {
el_mem_check()
let raw_source: String = fs_read(src_path)
let hdr_tokens: [Any] = lex(raw_source)
let hdr_stmts: [Map<String, Any>] = parse(hdr_tokens)
let hdr_sigs: [Map<String, Any>] = scan_fn_sigs_el(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)
emit_header_from_sigs(hdr_sigs, hdr_path)
el_release(hdr_sigs)
}
let source: String = resolve_imports(src_path)
+497 -94
View File
@@ -293,6 +293,48 @@ fn is_void_element(name: String) -> Bool {
false
}
// Collect all tokens as raw text until </tag_name> is encountered.
// Used for <style> and <script> elements to avoid parsing CSS/JS as El.
// Returns { "text": "...", "pos": p_after_closing_tag }
fn parse_raw_text_content(tokens: [Any], pos: Int, tag_name: String) -> Map<String, Any> {
let parts: [String] = native_list_empty()
let p = pos
let running = true
while running {
let k = tok_kind(tokens, p)
if str_eq(k, "Eof") {
let running = false
} else {
if str_eq(k, "Lt") {
let k2 = tok_kind(tokens, p + 1)
if str_eq(k2, "Slash") {
// Check if this is </tag_name>
let close_name = tok_value(tokens, p + 2)
if str_eq(close_name, tag_name) {
// consume </tag_name>
let p = p + 3
let p = expect(tokens, p, "Gt")
let running = false
} else {
let v = tok_value(tokens, p)
let parts = native_list_append(parts, v)
let p = p + 1
}
} else {
let v = tok_value(tokens, p)
let parts = native_list_append(parts, v)
let p = p + 1
}
} else {
let v = tok_value(tokens, p)
let parts = native_list_append(parts, v)
let p = p + 1
}
}
}
{ "text": str_join(parts, ""), "pos": p }
}
// Collect tokens as text content until we hit Lt, LBrace, Eof, or a
// closing-tag marker (Lt Slash). Returns { "text": "...", "pos": p }
fn parse_html_text_tokens(tokens: [Any], pos: Int) -> Map<String, Any> {
@@ -320,7 +362,7 @@ fn parse_html_text_tokens(tokens: [Any], pos: Int) -> Map<String, Any> {
}
}
}
{ "text": str_join(parts, " "), "pos": p }
{ "text": str_join(parts, ""), "pos": p }
}
// Parse an attribute list: (attrname | attrname="val" | attrname={expr})*
@@ -435,77 +477,125 @@ fn parse_html_children(tokens: [Any], pos: Int, parent_tag: String) -> Map<Strin
}
} else {
if str_eq(k, "LBrace") {
// Interpolation: {expr} or {#each ...} or {/each}
// Interpolation: {expr}, {#each ...}, {#if ...}, {#else}, {/each}, {/if}
// Note: '#' (ASCII 35) is skipped by the lexer, so {#each} lexes as
// LBrace Ident:"each" ... and {#if} lexes as LBrace If ... and
// {#else} lexes as LBrace Else RBrace.
let k2 = tok_kind(tokens, p + 1)
if str_eq(k2, "Hash") {
// {#each list as item}
let k3_v = tok_value(tokens, p + 2)
if str_eq(k3_v, "each") {
let p = p + 3
// parse list expr up to "as" keyword
if str_eq(k2, "Slash") {
// {/each} or {/if} end of block, stop
// skip { /
let p = p + 2
// skip the close-tag name token (each, if, etc.)
let p = p + 1
// skip }
let p = expect(tokens, p, "RBrace")
let running = false
} else {
if str_eq(k2, "If") {
// {#if condition} ... {/if}
// Skip { if (2 tokens; '#' was silently skipped by lexer)
let p = p + 2
// Parse condition expression (no block expr)
let prev_no_block: String = state_get("__no_block_expr")
state_set("__no_block_expr", "1")
let r_list = parse_expr(tokens, p)
let r_cond = parse_expr(tokens, p)
state_set("__no_block_expr", prev_no_block)
let list_expr = r_list["node"]
let p = r_list["pos"]
// r_list result map fully consumed release to free peak heap.
el_release(r_list)
// expect "as"
let p = expect(tokens, p, "As")
// item variable name
let item_name = tok_value(tokens, p)
let p = p + 1
let cond_expr = r_cond["node"]
let p = r_cond["pos"]
el_release(r_cond)
// consume closing }
let p = expect(tokens, p, "RBrace")
// parse body until {/each}
let r_body = parse_html_each_body(tokens, p)
let body_children = r_body["children"]
let p = r_body["pos"]
// r_body result map fully consumed release to free peak heap.
el_release(r_body)
let each_node: Map<String, Any> = { "html": "Each", "list": list_expr, "item": item_name, "body": body_children }
let children = native_list_append(children, each_node)
} else {
let p = p + 1
}
} else {
if str_eq(k2, "Slash") {
// {/each} end of each block, stop
// skip {/each}
let p = p + 2
// skip "each"
let p = p + 1
// skip }
let p = expect(tokens, p, "RBrace")
let running = false
} else {
// regular {expr}
let r = parse_expr(tokens, p + 1)
let interp_val = r["node"]
let p = r["pos"]
// r result map fully consumed release to free peak heap.
el_release(r)
let p = expect(tokens, p, "RBrace")
// Check if the expr is a call to raw()
let is_raw_call = false
let interp_kind: String = interp_val["expr"]
if str_eq(interp_kind, "Call") {
let fn_node = interp_val["func"]
let fn_kind: String = fn_node["expr"]
if str_eq(fn_kind, "Ident") {
let fn_name_v: String = fn_node["name"]
if str_eq(fn_name_v, "raw") {
let is_raw_call = true
}
// parse then-children until {#else} or {/if}
let r_then = parse_html_children(tokens, p, "__if_then__")
let then_children = r_then["children"]
let p = r_then["pos"]
el_release(r_then)
// check for {#else} lexed as LBrace Else RBrace
let else_children: [Map<String, Any>] = native_list_empty()
let ck = tok_kind(tokens, p)
if str_eq(ck, "LBrace") {
let ck2 = tok_kind(tokens, p + 1)
if str_eq(ck2, "Else") {
// consume { else }
let p = p + 2
let p = expect(tokens, p, "RBrace")
// parse else-children until {/if}
let r_else = parse_html_children(tokens, p, "__if_else__")
let else_children = r_else["children"]
let p = r_else["pos"]
el_release(r_else)
}
}
if is_raw_call {
let raw_args = interp_val["args"]
let raw_inner = native_list_get(raw_args, 0)
let children = native_list_append(children, { "html": "Raw", "value": raw_inner })
let if_node: Map<String, Any> = { "html": "HtmlIf", "cond": cond_expr, "then": then_children, "else": else_children }
let children = native_list_append(children, if_node)
} else {
if str_eq(k2, "Else") {
// {#else} sentinel lexed as LBrace Else RBrace
// Do NOT consume leave position for caller ({#if} handler checks for it)
let running = false
} else {
let children = native_list_append(children, { "html": "Interp", "value": interp_val })
// Check for {#each list as item} lexed as LBrace Ident:"each" ...
let k2_v = tok_value(tokens, p + 1)
if str_eq(k2_v, "each") {
let p = p + 2
// parse list expr up to "as" keyword
let prev_no_block: String = state_get("__no_block_expr")
state_set("__no_block_expr", "1")
let r_list = parse_expr(tokens, p)
state_set("__no_block_expr", prev_no_block)
let list_expr = r_list["node"]
let p = r_list["pos"]
// r_list result map fully consumed release to free peak heap.
el_release(r_list)
// expect "as"
let p = expect(tokens, p, "As")
// item variable name
let item_name = tok_value(tokens, p)
let p = p + 1
// consume closing }
let p = expect(tokens, p, "RBrace")
// parse body until {/each}
let r_body = parse_html_each_body(tokens, p)
let body_children = r_body["children"]
let p = r_body["pos"]
// r_body result map fully consumed release to free peak heap.
el_release(r_body)
let each_node: Map<String, Any> = { "html": "Each", "list": list_expr, "item": item_name, "body": body_children }
let children = native_list_append(children, each_node)
} else {
// regular {expr} disable map-literal parsing so {fn(a,b)}
// does not trigger the LBracemap path inside parse_primary
let prev_no_block: String = state_get("__no_block_expr")
state_set("__no_block_expr", "1")
let r = parse_expr(tokens, p + 1)
state_set("__no_block_expr", prev_no_block)
let interp_val = r["node"]
let p = r["pos"]
// r result map fully consumed release to free peak heap.
el_release(r)
let p = expect(tokens, p, "RBrace")
// Check if the expr is a call to raw()
let is_raw_call = false
let interp_kind: String = interp_val["expr"]
if str_eq(interp_kind, "Call") {
let fn_node = interp_val["func"]
let fn_kind: String = fn_node["expr"]
if str_eq(fn_kind, "Ident") {
let fn_name_v: String = fn_node["name"]
if str_eq(fn_name_v, "raw") {
let is_raw_call = true
}
}
}
if is_raw_call {
let raw_args = interp_val["args"]
let raw_inner = native_list_get(raw_args, 0)
let children = native_list_append(children, { "html": "Raw", "value": raw_inner })
} else {
let children = native_list_append(children, { "html": "Interp", "value": interp_val })
}
}
}
}
}
@@ -565,6 +655,27 @@ fn parse_html_element(tokens: [Any], pos: Int) -> Map<String, Any> {
if is_void_element(tag_name) {
return make_result({ "html": "Element", "tag": tag_name, "attrs": attrs, "children": native_list_empty(), "self_closing": true }, p)
}
// raw-text mode for style/script collect content as plain text without parsing CSS/JS as El
if str_eq(tag_name, "style") {
let r_raw = parse_raw_text_content(tokens, p, "style")
let raw_text: String = r_raw["text"]
let p = r_raw["pos"]
el_release(r_raw)
let raw_child: Map<String, Any> = { "html": "Text", "text": raw_text }
let raw_children: [Map<String, Any>] = native_list_empty()
let raw_children = native_list_append(raw_children, raw_child)
return make_result({ "html": "Element", "tag": tag_name, "attrs": attrs, "children": raw_children, "self_closing": false }, p)
}
if str_eq(tag_name, "script") {
let r_raw = parse_raw_text_content(tokens, p, "script")
let raw_text: String = r_raw["text"]
let p = r_raw["pos"]
el_release(r_raw)
let raw_child: Map<String, Any> = { "html": "Text", "text": raw_text }
let raw_children: [Map<String, Any>] = native_list_empty()
let raw_children = native_list_append(raw_children, raw_child)
return make_result({ "html": "Element", "tag": tag_name, "attrs": attrs, "children": raw_children, "self_closing": false }, p)
}
// parse children
let r_children = parse_html_children(tokens, p, tag_name)
let children = r_children["children"]
@@ -718,44 +829,123 @@ fn parse_primary(tokens: [Any], pos: Int) -> Map<String, Any> {
// as the start of the block they're expecting.
return make_result({ "expr": "Nil" }, pos)
}
let p = pos + 1
let pairs: [Map<String, Any>] = native_list_empty()
let running = true
while running {
let k2 = tok_kind(tokens, p)
if k2 == "RBrace" {
let running = false
} else {
if k2 == "Eof" {
// Distinguish map literal from interpolation chain.
// A map literal requires { key: value } the second token inside { must be Colon.
// An empty {} is a map literal. Everything else is an interpolation chain.
let first_k: String = tok_kind(tokens, pos + 1)
let second_k: String = tok_kind(tokens, pos + 2)
if str_eq(first_k, "RBrace") {
// Empty map literal {}
return make_result({ "expr": "Map", "pairs": native_list_empty() }, pos + 2)
}
if str_eq(second_k, "Colon") {
// MAP LITERAL: { key: value, ... }
let p = pos + 1
let pairs: [Map<String, Any>] = native_list_empty()
let running = true
while running {
let k2 = tok_kind(tokens, p)
if k2 == "RBrace" {
let running = false
} else {
// key: Str token
let key = tok_value(tokens, p)
let new_p: Int = p + 1
let new_p = expect(tokens, new_p, "Colon")
let r = parse_expr(tokens, new_p)
let val_node = r["node"]
let new_p = r["pos"]
// r result map fully consumed release to free peak heap.
el_release(r)
let pair = { "key": key, "value": val_node }
let pairs = native_list_append(pairs, pair)
let k3 = tok_kind(tokens, new_p)
if k3 == "Comma" {
let new_p = new_p + 1
}
// Non-progress guard: malformed map content can leave
// parse_expr returning the same pos. Force advance.
if new_p <= p {
let p = p + 1
if k2 == "Eof" {
let running = false
} else {
let p = new_p
// key: Str or Ident token
let key = tok_value(tokens, p)
let new_p: Int = p + 1
let new_p = expect(tokens, new_p, "Colon")
let r = parse_expr(tokens, new_p)
let val_node = r["node"]
let new_p = r["pos"]
// r result map fully consumed release to free peak heap.
el_release(r)
let pair = { "key": key, "value": val_node }
let pairs = native_list_append(pairs, pair)
let k3 = tok_kind(tokens, new_p)
if k3 == "Comma" {
let new_p = new_p + 1
}
// Non-progress guard: malformed map content can leave
// parse_expr returning the same pos. Force advance.
if new_p <= p {
let p = p + 1
} else {
let p = new_p
}
}
}
}
let p = expect(tokens, p, "RBrace")
return make_result({ "expr": "Map", "pairs": pairs }, p)
}
let p = expect(tokens, p, "RBrace")
return make_result({ "expr": "Map", "pairs": pairs }, p)
// INTERPOLATION CHAIN: {expr}, {expr}{expr}, {expr}<html>, etc.
// Build a BinOp(Plus, ...) concatenation chain.
let p = pos
let chain_node: Map<String, Any> = { "expr": "Nil" }
let chain_started = false
let chain_running = true
while chain_running {
let ck: String = tok_kind(tokens, p)
if str_eq(ck, "LBrace") {
let prev_no_block: String = state_get("__no_block_expr")
state_set("__no_block_expr", "1")
let r = parse_expr(tokens, p + 1)
state_set("__no_block_expr", prev_no_block)
let part = r["node"]
let p = r["pos"]
// r result map fully consumed release to free peak heap.
el_release(r)
let p = expect(tokens, p, "RBrace")
if !chain_started {
let chain_node = part
let chain_started = true
} else {
let chain_node: Map<String, Any> = { "expr": "BinOp", "op": "Plus", "left": chain_node, "right": part }
}
} else {
if str_eq(ck, "Lt") {
let ck2: String = tok_kind(tokens, p + 1)
if str_eq(ck2, "Not") {
let r = parse_html_template(tokens, p)
let part = r["node"]
let p = r["pos"]
// r result map fully consumed release to free peak heap.
el_release(r)
if !chain_started {
let chain_node = part
let chain_started = true
} else {
let chain_node: Map<String, Any> = { "expr": "BinOp", "op": "Plus", "left": chain_node, "right": part }
}
} else {
if str_eq(ck2, "Ident") {
let tag_candidate: String = tok_value(tokens, p + 1)
if is_html_tag_name(tag_candidate) {
let r = parse_html_template(tokens, p)
let part = r["node"]
let p = r["pos"]
// r result map fully consumed release to free peak heap.
el_release(r)
if !chain_started {
let chain_node = part
let chain_started = true
} else {
let chain_node: Map<String, Any> = { "expr": "BinOp", "op": "Plus", "left": chain_node, "right": part }
}
} else {
let chain_running = false
}
} else {
let chain_running = false
}
}
} else {
let chain_running = false
}
}
}
return make_result(chain_node, p)
}
// if expression
@@ -1875,6 +2065,219 @@ fn skip_expr_to_stmt_boundary(tokens: [Any], pos: Int) -> Int {
p
}
// scan_type_el read a type annotation starting at pos and return its El
// source representation as a string, plus the new position.
// Returns { "el": String, "pos": Int }.
// Handles: Ident, [Type], Map<K,V>, Type?, Type<T,...> (same shapes as skip_type).
fn scan_type_el(tokens: [Any], pos: Int) -> Map<String, Any> {
let k: String = tok_kind(tokens, pos)
// Array type: [Type]
if str_eq(k, "LBracket") {
let p: Int = pos + 1
let inner = scan_type_el(tokens, p)
let inner_str: String = inner["el"]
let p = inner["pos"]
el_release(inner)
let p = expect(tokens, p, "RBracket")
return { "el": "[" + inner_str + "]", "pos": p }
}
// Named type (possibly generic or optional)
if str_eq(k, "Ident") {
let name: String = tok_value(tokens, pos)
let p: Int = pos + 1
let k2: String = tok_kind(tokens, p)
if str_eq(k2, "Lt") {
// Generic params: collect until matching >
let p = p + 1
let depth: Int = 1
let parts: [String] = native_list_empty()
let parts = native_list_append(parts, name + "<")
let running: Bool = true
while running {
let kk: String = tok_kind(tokens, p)
if str_eq(kk, "Eof") {
let running = false
} else {
if str_eq(kk, "Lt") {
let depth = depth + 1
let parts = native_list_append(parts, "<")
let p = p + 1
} else {
if str_eq(kk, "Gt") {
let depth = depth - 1
let p = p + 1
if depth <= 0 {
let parts = native_list_append(parts, ">")
let running = false
} else {
let parts = native_list_append(parts, ">")
}
} else {
if str_eq(kk, "Comma") {
let parts = native_list_append(parts, ", ")
let p = p + 1
} else {
let parts = native_list_append(parts, tok_value(tokens, p))
let p = p + 1
}
}
}
}
}
let k3: String = tok_kind(tokens, p)
if str_eq(k3, "QuestionMark") { let p = p + 1 }
let result: String = str_join(parts, "")
el_release(parts)
return { "el": result, "pos": p }
}
// Optional marker
if str_eq(k2, "QuestionMark") {
return { "el": name + "?", "pos": p + 1 }
}
return { "el": name, "pos": p }
}
// Fallback: unknown token, treat as Any
{ "el": "Any", "pos": pos + 1 }
}
// scan_params_el scan a parameter list `(name: Type, ...)` starting at
// position `pos` (which should point at LParen) and return the El parameter
// declaration string (e.g. "a: String, b: Int") along with the new position.
// Returns { "el": String, "pos": Int }.
// Used by scan_fn_sigs_el for --emit-header without building full AST.
fn scan_params_el(tokens: [Any], pos: Int) -> Map<String, Any> {
let p: Int = expect(tokens, pos, "LParen")
let parts: [String] = native_list_empty()
let going: Bool = true
while going {
let kk: String = tok_kind(tokens, p)
if str_eq(kk, "RParen") {
let going = false
} else {
if str_eq(kk, "Eof") {
let going = false
} else {
let pname: String = tok_value(tokens, p)
let p = p + 1
let p = expect(tokens, p, "Colon")
let tr = scan_type_el(tokens, p)
let ptype: String = tr["el"]
let p = tr["pos"]
el_release(tr)
let parts = native_list_append(parts, pname + ": " + ptype)
let k2: String = tok_kind(tokens, p)
if str_eq(k2, "Comma") {
let p = p + 1
}
}
}
}
let p = expect(tokens, p, "RParen")
let el_str: String = str_join(parts, ", ")
el_release(parts)
{ "el": el_str, "pos": p }
}
// scan_fn_sigs_el lightweight token-level pre-scan for --emit-header.
//
// Like scan_fn_sigs but captures El-style type strings instead of C types.
// Only records fn/extern_fn entries (header generation ignores lets/blocks).
//
// Descriptor shape:
// { "kind": "fn"|"extern_fn", "name": String,
// "params_el": String, <- El param list, e.g. "a: String, b: Int"
// "ret_el": String } <- El return type, e.g. "String" or "Void"
//
// Peak memory: O(tokens) with no expression AST allocation.
fn scan_fn_sigs_el(tokens: [Any]) -> [Map<String, Any>] {
let total: Int = native_list_len(tokens) / 2
let sigs: [Map<String, Any>] = native_list_empty()
let pos: Int = 0
let going: Bool = true
while going {
if pos >= total {
let going = false
} else {
let k: String = tok_kind(tokens, pos)
if str_eq(k, "Eof") {
let going = false
} else {
// --- fn definition ---
if str_eq(k, "Fn") {
let p: Int = pos + 1
let name: String = tok_value(tokens, p)
let p = p + 1
let pr = scan_params_el(tokens, p)
let params_el: String = pr["el"]
let p = pr["pos"]
el_release(pr)
// read return type
let ret_el: String = "Any"
let k2: String = tok_kind(tokens, p)
if str_eq(k2, "Arrow") {
let p = p + 1
let tr = scan_type_el(tokens, p)
let ret_el = tr["el"]
let p = tr["pos"]
el_release(tr)
}
// skip body
let k3: String = tok_kind(tokens, p)
if str_eq(k3, "LBrace") {
let p = skip_to_rbrace(tokens, p)
}
if !str_eq(name, "main") {
let sigs = native_list_append(sigs, {
"kind": "fn",
"name": name,
"params_el": params_el,
"ret_el": ret_el
})
}
let pos = p
} else {
// --- extern fn ---
if str_eq(k, "Extern") {
let p: Int = pos + 1
let k2: String = tok_kind(tokens, p)
if str_eq(k2, "Fn") {
let p = p + 1
let name: String = tok_value(tokens, p)
let p = p + 1
let pr = scan_params_el(tokens, p)
let params_el: String = pr["el"]
let p = pr["pos"]
el_release(pr)
let ret_el: String = "Any"
let k3: String = tok_kind(tokens, p)
if str_eq(k3, "Arrow") {
let p = p + 1
let tr = scan_type_el(tokens, p)
let ret_el = tr["el"]
let p = tr["pos"]
el_release(tr)
}
let sigs = native_list_append(sigs, {
"kind": "extern_fn",
"name": name,
"params_el": params_el,
"ret_el": ret_el
})
let pos = p
} else {
let pos = pos + 1
}
} else {
// Let, Cgi, Service, Import, Type, Enum, From skip to boundary.
let p: Int = pos + 1
let p = skip_expr_to_stmt_boundary(tokens, p)
let pos = p
}}}
}
}
sigs
}
// scan_params_c scan a parameter list `(name: Type, ...)` starting at
// position `pos` (which should point at LParen) and return the C parameter
// declaration string along with the new position.
+3 -43
View File
@@ -6,55 +6,15 @@
//
// Dependencies: runtime/string.el, runtime/json.el
// --- Validation (defense in depth) ---
// el_val_t is an untyped machine word, so a wrong TYPE can't be caught here but a
// wrong VALUE can (a tier in the node_type slot, an empty/garbage string, an int, a
// path, a model name, a cgi id). Reject loudly instead of silently writing junk.
fn engram_valid_node_type(t: String) -> Bool {
return str_eq(t, "Memory") || str_eq(t, "Knowledge") || str_eq(t, "Belief")
|| str_eq(t, "Project") || str_eq(t, "Tag") || str_eq(t, "BacklogItem")
|| str_eq(t, "Artifact") || str_eq(t, "Conversation") || str_eq(t, "ExecutionContext")
|| str_eq(t, "InternalStateEvent") || str_eq(t, "Self") || str_eq(t, "Entity")
|| str_eq(t, "Process") || str_eq(t, "ConfigEntry") || str_eq(t, "Concept") || str_eq(t, "Imprint")
|| str_eq(t, "SessionSummary")
}
fn engram_valid_tier(t: String) -> Bool {
return str_eq(t, "Semantic") || str_eq(t, "Episodic") || str_eq(t, "Working")
|| str_eq(t, "Procedural") || str_eq(t, "Canonical") || str_eq(t, "Note") || str_eq(t, "Lesson")
}
// --- Node creation ---
fn engram_node(content: String, node_type: String, salience: Float) -> String {
if !engram_valid_node_type(node_type) {
__println("[engram] REJECTED node write — invalid node_type '" + node_type + "'")
return ""
}
return __engram_node(content, node_type, salience)
}
// Signature MUST match the C primitive __engram_node_full exactly (el_seed.h):
// (content, node_type, label, salience, importance, confidence, tier, tags)
// The previous wrapper declared a stale 8-arg schema with wrong names AND types
// (sal:Float at the label slot, ts:Int at the tier slot). Because el_val_t is an
// untyped machine word, the EL compiler coerced caller args to those wrong param
// types and then forwarded them BY POSITION into the C function so tier received
// an int, importance/confidence received strings, label received a float, etc.
// That is the field-corruption bug. Match the contract 1:1 no coercion, no reorder.
fn engram_node_full(content: String, node_type: String, label: String,
salience: Float, importance: Float, confidence: Float,
tier: String, tags: String) -> String {
if !engram_valid_node_type(node_type) {
__println("[engram] REJECTED node write — invalid node_type '" + node_type + "' (label=" + label + ")")
return ""
}
if !engram_valid_tier(tier) {
__println("[engram] REJECTED node write — invalid tier '" + tier + "' (node_type=" + node_type + ", label=" + label + ")")
return ""
}
return __engram_node_full(content, node_type, label, salience, importance, confidence, tier, tags)
fn engram_node_full(content: String, nt: String, sal: Float, imp: Float,
source: String, lang: String, ts: Int, tags: String) -> String {
return __engram_node_full(content, nt, sal, imp, source, lang, ts, tags)
}
// --- Node retrieval ---