tooling improvements

seed.ce

@@ -0,0 +1,116 @@
+// seed.ce — regenerate boot seed files in boot/
+//
+// Usage:
+//   pit seed                Regenerate all boot seeds
+//   pit seed --clean        Also clear the build cache after
+//
+// Seeds are the pre-compiled mcode IR (JSON) files that bootstrap the
+// compilation pipeline on cold start. They only need regenerating when:
+//   - The pipeline source (tokenize, parse, fold, mcode, streamline) changes
+//     in a way that the boot seeds can't compile the new source
+//   - You want the seed files to match the current pipeline output
+//     (e.g. before a release or distribution)
+//
+// The engine already recompiles pipeline modules from source when their
+// content changes (content-addressed cache). Seeds are a fallback for
+// cold start when no cache exists.
+
+var fd = use("fd")
+var json = use("json")
+var os = use("os")
+var shop = use("internal/shop")
+var tokenize = use("tokenize")
+var parse = use("parse")
+var fold = use("fold")
+var mcode = use("mcode")
+var streamline = use("streamline")
+
+var clean = false
+var i = 0
+
+for (i = 0; i < length(args); i++) {
+  if (args[i] == '--clean') {
+    clean = true
+  } else if (args[i] == '--help' || args[i] == '-h') {
+    print("usage: pit seed [--clean]")
+    print("")
+    print("  Regenerate boot seed files in boot/")
+    print("  --clean    Also clear the build cache after")
+    $stop()
+  }
+}
+
+var core_dir = shop.get_package_dir('core')
+var boot_dir = core_dir + '/boot'
+var pipeline_modules = ['tokenize', 'parse', 'fold', 'mcode', 'streamline']
+var generated = 0
+var name = null
+var src_path = null
+var src = null
+var tok = null
+var ast = null
+var folded = null
+var compiled = null
+var optimized = null
+var mcode_json = null
+var out_path = null
+
+// Regenerate pipeline module seeds
+for (i = 0; i < length(pipeline_modules); i++) {
+  name = pipeline_modules[i]
+  src_path = core_dir + '/' + name + '.cm'
+
+  if (!fd.is_file(src_path)) {
+    print('WARNING: source not found: ' + src_path)
+    continue
+  }
+
+  print('Seeding ' + name + '.cm ...')
+  src = text(fd.slurp(src_path))
+  tok = tokenize(src, src_path)
+  ast = parse(tok.tokens, src, src_path, tokenize)
+  folded = fold(ast)
+  compiled = mcode(folded)
+  optimized = streamline(compiled)
+  mcode_json = json.encode(optimized)
+
+  out_path = boot_dir + '/' + name + '.cm.mcode'
+  fd.slurpwrite(out_path, stone(blob(mcode_json)))
+  print('  -> ' + out_path + ' (' + text(length(mcode_json)) + ' bytes)')
+  generated = generated + 1
+}
+
+// Regenerate bootstrap.cm seed
+var bootstrap_path = core_dir + '/internal/bootstrap.cm'
+if (fd.is_file(bootstrap_path)) {
+  print('Seeding bootstrap.cm ...')
+  src = text(fd.slurp(bootstrap_path))
+  tok = tokenize(src, bootstrap_path)
+  ast = parse(tok.tokens, src, bootstrap_path, tokenize)
+  folded = fold(ast)
+  compiled = mcode(folded)
+  optimized = streamline(compiled)
+  mcode_json = json.encode(optimized)
+
+  out_path = boot_dir + '/bootstrap.cm.mcode'
+  fd.slurpwrite(out_path, stone(blob(mcode_json)))
+  print('  -> ' + out_path + ' (' + text(length(mcode_json)) + ' bytes)')
+  generated = generated + 1
+} else {
+  print('WARNING: bootstrap source not found: ' + bootstrap_path)
+}
+
+print('\nRegenerated ' + text(generated) + ' seed(s)')
+
+if (clean) {
+  var build_dir = shop.get_build_dir()
+  if (fd.is_dir(build_dir)) {
+    print('Clearing build cache: ' + build_dir)
+    os.system('rm -rf "' + build_dir + '"')
+    print('Build cache cleared. Next run will recompile from new seeds.')
+  } else {
+    print('No build cache to clear.')
+  }
+}
+
+$stop()