Be clearer about when jsonapi's need_escapes is needed

adunstan · adunstan · commit ea5ff5833c7d · 2025-01-19T09:09:58.000-05:00
Most operations beyond pure json parsing need to set need_escapes to true to get access to field names and string scalars. Document this fact more explicitly. Slightly tweaked patch from: Author: Corey Huinker <corey.huinker@gmail.com> Discussion: https://postgr.es/m/CADkLM=c49Vkfg2+A8ubSuEtaGEjuaKZXCA6SrXA8kdwHjx3uxQ@mail.gmail.com
diff --git a/src/common/jsonapi.c b/src/common/jsonapi.c
@@ -1297,7 +1297,10 @@ parse_scalar(JsonLexContext *lex, const JsonSemAction *sem)
 		return result;
 	}
 
-	/* invoke the callback, which may take ownership of val */
+	/*
+	 * invoke the callback, which may take ownership of val. For string
+	 * values, val is NULL if need_escapes is false.
+	 */
 	result = (*sfunc) (sem->semstate, val, tok);
 
 	if (lex->flags & JSONLEX_CTX_OWNS_TOKENS)
@@ -1326,6 +1329,7 @@ parse_object_field(JsonLexContext *lex, const JsonSemAction *sem)
 		return report_parse_error(JSON_PARSE_STRING, lex);
 	if ((ostart != NULL || oend != NULL) && lex->need_escapes)
 	{
+		/* fname is NULL if need_escapes is false */
 		fname = STRDUP(lex->strval->data);
 		if (fname == NULL)
 			return JSON_OUT_OF_MEMORY;
diff --git a/src/include/common/jsonapi.h b/src/include/common/jsonapi.h
@@ -118,6 +118,12 @@ typedef struct JsonLexContext
 	struct jsonapi_StrValType *errormsg;
 } JsonLexContext;
 
+/*
+ * Function types for custom json parsing actions.
+ *
+ * fname will be NULL if the context has need_escapes=false, as will token for
+ * string type values.
+ */
 typedef JsonParseErrorType (*json_struct_action) (void *state);
 typedef JsonParseErrorType (*json_ofield_action) (void *state, char *fname, bool isnull);
 typedef JsonParseErrorType (*json_aelem_action) (void *state, bool isnull);
@@ -197,12 +203,17 @@ extern JsonParseErrorType json_count_array_elements(JsonLexContext *lex,
  * struct is allocated.
  *
  * If need_escapes is true, ->strval stores the unescaped lexemes.
+ *
+ * Setting need_escapes to true is necessary if the operation needs
+ * to reference field names or scalar string values. This is true of most
+ * operations beyond purely checking the json-validity of the source
+ * document.
+ *
  * Unescaping is expensive, so only request it when necessary.
  *
  * If need_escapes is true or lex was given as NULL, then the caller is
  * responsible for freeing the returned struct, either by calling
- * freeJsonLexContext() or (in backend environment) via memory context
- * cleanup.
+ * freeJsonLexContext() or (in backends) via memory context cleanup.
  */
 extern JsonLexContext *makeJsonLexContextCstringLen(JsonLexContext *lex,
 													const char *json,
