From bdff981eef4a7434291941e6bca662357498bc59 Mon Sep 17 00:00:00 2001
From: Vincent Sanders <vince@netsurf-browser.org>
Date: Tue, 10 Jul 2012 11:23:57 -0600
Subject: Add ability to pause tokenisation

---
 src/parser.c              | 12 ++++++++++++
 src/tokeniser/tokeniser.c | 24 +++++++++++++++++++++++-
 src/tokeniser/tokeniser.h |  5 ++++-
 src/utils/errors.c        |  3 +++
 4 files changed, 42 insertions(+), 2 deletions(-)

(limited to 'src')

diff --git a/src/parser.c b/src/parser.c
index 8da9d79..5423922 100644
--- a/src/parser.c
+++ b/src/parser.c
@@ -152,6 +152,7 @@ hubbub_error hubbub_parser_setopt(hubbub_parser *parser,
 				HUBBUB_TOKENISER_TOKEN_HANDLER,
 				(hubbub_tokeniser_optparams *) params);
 		break;
+
 	case HUBBUB_PARSER_ERROR_HANDLER:
 		/* The error handler does not cascade, so tell both the
 		 * treebuilder (if extant) and the tokeniser. */
@@ -166,11 +167,19 @@ hubbub_error hubbub_parser_setopt(hubbub_parser *parser,
 					(hubbub_tokeniser_optparams *) params);
 		}
 		break;
+
 	case HUBBUB_PARSER_CONTENT_MODEL:
 		result = hubbub_tokeniser_setopt(parser->tok,
 				HUBBUB_TOKENISER_CONTENT_MODEL,
 				(hubbub_tokeniser_optparams *) params);
 		break;
+
+	case HUBBUB_PARSER_PAUSE:
+		result = hubbub_tokeniser_setopt(parser->tok,
+				HUBBUB_TOKENISER_PAUSE,
+				(hubbub_tokeniser_optparams *) params);
+		break;
+
 	case HUBBUB_PARSER_TREE_HANDLER:
 		if (parser->tb != NULL) {
 			result = hubbub_treebuilder_setopt(parser->tb,
@@ -178,6 +187,7 @@ hubbub_error hubbub_parser_setopt(hubbub_parser *parser,
 					(hubbub_treebuilder_optparams *) params);
 		}
 		break;
+
 	case HUBBUB_PARSER_DOCUMENT_NODE:
 		if (parser->tb != NULL) {
 			result = hubbub_treebuilder_setopt(parser->tb,
@@ -185,6 +195,7 @@ hubbub_error hubbub_parser_setopt(hubbub_parser *parser,
 					(hubbub_treebuilder_optparams *) params);
 		}
 		break;
+
 	case HUBBUB_PARSER_ENABLE_SCRIPTING:
 		if (parser->tb != NULL) {
 			result = hubbub_treebuilder_setopt(parser->tb,
@@ -192,6 +203,7 @@ hubbub_error hubbub_parser_setopt(hubbub_parser *parser,
 					(hubbub_treebuilder_optparams *) params);
 		}
 		break;
+
 	default:
 		result = HUBBUB_INVALID;
 	}
diff --git a/src/tokeniser/tokeniser.c b/src/tokeniser/tokeniser.c
index 549a89a..04f4ea7 100644
--- a/src/tokeniser/tokeniser.c
+++ b/src/tokeniser/tokeniser.c
@@ -16,6 +16,7 @@
 #include "utils/parserutilserror.h"
 #include "utils/utils.h"
 
+#include "hubbub/errors.h"
 #include "tokeniser/entities.h"
 #include "tokeniser/tokeniser.h"
 
@@ -155,6 +156,7 @@ typedef struct hubbub_tokeniser_context {
 	} position;				/**< Position in source data */
 
 	uint32_t allowed_char;			/**< Used for quote matching */
+
 } hubbub_tokeniser_context;
 
 /**
@@ -166,6 +168,7 @@ struct hubbub_tokeniser {
 						 * model flag */
 	bool escape_flag;		/**< Escape flag **/
 	bool process_cdata_section;	/**< Whether to process CDATA sections*/
+	bool paused; /**< flag for if parsing is currently paused */
 
 	parserutils_inputstream *input;	/**< Input stream */
 	parserutils_buffer *buffer;	/**< Input buffer */
@@ -372,6 +375,8 @@ hubbub_error hubbub_tokeniser_setopt(hubbub_tokeniser *tokeniser,
 		hubbub_tokeniser_opttype type,
 		hubbub_tokeniser_optparams *params)
 {
+	hubbub_error err = HUBBUB_OK;
+
 	if (tokeniser == NULL || params == NULL)
 		return HUBBUB_BADPARM;
 
@@ -390,9 +395,18 @@ hubbub_error hubbub_tokeniser_setopt(hubbub_tokeniser *tokeniser,
 	case HUBBUB_TOKENISER_PROCESS_CDATA:
 		tokeniser->process_cdata_section = params->process_cdata;
 		break;
+	case HUBBUB_TOKENISER_PAUSE:
+		if (params->pause_parse == true) {
+			tokeniser->paused = true;
+		} else {
+			if (tokeniser->paused == true) {
+				tokeniser->paused = false;
+				err = hubbub_tokeniser_run(tokeniser);
+			}
+		}
 	}
 
-	return HUBBUB_OK;
+	return err;
 }
 
 /**
@@ -434,6 +448,9 @@ hubbub_error hubbub_tokeniser_run(hubbub_tokeniser *tokeniser)
 	if (tokeniser == NULL)
 		return HUBBUB_BADPARM;
 
+	if (tokeniser->paused == true)
+		return HUBBUB_PAUSED;
+
 #if 0
 #define state(x) \
 		case x: \
@@ -3416,5 +3433,10 @@ hubbub_error hubbub_tokeniser_emit_token(hubbub_tokeniser *tokeniser,
 				tokeniser->insert_buf->length);
 	}
 
+	/* Ensure callback can pause the tokenise */
+	if (err == HUBBUB_PAUSED) {
+		tokeniser->paused = true;
+	}
+
 	return err;
 }
diff --git a/src/tokeniser/tokeniser.h b/src/tokeniser/tokeniser.h
index e13a03a..ed46d60 100644
--- a/src/tokeniser/tokeniser.h
+++ b/src/tokeniser/tokeniser.h
@@ -26,7 +26,8 @@ typedef enum hubbub_tokeniser_opttype {
 	HUBBUB_TOKENISER_TOKEN_HANDLER,
 	HUBBUB_TOKENISER_ERROR_HANDLER,
 	HUBBUB_TOKENISER_CONTENT_MODEL,
-	HUBBUB_TOKENISER_PROCESS_CDATA
+	HUBBUB_TOKENISER_PROCESS_CDATA,
+	HUBBUB_TOKENISER_PAUSE
 } hubbub_tokeniser_opttype;
 
 /**
@@ -48,6 +49,8 @@ typedef union hubbub_tokeniser_optparams {
 	} content_model;		/**< Current content model */
 
 	bool process_cdata;		/**< Whether to process CDATA sections*/
+
+	bool pause_parse;		/**< Pause parsing */
 } hubbub_tokeniser_optparams;
 
 /* Create a hubbub tokeniser */
diff --git a/src/utils/errors.c b/src/utils/errors.c
index 766b181..e7a797f 100644
--- a/src/utils/errors.c
+++ b/src/utils/errors.c
@@ -29,6 +29,9 @@ const char *hubbub_error_to_string(hubbub_error error)
 	case HUBBUB_ENCODINGCHANGE:
 		result = "Encoding of document has changed";
 		break;
+	case HUBBUB_PAUSED:
+		result = "Parser is paused";
+		break;
 	case HUBBUB_NOMEM:
 		result = "Insufficient memory";
 		break;
-- 
cgit v1.2.3