From b20ec087e0c528a1b9a76b7887023d6da50bac47 Mon Sep 17 00:00:00 2001
From: John-Mark Bell <jmb@netsurf-browser.org>
Date: Thu, 5 Jul 2012 21:21:14 +0100
Subject: Insert data at correct point in input stream.

---
 src/parser.c              | 10 ++--------
 src/tokeniser/tokeniser.c | 45 +++++++++++++++++++++++++++++++++++++++++++++
 src/tokeniser/tokeniser.h |  4 ++++
 3 files changed, 51 insertions(+), 8 deletions(-)

(limited to 'src')

diff --git a/src/parser.c b/src/parser.c
index 95216a3..bf6cca4 100644
--- a/src/parser.c
+++ b/src/parser.c
@@ -207,23 +207,17 @@ hubbub_error hubbub_parser_setopt(hubbub_parser *parser,
  * useful to allow hubbub callbacks to add computed data to the input.
  * 
  * \param parser  Parser instance to use
- * \param data    Data to parse (encoded in the input charset)
+ * \param data    Data to parse (encoded in UTF-8)
  * \param len     Length, in bytes, of data
  * \return HUBBUB_OK on success, appropriate error otherwise
  */
 hubbub_error hubbub_parser_insert_chunk(hubbub_parser *parser,
 		const uint8_t *data, size_t len)
 {
-	parserutils_error perror;
-
 	if (parser == NULL || data == NULL)
 		return HUBBUB_BADPARM;
 
-	perror = parserutils_inputstream_insert(parser->stream, data, len);
-	if (perror != PARSERUTILS_OK)
-		return hubbub_error_from_parserutils_error(perror);
-
-	return HUBBUB_OK;
+	return hubbub_tokeniser_insert_chunk(parser->tok, data, len);
 }
 
 /**
diff --git a/src/tokeniser/tokeniser.c b/src/tokeniser/tokeniser.c
index 949f433..549a89a 100644
--- a/src/tokeniser/tokeniser.c
+++ b/src/tokeniser/tokeniser.c
@@ -169,6 +169,7 @@ struct hubbub_tokeniser {
 
 	parserutils_inputstream *input;	/**< Input stream */
 	parserutils_buffer *buffer;	/**< Input buffer */
+	parserutils_buffer *insert_buf; /**< Stream insertion buffer */
 
 	hubbub_tokeniser_context context;	/**< Tokeniser context */
 
@@ -303,6 +304,13 @@ hubbub_error hubbub_tokeniser_create(parserutils_inputstream *input,
 		return hubbub_error_from_parserutils_error(perror);
 	}
 
+	perror = parserutils_buffer_create(alloc, pw, &tok->insert_buf);
+	if (perror != PARSERUTILS_OK) {
+		parserutils_buffer_destroy(tok->buffer);
+		alloc(tok, 0, pw);
+		return hubbub_error_from_parserutils_error(perror);
+	}
+
 	tok->state = STATE_DATA;
 	tok->content_model = HUBBUB_CONTENT_MODEL_PCDATA;
 
@@ -343,6 +351,8 @@ hubbub_error hubbub_tokeniser_destroy(hubbub_tokeniser *tokeniser)
 				0, tokeniser->alloc_pw);
 	}
 
+	parserutils_buffer_destroy(tokeniser->insert_buf);
+
 	parserutils_buffer_destroy(tokeniser->buffer);
 
 	tokeniser->alloc(tokeniser, 0, tokeniser->alloc_pw);
@@ -385,6 +395,32 @@ hubbub_error hubbub_tokeniser_setopt(hubbub_tokeniser *tokeniser,
 	return HUBBUB_OK;
 }
 
+/**
+ * Insert a chunk of data into the input stream.
+ *
+ * Inserts the given data into the input stream ready for parsing but
+ * does not cause any additional processing of the input.
+ *
+ * \param tokeniser  Tokeniser instance
+ * \param data       Data to insert (UTF-8 encoded)
+ * \param len        Length, in bytes, of data
+ * \return HUBBUB_OK on success, appropriate error otherwise
+ */
+hubbub_error hubbub_tokeniser_insert_chunk(hubbub_tokeniser *tokeniser,
+		const uint8_t *data, size_t len)
+{
+	parserutils_error perror;
+
+	if (tokeniser == NULL || data == NULL)
+		return HUBBUB_BADPARM;
+
+	perror = parserutils_buffer_append(tokeniser->insert_buf, data, len);
+	if (perror != PARSERUTILS_OK)
+		return hubbub_error_from_parserutils_error(perror);
+
+	return HUBBUB_OK;
+}
+
 /**
  * Process remaining data in the input stream
  *
@@ -3310,6 +3346,7 @@ hubbub_error hubbub_tokeniser_emit_token(hubbub_tokeniser *tokeniser,
 
 	assert(tokeniser != NULL);
 	assert(token != NULL);
+	assert(tokeniser->insert_buf->length == 0);
 
 #ifndef NDEBUG
 	/* Sanity checks */
@@ -3371,5 +3408,13 @@ hubbub_error hubbub_tokeniser_emit_token(hubbub_tokeniser *tokeniser,
 		tokeniser->context.pending = 0;
 	}
 
+	if (tokeniser->insert_buf->length > 0) {
+		parserutils_inputstream_insert(tokeniser->input,
+				tokeniser->insert_buf->data,
+				tokeniser->insert_buf->length);
+		parserutils_buffer_discard(tokeniser->insert_buf, 0,
+				tokeniser->insert_buf->length);
+	}
+
 	return err;
 }
diff --git a/src/tokeniser/tokeniser.h b/src/tokeniser/tokeniser.h
index 4a9f362..e13a03a 100644
--- a/src/tokeniser/tokeniser.h
+++ b/src/tokeniser/tokeniser.h
@@ -62,6 +62,10 @@ hubbub_error hubbub_tokeniser_setopt(hubbub_tokeniser *tokeniser,
 		hubbub_tokeniser_opttype type,
 		hubbub_tokeniser_optparams *params);
 
+/* Insert a chunk of data into the input stream */
+hubbub_error hubbub_tokeniser_insert_chunk(hubbub_tokeniser *tokeniser,
+		const uint8_t *data, size_t len);
+
 /* Process remaining data in the input stream */
 hubbub_error hubbub_tokeniser_run(hubbub_tokeniser *tokeniser);
 
-- 
cgit v1.2.3