fix(extract): parse PHP past inline HTML so symbols after ?> reach the graph (#2000)

Makefile.cbm

@@ -348,6 +348,7 @@ EXTRACTION_SRCS = \
     $(CBM_DIR)/extract_channels.c \
     $(CBM_DIR)/extract_k8s.c \
     $(CBM_DIR)/extract_dbt.c \
+    $(CBM_DIR)/php_inline_html.c \
     $(CBM_DIR)/helpers.c \
     $(CBM_DIR)/result_compact.c \
     $(CBM_DIR)/result_spill.c \

internal/cbm/cbm.c

@@ -2057,6 +2057,18 @@ CBMFileResult *cbm_extract_file(const char *source, int source_len, CBMLanguage
     return r;
 }
 
+/* The bytes the grammar parses. PHP that leaves PHP mode (`?> markup <?php`)
+ * is parsed with its markup blanked (#2000); the rewrite keeps the length and
+ * every line break, so offsets and line numbers still refer to the file on
+ * disk. Every other language parses the file as it is. */
+static const char *extract_parse_bytes(CBMArena *a, CBMLanguage language, const char *source,
+                                       int source_len) {
+    if (language == CBM_LANG_PHP) {
+        return cbm_php_mask_inline_html(a, source, source_len);
+    }
+    return source;
+}
+
 /* Initial block for the per-file traversal scratch arena, chosen by measuring
  * arena_grow on a 14k-file TypeScript tree: it fires on one file in 12,000 at
  * both this size and at 1 MB, and on most files at 256 KB, where the two
@@ -2122,6 +2134,8 @@ static CBMFileResult *extract_file_ex_body(const char *source, int source_len, C
         return result;
     }
 
+    source = extract_parse_bytes(a, language, source, source_len);
+
     // Skip pathologically nested Perl before tree-sitter's recursive GLR stack
     // merge overflows a small stack during the parse (see
     // cbm_source_nesting_exceeds). Scoped to Perl: its ambiguous call grammar is

internal/cbm/cbm.h

@@ -891,6 +891,10 @@ void cbm_extract_definitions_without_module(CBMExtractCtx *ctx);
 // per ref()/source() call. No-op unless the file parses as SQL and actually
 // contains a dbt builtin call. Defined in extract_dbt.c.
 void cbm_extract_dbt(CBMExtractCtx *ctx);
+// PHP source with its inline-HTML regions blanked so the php_only grammar can
+// parse past `?>` (#2000). Same length and line breaks as the input; returns
+// `source` itself when nothing needs rewriting. Defined in php_inline_html.c.
+const char *cbm_php_mask_inline_html(CBMArena *arena, const char *source, int source_len);
 void cbm_extract_imports(CBMExtractCtx *ctx);
 void cbm_extract_usages(CBMExtractCtx *ctx);
 void cbm_extract_semantic(CBMExtractCtx *ctx);

internal/cbm/lsp/php_lsp.c

@@ -19,6 +19,7 @@
 
 #include "php_lsp.h"
 #include "lsp_node_iter.h"
+#include "../cbm.h"
 #include "../helpers.h"
 #include <ctype.h>
 #include <stdio.h>
@@ -4489,6 +4490,9 @@ void cbm_run_php_lsp_cross(CBMArena *arena, const char *source, int source_len,
                            TSTree *cached_tree, CBMResolvedCallArray *out) {
     if (!arena || !source || source_len <= 0 || !out)
         return;
+    /* The retained source is the file on disk; read it the way the per-file
+     * extract did, so the cached tree and node text agree (#2000). */
+    source = cbm_php_mask_inline_html(arena, source, source_len);
 
     TSParser *parser = NULL;
     TSTree *tree = cached_tree;