Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(22)

Issues Search
    My Issues | Starred     Open | Closed | All

Unified Diff: third_party/libxml/patches/icu

Issue 2951008: Update libxml to 2.7.7. (Closed) Base URL: http://src.chromium.org/git/chromium.git
Patch Set: Created 10 years, 5 months ago
Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.
Jump to:
View side-by-side diff with in-line comments
Download patch
« no previous file with comments | « third_party/libxml/patches/LoadLibraryA ('k') | third_party/libxml/patches/icu-configure » ('j') | no next file with comments »
Expand Comments ('e') | Collapse Comments ('c') | Show Comments Hide Comments ('s')
Index: third_party/libxml/patches/icu
diff --git a/third_party/libxml/patches/icu b/third_party/libxml/patches/icu
new file mode 100644
index 0000000000000000000000000000000000000000..814d23c9832a9ad6db9eafad58449307814a144f
--- /dev/null
+++ b/third_party/libxml/patches/icu
@@ -0,0 +1,892 @@
+Add code support for ICU.
+
+diff --git a/third_party/libxml/encoding.c b/third_party/libxml/encoding.c
+index b86a547..0f41df9 100644
+--- a/third_party/libxml/encoding.c
++++ b/third_party/libxml/encoding.c
+@@ -58,7 +58,7 @@ static xmlCharEncodingAliasPtr xmlCharEncodingAliases = NULL;
+ static int xmlCharEncodingAliasesNb = 0;
+ static int xmlCharEncodingAliasesMax = 0;
+
+-#ifdef LIBXML_ICONV_ENABLED
++#if defined(LIBXML_ICONV_ENABLED) || defined(LIBXML_ICU_ENABLED)
+ #if 0
+ #define DEBUG_ENCODING /* Define this to get encoding traces */
+ #endif
+@@ -97,6 +97,54 @@ xmlEncodingErr(xmlParserErrors error, const char *msg, const char *val)
+ NULL, 0, val, NULL, NULL, 0, 0, msg, val);
+ }
+
++#ifdef LIBXML_ICU_ENABLED
++static uconv_t*
++openIcuConverter(const char* name, int toUnicode)
++{
++ UErrorCode status = U_ZERO_ERROR;
++ uconv_t *conv = (uconv_t *) xmlMalloc(sizeof(uconv_t));
++ if (conv == NULL)
++ return NULL;
++
++ conv->uconv = ucnv_open(name, &status);
++ if (U_FAILURE(status))
++ goto error;
++
++ status = U_ZERO_ERROR;
++ if (toUnicode) {
++ ucnv_setToUCallBack(conv->uconv, UCNV_TO_U_CALLBACK_STOP,
++ NULL, NULL, NULL, &status);
++ }
++ else {
++ ucnv_setFromUCallBack(conv->uconv, UCNV_FROM_U_CALLBACK_STOP,
++ NULL, NULL, NULL, &status);
++ }
++ if (U_FAILURE(status))
++ goto error;
++
++ status = U_ZERO_ERROR;
++ conv->utf8 = ucnv_open("UTF-8", &status);
++ if (U_SUCCESS(status))
++ return conv;
++
++error:
++ if (conv->uconv)
++ ucnv_close(conv->uconv);
++ xmlFree(conv);
++ return NULL;
++}
++
++static void
++closeIcuConverter(uconv_t *conv)
++{
++ if (conv != NULL) {
++ ucnv_close(conv->uconv);
++ ucnv_close(conv->utf8);
++ xmlFree(conv);
++ }
++}
++#endif /* LIBXML_ICU_ENABLED */
++
+ /************************************************************************
+ * *
+ * Conversions To/From UTF8 encoding *
+@@ -1306,7 +1354,11 @@ xmlNewCharEncodingHandler(const char *name,
+ #ifdef LIBXML_ICONV_ENABLED
+ handler->iconv_in = NULL;
+ handler->iconv_out = NULL;
+-#endif /* LIBXML_ICONV_ENABLED */
++#endif
++#ifdef LIBXML_ICU_ENABLED
++ handler->uconv_in = NULL;
++ handler->uconv_out = NULL;
++#endif
+
+ /*
+ * registers and returns the handler.
+@@ -1371,7 +1423,7 @@ xmlInitCharEncodingHandlers(void) {
+ xmlNewCharEncodingHandler("ASCII", asciiToUTF8, NULL);
+ xmlNewCharEncodingHandler("US-ASCII", asciiToUTF8, NULL);
+ #endif /* LIBXML_OUTPUT_ENABLED */
+-#ifndef LIBXML_ICONV_ENABLED
++#if !defined(LIBXML_ICONV_ENABLED) && !defined(LIBXML_ICU_ENABLED)
+ #ifdef LIBXML_ISO8859X_ENABLED
+ xmlRegisterCharEncodingHandlersISO8859x ();
+ #endif
+@@ -1578,6 +1630,10 @@ xmlFindCharEncodingHandler(const char *name) {
+ xmlCharEncodingHandlerPtr enc;
+ iconv_t icv_in, icv_out;
+ #endif /* LIBXML_ICONV_ENABLED */
++#ifdef LIBXML_ICU_ENABLED
++ xmlCharEncodingHandlerPtr enc;
++ uconv_t *ucv_in, *ucv_out;
++#endif /* LIBXML_ICU_ENABLED */
+ char upper[100];
+ int i;
+
+@@ -1647,6 +1703,35 @@ xmlFindCharEncodingHandler(const char *name) {
+ "iconv : problems with filters for '%s'\n", name);
+ }
+ #endif /* LIBXML_ICONV_ENABLED */
++#ifdef LIBXML_ICU_ENABLED
++ /* check whether icu can handle this */
++ ucv_in = openIcuConverter(name, 1);
++ ucv_out = openIcuConverter(name, 0);
++ if (ucv_in != NULL && ucv_out != NULL) {
++ enc = (xmlCharEncodingHandlerPtr)
++ xmlMalloc(sizeof(xmlCharEncodingHandler));
++ if (enc == NULL) {
++ closeIcuConverter(ucv_in);
++ closeIcuConverter(ucv_out);
++ return(NULL);
++ }
++ enc->name = xmlMemStrdup(name);
++ enc->input = NULL;
++ enc->output = NULL;
++ enc->uconv_in = ucv_in;
++ enc->uconv_out = ucv_out;
++#ifdef DEBUG_ENCODING
++ xmlGenericError(xmlGenericErrorContext,
++ "Found ICU converter handler for encoding %s\n", name);
++#endif
++ return enc;
++ } else if (ucv_in != NULL || ucv_out != NULL) {
++ closeIcuConverter(ucv_in);
++ closeIcuConverter(ucv_out);
++ xmlEncodingErr(XML_ERR_INTERNAL_ERROR,
++ "ICU converter : problems with filters for '%s'\n", name);
++ }
++#endif /* LIBXML_ICU_ENABLED */
+
+ #ifdef DEBUG_ENCODING
+ xmlGenericError(xmlGenericErrorContext,
+@@ -1737,6 +1822,75 @@ xmlIconvWrapper(iconv_t cd, unsigned char *out, int *outlen,
+
+ /************************************************************************
+ * *
++ * ICU based generic conversion functions *
++ * *
++ ************************************************************************/
++
++#ifdef LIBXML_ICU_ENABLED
++/**
++ * xmlUconvWrapper:
++ * @cd: ICU uconverter data structure
++ * @toUnicode : non-zero if toUnicode. 0 otherwise.
++ * @out: a pointer to an array of bytes to store the result
++ * @outlen: the length of @out
++ * @in: a pointer to an array of ISO Latin 1 chars
++ * @inlen: the length of @in
++ *
++ * Returns 0 if success, or
++ * -1 by lack of space, or
++ * -2 if the transcoding fails (for *in is not valid utf8 string or
++ * the result of transformation can't fit into the encoding we want), or
++ * -3 if there the last byte can't form a single output char.
++ *
++ * The value of @inlen after return is the number of octets consumed
++ * as the return value is positive, else unpredictable.
++ * The value of @outlen after return is the number of ocetes consumed.
++ */
++static int
++xmlUconvWrapper(uconv_t *cd, int toUnicode, unsigned char *out, int *outlen,
++ const unsigned char *in, int *inlen) {
++ const char *ucv_in = (const char *) in;
++ char *ucv_out = (char *) out;
++ UErrorCode err = U_ZERO_ERROR;
++
++ if ((out == NULL) || (outlen == NULL) || (inlen == NULL) || (in == NULL)) {
++ if (outlen != NULL) *outlen = 0;
++ return(-1);
++ }
++
++ /*
++ * TODO(jungshik)
++ * 1. is ucnv_convert(To|From)Algorithmic better?
++ * 2. had we better use an explicit pivot buffer?
++ * 3. error returned comes from 'fromUnicode' only even
++ * when toUnicode is true !
++ */
++ if (toUnicode) {
++ /* encoding => UTF-16 => UTF-8 */
++ ucnv_convertEx(cd->utf8, cd->uconv, &ucv_out, ucv_out + *outlen,
++ &ucv_in, ucv_in + *inlen, NULL, NULL, NULL, NULL,
++ 0, TRUE, &err);
++ } else {
++ /* UTF-8 => UTF-16 => encoding */
++ ucnv_convertEx(cd->uconv, cd->utf8, &ucv_out, ucv_out + *outlen,
++ &ucv_in, ucv_in + *inlen, NULL, NULL, NULL, NULL,
++ 0, TRUE, &err);
++ }
++ *inlen = ucv_in - (const char*) in;
++ *outlen = ucv_out - (char *) out;
++ if (U_SUCCESS(err))
++ return 0;
++ if (err == U_BUFFER_OVERFLOW_ERROR)
++ return -1;
++ if (err == U_INVALID_CHAR_FOUND || err == U_ILLEGAL_CHAR_FOUND)
++ return -2;
++ /* if (err == U_TRUNCATED_CHAR_FOUND) */
++ return -3;
++}
++#endif /* LIBXML_ICU_ENABLED */
++
++/************************************************************************
++ * *
+ * The real API used by libxml for on-the-fly conversion *
+ * *
+ ************************************************************************/
+@@ -1810,6 +1964,16 @@ xmlCharEncFirstLineInt(xmlCharEncodingHandler *handler, xmlBufferPtr out,
+ if (ret == -1) ret = -3;
+ }
+ #endif /* LIBXML_ICONV_ENABLED */
++#ifdef LIBXML_ICU_ENABLED
++ else if (handler->uconv_in != NULL) {
++ ret = xmlUconvWrapper(handler->uconv_in, 1, &out->content[out->use],
++ &written, in->content, &toconv);
++ xmlBufferShrink(in, toconv);
++ out->use += written;
++ out->content[out->use] = 0;
++ if (ret == -1) ret = -3;
++ }
++#endif /* LIBXML_ICU_ENABLED */
+ #ifdef DEBUG_ENCODING
+ switch (ret) {
+ case 0:
+@@ -1915,6 +2079,17 @@ xmlCharEncInFunc(xmlCharEncodingHandler * handler, xmlBufferPtr out,
+ ret = -3;
+ }
+ #endif /* LIBXML_ICONV_ENABLED */
++#ifdef LIBXML_ICU_ENABLED
++ else if (handler->uconv_in != NULL) {
++ ret = xmlUconvWrapper(handler->uconv_in, 1, &out->content[out->use],
++ &written, in->content, &toconv);
++ xmlBufferShrink(in, toconv);
++ out->use += written;
++ out->content[out->use] = 0;
++ if (ret == -1)
++ ret = -3;
++ }
++#endif /* LIBXML_ICU_ENABLED */
+ switch (ret) {
+ case 0:
+ #ifdef DEBUG_ENCODING
+@@ -2015,6 +2190,15 @@ retry:
+ out->content[out->use] = 0;
+ }
+ #endif /* LIBXML_ICONV_ENABLED */
++#ifdef LIBXML_ICU_ENABLED
++ else if (handler->uconv_out != NULL) {
++ ret = xmlUconvWrapper(handler->uconv_out, 0,
++ &out->content[out->use],
++ &written, NULL, &toconv);
++ out->use += written;
++ out->content[out->use] = 0;
++ }
++#endif /* LIBXML_ICU_ENABLED */
+ #ifdef DEBUG_ENCODING
+ xmlGenericError(xmlGenericErrorContext,
+ "initialized encoder\n");
+@@ -2061,6 +2245,26 @@ retry:
+ }
+ }
+ #endif /* LIBXML_ICONV_ENABLED */
++#ifdef LIBXML_ICU_ENABLED
++ else if (handler->uconv_out != NULL) {
++ ret = xmlUconvWrapper(handler->uconv_out, 0,
++ &out->content[out->use],
++ &written, in->content, &toconv);
++ xmlBufferShrink(in, toconv);
++ out->use += written;
++ writtentot += written;
++ out->content[out->use] = 0;
++ if (ret == -1) {
++ if (written > 0) {
++ /*
++ * Can be a limitation of iconv
++ */
++ goto retry;
++ }
++ ret = -3;
++ }
++ }
++#endif /* LIBXML_ICU_ENABLED */
+ else {
+ xmlEncodingErr(XML_I18N_NO_OUTPUT,
+ "xmlCharEncOutFunc: no output function !\n", NULL);
+@@ -2173,6 +2377,22 @@ xmlCharEncCloseFunc(xmlCharEncodingHandler *handler) {
+ xmlFree(handler);
+ }
+ #endif /* LIBXML_ICONV_ENABLED */
++#ifdef LIBXML_ICU_ENABLED
++ if ((handler->uconv_out != NULL) || (handler->uconv_in != NULL)) {
++ if (handler->name != NULL)
++ xmlFree(handler->name);
++ handler->name = NULL;
++ if (handler->uconv_out != NULL) {
++ closeIcuConverter(handler->uconv_out);
++ handler->uconv_out = NULL;
++ }
++ if (handler->uconv_in != NULL) {
++ closeIcuConverter(handler->uconv_in);
++ handler->uconv_in = NULL;
++ }
++ xmlFree(handler);
++ }
++#endif
+ #ifdef DEBUG_ENCODING
+ if (ret)
+ xmlGenericError(xmlGenericErrorContext,
+@@ -2248,6 +2468,22 @@ xmlByteConsumed(xmlParserCtxtPtr ctxt) {
+ cur += toconv;
+ } while (ret == -2);
+ #endif
++#ifdef LIBXML_ICU_ENABLED
++ } else if (handler->uconv_out != NULL) {
++ do {
++ toconv = in->end - cur;
++ written = 32000;
++ ret = xmlUconvWrapper(handler->uconv_out, 0, &convbuf[0],
++ &written, cur, &toconv);
++ if (ret < 0) {
++ if (written > 0)
++ ret = -2;
++ else
++ return(-1);
++ }
++ unused += written;
++ cur += toconv;
++ } while (ret == -2);
+ } else {
+ /* could not find a converter */
+ return(-1);
+@@ -2259,8 +2495,9 @@ xmlByteConsumed(xmlParserCtxtPtr ctxt) {
+ }
+ return(in->consumed + (in->cur - in->base));
+ }
++#endif
+
+-#ifndef LIBXML_ICONV_ENABLED
++#if !defined(LIBXML_ICONV_ENABLED) && !defined(LIBXML_ICU_ENABLED)
+ #ifdef LIBXML_ISO8859X_ENABLED
+
+ /**
+diff --git a/third_party/libxml/include/libxml/encoding.h b/third_party/libxml/include/libxml/encoding.h
+index c74b25f..b5f8b48 100644
+--- a/third_party/libxml/include/libxml/encoding.h
++++ b/third_party/libxml/include/libxml/encoding.h
+@@ -26,6 +26,24 @@
+
+ #ifdef LIBXML_ICONV_ENABLED
+ #include <iconv.h>
++#else
++#ifdef LIBXML_ICU_ENABLED
++#include <unicode/ucnv.h>
++#if 0
++/* Forward-declare UConverter here rather than pulling in <unicode/ucnv.h>
++ * to prevent unwanted ICU symbols being exposed to users of libxml2.
++ * One particular case is Qt4 conflicting on UChar32.
++ */
++#include <stdint.h>
++struct UConverter;
++typedef struct UConverter UConverter;
++#ifdef _MSC_VER
++typedef wchar_t UChar;
++#else
++typedef uint16_t UChar;
++#endif
++#endif
++#endif
+ #endif
+ #ifdef __cplusplus
+ extern "C" {
+@@ -125,6 +143,13 @@ typedef int (* xmlCharEncodingOutputFunc)(unsigned char *out, int *outlen,
+ * Block defining the handlers for non UTF-8 encodings.
+ * If iconv is supported, there are two extra fields.
+ */
++#ifdef LIBXML_ICU_ENABLED
++struct _uconv_t {
++ UConverter *uconv; /* for conversion between an encoding and UTF-16 */
++ UConverter *utf8; /* for conversion between UTF-8 and UTF-16 */
++};
++typedef struct _uconv_t uconv_t;
++#endif
+
+ typedef struct _xmlCharEncodingHandler xmlCharEncodingHandler;
+ typedef xmlCharEncodingHandler *xmlCharEncodingHandlerPtr;
+@@ -136,6 +161,10 @@ struct _xmlCharEncodingHandler {
+ iconv_t iconv_in;
+ iconv_t iconv_out;
+ #endif /* LIBXML_ICONV_ENABLED */
++#ifdef LIBXML_ICU_ENABLED
++ uconv_t *uconv_in;
++ uconv_t *uconv_out;
++#endif /* LIBXML_ICU_ENABLED */
+ };
+
+ #ifdef __cplusplus
+diff --git a/third_party/libxml/include/libxml/parser.h b/third_party/libxml/include/libxml/parser.h
+index dd79c42..3580b63 100644
+--- a/third_party/libxml/include/libxml/parser.h
++++ b/third_party/libxml/include/libxml/parser.h
+@@ -1222,6 +1222,7 @@ typedef enum {
+ XML_WITH_DEBUG_MEM = 29,
+ XML_WITH_DEBUG_RUN = 30,
+ XML_WITH_ZLIB = 31,
++ XML_WITH_ICU = 32,
+ XML_WITH_NONE = 99999 /* just to be sure of allocation size */
+ } xmlFeature;
+
+diff --git a/third_party/libxml/include/libxml/xmlversion.h.in b/third_party/libxml/include/libxml/xmlversion.h.in
+index 4739f3a..de310ab 100644
+--- a/third_party/libxml/include/libxml/xmlversion.h.in
++++ b/third_party/libxml/include/libxml/xmlversion.h.in
+@@ -269,6 +269,15 @@ XMLPUBFUN void XMLCALL xmlCheckVersion(int version);
+ #endif
+
+ /**
++ * LIBXML_ICU_ENABLED:
++ *
++ * Whether icu support is available
++ */
++#if @WITH_ICU@
++#define LIBXML_ICU_ENABLED
++#endif
++
++/**
+ * LIBXML_ISO8859X_ENABLED:
+ *
+ * Whether ISO-8859-* support is made available in case iconv is not
+diff --git a/third_party/libxml/parser.c b/third_party/libxml/parser.c
+index 85e7599..3ba2a06 100644
+--- a/third_party/libxml/parser.c
++++ b/third_party/libxml/parser.c
+@@ -954,6 +954,12 @@ xmlHasFeature(xmlFeature feature)
+ #else
+ return(0);
+ #endif
++ case XML_WITH_ICU:
++#ifdef LIBXML_ICU_ENABLED
++ return(1);
++#else
++ return(0);
++#endif
+ default:
+ break;
+ }
+diff --git a/third_party/libxml/patches/icu b/third_party/libxml/patches/icu
jungshik at Google 2010/07/12 18:23:29 I guess you don't want to include this diff in the
+index 324cea3..6c22c3c 100644
+--- a/third_party/libxml/patches/icu
++++ b/third_party/libxml/patches/icu
+@@ -1,434 +0,0 @@
+-Code support for ICU. Note that this relies on modifications to the
+-build environment (either configure or configure.js on Windows).
+-
+-Index: libxml/encoding.c
+-===================================================================
+---- libxml.orig/encoding.c 2010-07-09 14:48:28.881863834 -0700
+-+++ libxml/encoding.c 2010-07-09 14:49:23.479741318 -0700
+-@@ -58,7 +58,7 @@
+- static int xmlCharEncodingAliasesNb = 0;
+- static int xmlCharEncodingAliasesMax = 0;
+-
+--#ifdef LIBXML_ICONV_ENABLED
+-+#if defined(LIBXML_ICONV_ENABLED) || defined(LIBXML_ICU_ENABLED)
+- #if 0
+- #define DEBUG_ENCODING /* Define this to get encoding traces */
+- #endif
+-@@ -97,6 +97,54 @@
+- NULL, 0, val, NULL, NULL, 0, 0, msg, val);
+- }
+-
+-+#ifdef LIBXML_ICU_ENABLED
+-+static uconv_t*
+-+openIcuConverter(const char* name, int toUnicode)
+-+{
+-+ UErrorCode status = U_ZERO_ERROR;
+-+ uconv_t *conv = (uconv_t *) xmlMalloc(sizeof(uconv_t));
+-+ if (conv == NULL)
+-+ return NULL;
+-+
+-+ conv->uconv = ucnv_open(name, &status);
+-+ if (U_FAILURE(status))
+-+ goto error;
+-+
+-+ status = U_ZERO_ERROR;
+-+ if (toUnicode) {
+-+ ucnv_setToUCallBack(conv->uconv, UCNV_TO_U_CALLBACK_STOP,
+-+ NULL, NULL, NULL, &status);
+-+ }
+-+ else {
+-+ ucnv_setFromUCallBack(conv->uconv, UCNV_FROM_U_CALLBACK_STOP,
+-+ NULL, NULL, NULL, &status);
+-+ }
+-+ if (U_FAILURE(status))
+-+ goto error;
+-+
+-+ status = U_ZERO_ERROR;
+-+ conv->utf8 = ucnv_open("UTF-8", &status);
+-+ if (U_SUCCESS(status))
+-+ return conv;
+-+
+-+error:
+-+ if (conv->uconv)
+-+ ucnv_close(conv->uconv);
+-+ xmlFree(conv);
+-+ return NULL;
+-+}
+-+
+-+static void
+-+closeIcuConverter(uconv_t *conv)
+-+{
+-+ if (conv != NULL) {
+-+ ucnv_close(conv->uconv);
+-+ ucnv_close(conv->utf8);
+-+ xmlFree(conv);
+-+ }
+-+}
+-+#endif /* LIBXML_ICU_ENABLED */
+-+
+- /************************************************************************
+- * *
+- * Conversions To/From UTF8 encoding *
+-@@ -1306,7 +1354,11 @@
+- #ifdef LIBXML_ICONV_ENABLED
+- handler->iconv_in = NULL;
+- handler->iconv_out = NULL;
+--#endif /* LIBXML_ICONV_ENABLED */
+-+#endif
+-+#ifdef LIBXML_ICU_ENABLED
+-+ handler->uconv_in = NULL;
+-+ handler->uconv_out = NULL;
+-+#endif
+-
+- /*
+- * registers and returns the handler.
+-@@ -1371,7 +1423,7 @@
+- xmlNewCharEncodingHandler("ASCII", asciiToUTF8, NULL);
+- xmlNewCharEncodingHandler("US-ASCII", asciiToUTF8, NULL);
+- #endif /* LIBXML_OUTPUT_ENABLED */
+--#ifndef LIBXML_ICONV_ENABLED
+-+#if !defined(LIBXML_ICONV_ENABLED) && !defined(LIBXML_ICU_ENABLED)
+- #ifdef LIBXML_ISO8859X_ENABLED
+- xmlRegisterCharEncodingHandlersISO8859x ();
+- #endif
+-@@ -1578,6 +1630,10 @@
+- xmlCharEncodingHandlerPtr enc;
+- iconv_t icv_in, icv_out;
+- #endif /* LIBXML_ICONV_ENABLED */
+-+#ifdef LIBXML_ICU_ENABLED
+-+ xmlCharEncodingHandlerPtr enc;
+-+ uconv_t *ucv_in, *ucv_out;
+-+#endif /* LIBXML_ICU_ENABLED */
+- char upper[100];
+- int i;
+-
+-@@ -1647,6 +1703,35 @@
+- "iconv : problems with filters for '%s'\n", name);
+- }
+- #endif /* LIBXML_ICONV_ENABLED */
+-+#ifdef LIBXML_ICU_ENABLED
+-+ /* check whether icu can handle this */
+-+ ucv_in = openIcuConverter(name, 1);
+-+ ucv_out = openIcuConverter(name, 0);
+-+ if (ucv_in != NULL && ucv_out != NULL) {
+-+ enc = (xmlCharEncodingHandlerPtr)
+-+ xmlMalloc(sizeof(xmlCharEncodingHandler));
+-+ if (enc == NULL) {
+-+ closeIcuConverter(ucv_in);
+-+ closeIcuConverter(ucv_out);
+-+ return(NULL);
+-+ }
+-+ enc->name = xmlMemStrdup(name);
+-+ enc->input = NULL;
+-+ enc->output = NULL;
+-+ enc->uconv_in = ucv_in;
+-+ enc->uconv_out = ucv_out;
+-+#ifdef DEBUG_ENCODING
+-+ xmlGenericError(xmlGenericErrorContext,
+-+ "Found ICU converter handler for encoding %s\n", name);
+-+#endif
+-+ return enc;
+-+ } else if (ucv_in != NULL || ucv_out != NULL) {
+-+ closeIcuConverter(ucv_in);
+-+ closeIcuConverter(ucv_out);
+-+ xmlEncodingErr(XML_ERR_INTERNAL_ERROR,
+-+ "ICU converter : problems with filters for '%s'\n", name);
+-+ }
+-+#endif /* LIBXML_ICU_ENABLED */
+-
+- #ifdef DEBUG_ENCODING
+- xmlGenericError(xmlGenericErrorContext,
+-@@ -1737,6 +1822,75 @@
+-
+- /************************************************************************
+- * *
+-+ * ICU based generic conversion functions *
+-+ * *
+-+ ************************************************************************/
+-+
+-+#ifdef LIBXML_ICU_ENABLED
+-+/**
+-+ * xmlUconvWrapper:
+-+ * @cd: ICU uconverter data structure
+-+ * @toUnicode : non-zero if toUnicode. 0 otherwise.
+-+ * @out: a pointer to an array of bytes to store the result
+-+ * @outlen: the length of @out
+-+ * @in: a pointer to an array of ISO Latin 1 chars
+-+ * @inlen: the length of @in
+-+ *
+-+ * Returns 0 if success, or
+-+ * -1 by lack of space, or
+-+ * -2 if the transcoding fails (for *in is not valid utf8 string or
+-+ * the result of transformation can't fit into the encoding we want), or
+-+ * -3 if there the last byte can't form a single output char.
+-+ *
+-+ * The value of @inlen after return is the number of octets consumed
+-+ * as the return value is positive, else unpredictable.
+-+ * The value of @outlen after return is the number of ocetes consumed.
+-+ */
+-+static int
+-+xmlUconvWrapper(uconv_t *cd, int toUnicode, unsigned char *out, int *outlen,
+-+ const unsigned char *in, int *inlen) {
+-+ const char *ucv_in = (const char *) in;
+-+ char *ucv_out = (char *) out;
+-+ UErrorCode err = U_ZERO_ERROR;
+-+
+-+ if ((out == NULL) || (outlen == NULL) || (inlen == NULL) || (in == NULL)) {
+-+ if (outlen != NULL) *outlen = 0;
+-+ return(-1);
+-+ }
+-+
+-+ /*
+-+ * TODO(jungshik)
+-+ * 1. is ucnv_convert(To|From)Algorithmic better?
+-+ * 2. had we better use an explicit pivot buffer?
+-+ * 3. error returned comes from 'fromUnicode' only even
+-+ * when toUnicode is true !
+-+ */
+-+ if (toUnicode) {
+-+ /* encoding => UTF-16 => UTF-8 */
+-+ ucnv_convertEx(cd->utf8, cd->uconv, &ucv_out, ucv_out + *outlen,
+-+ &ucv_in, ucv_in + *inlen, NULL, NULL, NULL, NULL,
+-+ 0, TRUE, &err);
+-+ } else {
+-+ /* UTF-8 => UTF-16 => encoding */
+-+ ucnv_convertEx(cd->uconv, cd->utf8, &ucv_out, ucv_out + *outlen,
+-+ &ucv_in, ucv_in + *inlen, NULL, NULL, NULL, NULL,
+-+ 0, TRUE, &err);
+-+ }
+-+ *inlen = ucv_in - (const char*) in;
+-+ *outlen = ucv_out - (char *) out;
+-+ if (U_SUCCESS(err))
+-+ return 0;
+-+ if (err == U_BUFFER_OVERFLOW_ERROR)
+-+ return -1;
+-+ if (err == U_INVALID_CHAR_FOUND || err == U_ILLEGAL_CHAR_FOUND)
+-+ return -2;
+-+ /* if (err == U_TRUNCATED_CHAR_FOUND) */
+-+ return -3;
+-+}
+-+#endif /* LIBXML_ICU_ENABLED */
+-+
+-+/************************************************************************
+-+ * *
+- * The real API used by libxml for on-the-fly conversion *
+- * *
+- ************************************************************************/
+-@@ -1810,6 +1964,16 @@
+- if (ret == -1) ret = -3;
+- }
+- #endif /* LIBXML_ICONV_ENABLED */
+-+#ifdef LIBXML_ICU_ENABLED
+-+ else if (handler->uconv_in != NULL) {
+-+ ret = xmlUconvWrapper(handler->uconv_in, 1, &out->content[out->use],
+-+ &written, in->content, &toconv);
+-+ xmlBufferShrink(in, toconv);
+-+ out->use += written;
+-+ out->content[out->use] = 0;
+-+ if (ret == -1) ret = -3;
+-+ }
+-+#endif /* LIBXML_ICU_ENABLED */
+- #ifdef DEBUG_ENCODING
+- switch (ret) {
+- case 0:
+-@@ -1915,6 +2079,17 @@
+- ret = -3;
+- }
+- #endif /* LIBXML_ICONV_ENABLED */
+-+#ifdef LIBXML_ICU_ENABLED
+-+ else if (handler->uconv_in != NULL) {
+-+ ret = xmlUconvWrapper(handler->uconv_in, 1, &out->content[out->use],
+-+ &written, in->content, &toconv);
+-+ xmlBufferShrink(in, toconv);
+-+ out->use += written;
+-+ out->content[out->use] = 0;
+-+ if (ret == -1)
+-+ ret = -3;
+-+ }
+-+#endif /* LIBXML_ICU_ENABLED */
+- switch (ret) {
+- case 0:
+- #ifdef DEBUG_ENCODING
+-@@ -2015,6 +2190,15 @@
+- out->content[out->use] = 0;
+- }
+- #endif /* LIBXML_ICONV_ENABLED */
+-+#ifdef LIBXML_ICU_ENABLED
+-+ else if (handler->uconv_out != NULL) {
+-+ ret = xmlUconvWrapper(handler->uconv_out, 0,
+-+ &out->content[out->use],
+-+ &written, NULL, &toconv);
+-+ out->use += written;
+-+ out->content[out->use] = 0;
+-+ }
+-+#endif /* LIBXML_ICU_ENABLED */
+- #ifdef DEBUG_ENCODING
+- xmlGenericError(xmlGenericErrorContext,
+- "initialized encoder\n");
+-@@ -2061,6 +2245,26 @@
+- }
+- }
+- #endif /* LIBXML_ICONV_ENABLED */
+-+#ifdef LIBXML_ICU_ENABLED
+-+ else if (handler->uconv_out != NULL) {
+-+ ret = xmlUconvWrapper(handler->uconv_out, 0,
+-+ &out->content[out->use],
+-+ &written, in->content, &toconv);
+-+ xmlBufferShrink(in, toconv);
+-+ out->use += written;
+-+ writtentot += written;
+-+ out->content[out->use] = 0;
+-+ if (ret == -1) {
+-+ if (written > 0) {
+-+ /*
+-+ * Can be a limitation of iconv
+-+ */
+-+ goto retry;
+-+ }
+-+ ret = -3;
+-+ }
+-+ }
+-+#endif /* LIBXML_ICU_ENABLED */
+- else {
+- xmlEncodingErr(XML_I18N_NO_OUTPUT,
+- "xmlCharEncOutFunc: no output function !\n", NULL);
+-@@ -2173,6 +2377,22 @@
+- xmlFree(handler);
+- }
+- #endif /* LIBXML_ICONV_ENABLED */
+-+#ifdef LIBXML_ICU_ENABLED
+-+ if ((handler->uconv_out != NULL) || (handler->uconv_in != NULL)) {
+-+ if (handler->name != NULL)
+-+ xmlFree(handler->name);
+-+ handler->name = NULL;
+-+ if (handler->uconv_out != NULL) {
+-+ closeIcuConverter(handler->uconv_out);
+-+ handler->uconv_out = NULL;
+-+ }
+-+ if (handler->uconv_in != NULL) {
+-+ closeIcuConverter(handler->uconv_in);
+-+ handler->uconv_in = NULL;
+-+ }
+-+ xmlFree(handler);
+-+ }
+-+#endif
+- #ifdef DEBUG_ENCODING
+- if (ret)
+- xmlGenericError(xmlGenericErrorContext,
+-@@ -2248,6 +2468,22 @@
+- cur += toconv;
+- } while (ret == -2);
+- #endif
+-+#ifdef LIBXML_ICU_ENABLED
+-+ } else if (handler->uconv_out != NULL) {
+-+ do {
+-+ toconv = in->end - cur;
+-+ written = 32000;
+-+ ret = xmlUconvWrapper(handler->uconv_out, 0, &convbuf[0],
+-+ &written, cur, &toconv);
+-+ if (ret < 0) {
+-+ if (written > 0)
+-+ ret = -2;
+-+ else
+-+ return(-1);
+-+ }
+-+ unused += written;
+-+ cur += toconv;
+-+ } while (ret == -2);
+- } else {
+- /* could not find a converter */
+- return(-1);
+-@@ -2259,8 +2495,9 @@
+- }
+- return(in->consumed + (in->cur - in->base));
+- }
+-+#endif
+-
+--#ifndef LIBXML_ICONV_ENABLED
+-+#if !defined(LIBXML_ICONV_ENABLED) && !defined(LIBXML_ICU_ENABLED)
+- #ifdef LIBXML_ISO8859X_ENABLED
+-
+- /**
+-Index: libxml/include/libxml/encoding.h
+-===================================================================
+---- libxml.orig/include/libxml/encoding.h 2010-07-09 14:50:27.503114118 -0700
+-+++ libxml/include/libxml/encoding.h 2010-07-09 14:53:27.251611643 -0700
+-@@ -26,6 +26,24 @@
+-
+- #ifdef LIBXML_ICONV_ENABLED
+- #include <iconv.h>
+-+#else
+-+#ifdef LIBXML_ICU_ENABLED
+-+#include <unicode/ucnv.h>
+-+#if 0
+-+/* Forward-declare UConverter here rather than pulling in <unicode/ucnv.h>
+-+ * to prevent unwanted ICU symbols being exposed to users of libxml2.
+-+ * One particular case is Qt4 conflicting on UChar32.
+-+ */
+-+#include <stdint.h>
+-+struct UConverter;
+-+typedef struct UConverter UConverter;
+-+#ifdef _MSC_VER
+-+typedef wchar_t UChar;
+-+#else
+-+typedef uint16_t UChar;
+-+#endif
+-+#endif
+-+#endif
+- #endif
+- #ifdef __cplusplus
+- extern "C" {
+-@@ -125,6 +143,13 @@
+- * Block defining the handlers for non UTF-8 encodings.
+- * If iconv is supported, there are two extra fields.
+- */
+-+#ifdef LIBXML_ICU_ENABLED
+-+struct _uconv_t {
+-+ UConverter *uconv; /* for conversion between an encoding and UTF-16 */
+-+ UConverter *utf8; /* for conversion between UTF-8 and UTF-16 */
+-+};
+-+typedef struct _uconv_t uconv_t;
+-+#endif
+-
+- typedef struct _xmlCharEncodingHandler xmlCharEncodingHandler;
+- typedef xmlCharEncodingHandler *xmlCharEncodingHandlerPtr;
+-@@ -136,6 +161,10 @@
+- iconv_t iconv_in;
+- iconv_t iconv_out;
+- #endif /* LIBXML_ICONV_ENABLED */
+-+#ifdef LIBXML_ICU_ENABLED
+-+ uconv_t *uconv_in;
+-+ uconv_t *uconv_out;
+-+#endif /* LIBXML_ICU_ENABLED */
+- };
+-
+- #ifdef __cplusplus
+-Index: libxml/include/libxml/parser.h
+-===================================================================
+---- libxml.orig/include/libxml/parser.h 2010-07-09 14:51:21.190673740 -0700
+-+++ libxml/include/libxml/parser.h 2010-07-09 14:53:19.571862214 -0700
+-@@ -1222,6 +1222,7 @@
+- XML_WITH_DEBUG_MEM = 29,
+- XML_WITH_DEBUG_RUN = 30,
+- XML_WITH_ZLIB = 31,
+-+ XML_WITH_ICU = 32,
+- XML_WITH_NONE = 99999 /* just to be sure of allocation size */
+- } xmlFeature;
+-
+-Index: libxml/parser.c
+-===================================================================
+---- libxml.orig/parser.c 2010-07-09 14:52:15.150057108 -0700
+-+++ libxml/parser.c 2010-07-09 14:53:06.190137405 -0700
+-@@ -954,6 +954,12 @@
+- #else
+- return(0);
+- #endif
+-+ case XML_WITH_ICU:
+-+#ifdef LIBXML_ICU_ENABLED
+-+ return(1);
+-+#else
+-+ return(0);
+-+#endif
+- default:
+- break;
+- }
« no previous file with comments | « third_party/libxml/patches/LoadLibraryA ('k') | third_party/libxml/patches/icu-configure » ('j') | no next file with comments »

Powered by Google App Engine
This is Rietveld 408576698