4 * Deals with the XML information in WIM files. Uses the C library libxml2.
8 * Copyright (C) 2012, 2013 Eric Biggers
10 * This file is free software; you can redistribute it and/or modify it under
11 * the terms of the GNU Lesser General Public License as published by the Free
12 * Software Foundation; either version 3 of the License, or (at your option) any
15 * This file is distributed in the hope that it will be useful, but WITHOUT
16 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
17 * FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more
20 * You should have received a copy of the GNU Lesser General Public License
21 * along with this file; if not, see http://www.gnu.org/licenses/.
28 #include "wimlib/assert.h"
29 #include "wimlib/dentry.h"
30 #include "wimlib/encoding.h"
31 #include "wimlib/error.h"
32 #include "wimlib/file_io.h"
33 #include "wimlib/lookup_table.h"
34 #include "wimlib/metadata.h"
35 #include "wimlib/resource.h"
36 #include "wimlib/timestamp.h"
37 #include "wimlib/xml.h"
38 #include "wimlib/write.h"
40 #include <libxml/encoding.h>
41 #include <libxml/parser.h>
42 #include <libxml/tree.h>
43 #include <libxml/xmlwriter.h>
48 /* Structures used to form an in-memory representation of the XML data (other
49 * than the raw parse tree from libxml). */
51 struct windows_version {
63 tchar *installation_type;
64 tchar *pkeyconfigversion;
69 tchar *default_language;
72 bool windows_version_exists;
73 struct windows_version windows_version;
78 bool windows_info_exists;
84 u64 last_modification_time;
85 struct windows_info windows_info;
89 tchar *display_description;
93 /* Note: must update clone_image_info() if adding new fields here */
95 struct wim_lookup_table *lookup_table; /* temporary field */
98 /* A struct wim_info structure corresponds to the entire XML data for a WIM file. */
102 /* Array of `struct image_info's, one for each image in the WIM that is
103 * mentioned in the XML data. */
104 struct image_info *images;
107 struct xml_string_spec {
112 #define ELEM(STRING_NAME, MEMBER_NAME) \
113 {STRING_NAME, offsetof(struct image_info, MEMBER_NAME)}
114 static const struct xml_string_spec
115 image_info_xml_string_specs[] = {
117 ELEM("DESCRIPTION", description),
118 ELEM("DISPLAYNAME", display_name),
119 ELEM("DISPLAYDESCRIPTION", display_description),
120 ELEM("FLAGS", flags),
124 #define ELEM(STRING_NAME, MEMBER_NAME) \
125 {STRING_NAME, offsetof(struct windows_info, MEMBER_NAME)}
126 static const struct xml_string_spec
127 windows_info_xml_string_specs[] = {
128 ELEM("PRODUCTNAME", product_name),
129 ELEM("EDITIONID", edition_id),
130 ELEM("INSTALLATIONTYPE", installation_type),
132 ELEM("PRODUCTTYPE", product_type),
133 ELEM("PRODUCTSUITE", product_suite),
138 wim_info_get_total_bytes(const struct wim_info *info)
141 return info->total_bytes;
147 wim_info_get_image_hard_link_bytes(const struct wim_info *info, int image)
150 return info->images[image - 1].hard_link_bytes;
156 wim_info_get_image_total_bytes(const struct wim_info *info, int image)
159 return info->images[image - 1].total_bytes;
165 wim_info_get_num_images(const struct wim_info *info)
168 return info->num_images;
174 wim_info_set_wimboot(struct wim_info *info, int image, bool value)
176 info->images[image - 1].wimboot = value;
180 wim_info_get_wimboot(const struct wim_info *info, int image)
182 return info->images[image - 1].wimboot;
185 /* Architecture constants are from w64 mingw winnt.h */
186 #define PROCESSOR_ARCHITECTURE_INTEL 0
187 #define PROCESSOR_ARCHITECTURE_MIPS 1
188 #define PROCESSOR_ARCHITECTURE_ALPHA 2
189 #define PROCESSOR_ARCHITECTURE_PPC 3
190 #define PROCESSOR_ARCHITECTURE_SHX 4
191 #define PROCESSOR_ARCHITECTURE_ARM 5
192 #define PROCESSOR_ARCHITECTURE_IA64 6
193 #define PROCESSOR_ARCHITECTURE_ALPHA64 7
194 #define PROCESSOR_ARCHITECTURE_MSIL 8
195 #define PROCESSOR_ARCHITECTURE_AMD64 9
196 #define PROCESSOR_ARCHITECTURE_IA32_ON_WIN64 10
198 /* Returns a statically allocated string that is a string representation of the
199 * architecture number. */
204 case PROCESSOR_ARCHITECTURE_INTEL:
206 case PROCESSOR_ARCHITECTURE_MIPS:
208 case PROCESSOR_ARCHITECTURE_ARM:
210 case PROCESSOR_ARCHITECTURE_IA64:
212 case PROCESSOR_ARCHITECTURE_AMD64:
220 /* Iterate through the children of an xmlNode. */
221 #define for_node_child(parent, child) \
222 for (child = parent->children; child != NULL; child = child->next)
224 /* Utility functions for xmlNodes */
226 node_is_element(xmlNode *node)
228 return node->type == XML_ELEMENT_NODE;
232 node_is_text(xmlNode *node)
234 return node->type == XML_TEXT_NODE;
238 node_name_is(xmlNode *node, const char *name)
240 /* For now, both upper case and lower case element names are accepted. */
241 return strcasecmp((const char *)node->name, name) == 0;
245 node_get_number(const xmlNode *u64_node, int base)
248 for_node_child(u64_node, child)
249 if (node_is_text(child))
250 return strtoull(child->content, NULL, base);
254 /* Finds the text node that is a child of an element node and returns its
255 * content converted to a 64-bit unsigned integer. Returns 0 if no text node is
258 node_get_u64(const xmlNode *u64_node)
260 return node_get_number(u64_node, 10);
263 /* Like node_get_u64(), but expects a number in base 16. */
265 node_get_hex_u64(const xmlNode *u64_node)
267 return node_get_number(u64_node, 16);
271 node_get_string(const xmlNode *string_node, tchar **tstr_ret)
278 for_node_child(string_node, child)
279 if (node_is_text(child) && child->content)
280 return utf8_to_tstr_simple(child->content, tstr_ret);
284 /* Returns the timestamp from a time node. It has child elements <HIGHPART> and
285 * <LOWPART> that are then used to construct a 64-bit timestamp. */
287 node_get_timestamp(const xmlNode *time_node)
292 for_node_child(time_node, child) {
293 if (!node_is_element(child))
295 if (node_name_is(child, "HIGHPART"))
296 high_part = node_get_hex_u64(child);
297 else if (node_name_is(child, "LOWPART"))
298 low_part = node_get_hex_u64(child);
300 return (u64)low_part | ((u64)high_part << 32);
303 /* Used to sort an array of struct image_infos by their image indices. */
305 sort_by_index(const void *p1, const void *p2)
307 int index_1 = ((const struct image_info*)p1)->index;
308 int index_2 = ((const struct image_info*)p2)->index;
309 if (index_1 < index_2)
311 else if (index_1 > index_2)
318 /* Frees memory allocated inside a struct windows_info structure. */
320 destroy_windows_info(struct windows_info *windows_info)
322 FREE(windows_info->product_name);
323 FREE(windows_info->edition_id);
324 FREE(windows_info->installation_type);
325 FREE(windows_info->hal);
326 FREE(windows_info->product_type);
327 FREE(windows_info->product_suite);
328 FREE(windows_info->pkeyconfigversion);
329 for (size_t i = 0; i < windows_info->num_languages; i++)
330 FREE(windows_info->languages[i]);
331 FREE(windows_info->languages);
332 FREE(windows_info->default_language);
333 FREE(windows_info->system_root);
336 /* Frees memory allocated inside a struct image_info structure. */
338 destroy_image_info(struct image_info *image_info)
340 FREE(image_info->name);
341 FREE(image_info->description);
342 FREE(image_info->flags);
343 FREE(image_info->display_name);
344 FREE(image_info->display_description);
345 destroy_windows_info(&image_info->windows_info);
346 memset(image_info, 0, sizeof(struct image_info));
350 free_wim_info(struct wim_info *info)
354 for (int i = 0; i < info->num_images; i++)
355 destroy_image_info(&info->images[i]);
362 /* Reads the information from a <VERSION> element inside the <WINDOWS> element.
365 xml_read_windows_version(const xmlNode *version_node,
366 struct windows_version* windows_version)
369 for_node_child(version_node, child) {
370 if (!node_is_element(child))
372 if (node_name_is(child, "MAJOR"))
373 windows_version->major = node_get_u64(child);
374 else if (node_name_is(child, "MINOR"))
375 windows_version->minor = node_get_u64(child);
376 else if (node_name_is(child, "BUILD"))
377 windows_version->build = node_get_u64(child);
378 else if (node_name_is(child, "SPBUILD"))
379 windows_version->sp_build = node_get_u64(child);
380 else if (node_name_is(child, "SPLEVEL"))
381 windows_version->sp_level = node_get_u64(child);
385 /* Reads the information from a <LANGUAGE> element inside a <WINDOWS> element.
388 xml_read_languages(const xmlNode *languages_node,
389 tchar ***languages_ret,
390 size_t *num_languages_ret,
391 tchar **default_language_ret)
394 size_t num_languages = 0;
398 for_node_child(languages_node, child)
399 if (node_is_element(child) && node_name_is(child, "LANGUAGE"))
402 languages = CALLOC(num_languages, sizeof(languages[0]));
404 return WIMLIB_ERR_NOMEM;
406 *languages_ret = languages;
407 *num_languages_ret = num_languages;
410 for_node_child(languages_node, child) {
411 if (!node_is_element(child))
413 if (node_name_is(child, "LANGUAGE"))
414 ret = node_get_string(child, languages++);
415 else if (node_name_is(child, "DEFAULT"))
416 ret = node_get_string(child, default_language_ret);
423 /* Reads the information from a <WINDOWS> element inside an <IMAGE> element. */
425 xml_read_windows_info(const xmlNode *windows_node,
426 struct windows_info *windows_info)
431 for_node_child(windows_node, child) {
432 if (!node_is_element(child))
434 if (node_name_is(child, "ARCH")) {
435 windows_info->arch = node_get_u64(child);
436 } else if (node_name_is(child, "PRODUCTNAME")) {
437 ret = node_get_string(child,
438 &windows_info->product_name);
439 } else if (node_name_is(child, "EDITIONID")) {
440 ret = node_get_string(child,
441 &windows_info->edition_id);
442 } else if (node_name_is(child, "INSTALLATIONTYPE")) {
443 ret = node_get_string(child,
444 &windows_info->installation_type);
445 } else if (node_name_is(child, "PRODUCTTYPE")) {
446 ret = node_get_string(child,
447 &windows_info->product_type);
448 } else if (node_name_is(child, "PRODUCTSUITE")) {
449 ret = node_get_string(child,
450 &windows_info->product_suite);
451 } else if (node_name_is(child, "LANGUAGES")) {
452 ret = xml_read_languages(child,
453 &windows_info->languages,
454 &windows_info->num_languages,
455 &windows_info->default_language);
456 } else if (node_name_is(child, "VERSION")) {
457 xml_read_windows_version(child,
458 &windows_info->windows_version);
459 windows_info->windows_version_exists = true;
460 } else if (node_name_is(child, "SYSTEMROOT")) {
461 ret = node_get_string(child, &windows_info->system_root);
462 } else if (node_name_is(child, "HAL")) {
463 ret = node_get_string(child, &windows_info->hal);
464 } else if (node_name_is(child, "SERVICINGDATA")) {
467 for_node_child(child, grandchild) {
468 if (node_is_element(grandchild) &&
469 node_name_is(grandchild, "PKEYCONFIGVERSION"))
471 ret = node_get_string(grandchild,
472 &windows_info->pkeyconfigversion);
483 /* Reads the information from an <IMAGE> element. */
485 xml_read_image_info(xmlNode *image_node, struct image_info *image_info)
491 index_prop = xmlGetProp(image_node, "INDEX");
493 image_info->index = atoi(index_prop);
496 image_info->index = 1;
500 for_node_child(image_node, child) {
501 if (!node_is_element(child))
503 if (node_name_is(child, "DIRCOUNT"))
504 image_info->dir_count = node_get_u64(child);
505 else if (node_name_is(child, "FILECOUNT"))
506 image_info->file_count = node_get_u64(child);
507 else if (node_name_is(child, "TOTALBYTES"))
508 image_info->total_bytes = node_get_u64(child);
509 else if (node_name_is(child, "HARDLINKBYTES"))
510 image_info->hard_link_bytes = node_get_u64(child);
511 else if (node_name_is(child, "CREATIONTIME"))
512 image_info->creation_time = node_get_timestamp(child);
513 else if (node_name_is(child, "LASTMODIFICATIONTIME"))
514 image_info->last_modification_time = node_get_timestamp(child);
515 else if (node_name_is(child, "WINDOWS")) {
516 ret = xml_read_windows_info(child,
517 &image_info->windows_info);
518 image_info->windows_info_exists = true;
519 } else if (node_name_is(child, "NAME")) {
520 ret = node_get_string(child, &image_info->name);
521 } else if (node_name_is(child, "DESCRIPTION")) {
522 ret = node_get_string(child, &image_info->description);
523 } else if (node_name_is(child, "FLAGS")) {
524 ret = node_get_string(child, &image_info->flags);
525 } else if (node_name_is(child, "DISPLAYNAME")) {
526 ret = node_get_string(child, &image_info->display_name);
527 } else if (node_name_is(child, "DISPLAYDESCRIPTION")) {
528 ret = node_get_string(child, &image_info->display_description);
529 } else if (node_name_is(child, "WIMBOOT")) {
530 if (node_get_u64(child) == 1) {
531 image_info->wimboot = true;
537 if (!image_info->name) {
539 empty_name = MALLOC(sizeof(tchar));
541 return WIMLIB_ERR_NOMEM;
542 *empty_name = T('\0');
543 image_info->name = empty_name;
548 /* Reads the information from a <WIM> element, which should be the root element
549 * of the XML tree. */
551 xml_read_wim_info(const xmlNode *wim_node, struct wim_info **wim_info_ret)
553 struct wim_info *wim_info;
559 wim_info = CALLOC(1, sizeof(struct wim_info));
561 return WIMLIB_ERR_NOMEM;
563 /* Count how many images there are. */
565 for_node_child(wim_node, child) {
566 if (node_is_element(child) && node_name_is(child, "IMAGE")) {
567 if (unlikely(num_images == MAX_IMAGES)) {
568 ret = WIMLIB_ERR_IMAGE_COUNT;
575 if (num_images > 0) {
576 /* Allocate the array of struct image_infos and fill them in. */
577 wim_info->images = CALLOC(num_images, sizeof(wim_info->images[0]));
578 if (!wim_info->images) {
579 ret = WIMLIB_ERR_NOMEM;
582 wim_info->num_images = num_images;
584 for_node_child(wim_node, child) {
585 if (!node_is_element(child))
587 if (node_name_is(child, "IMAGE")) {
588 ret = xml_read_image_info(child,
589 &wim_info->images[i]);
593 } else if (node_name_is(child, "TOTALBYTES")) {
594 wim_info->total_bytes = node_get_u64(child);
595 } else if (node_name_is(child, "ESD")) {
597 for_node_child(child, esdchild) {
598 if (node_is_element(esdchild) &&
599 node_name_is(esdchild, "ENCRYPTED"))
601 ret = WIMLIB_ERR_WIM_IS_ENCRYPTED;
608 /* Sort the array of image info by image index. */
609 qsort(wim_info->images, num_images,
610 sizeof(struct image_info), sort_by_index);
612 /* Make sure the image indices make sense */
613 for (i = 0; i < num_images; i++) {
614 if (wim_info->images[i].index != i + 1) {
615 ERROR("WIM images are not indexed [1...%d] "
616 "in XML data as expected",
618 ret = WIMLIB_ERR_IMAGE_COUNT;
624 *wim_info_ret = wim_info;
627 free_wim_info(wim_info);
631 /* Prints the information contained in a `struct windows_info'. */
633 print_windows_info(const struct windows_info *windows_info)
635 const struct windows_version *windows_version;
637 tprintf(T("Architecture: %"TS"\n"),
638 get_arch(windows_info->arch));
640 if (windows_info->product_name) {
641 tprintf(T("Product Name: %"TS"\n"),
642 windows_info->product_name);
645 if (windows_info->edition_id) {
646 tprintf(T("Edition ID: %"TS"\n"),
647 windows_info->edition_id);
650 if (windows_info->installation_type) {
651 tprintf(T("Installation Type: %"TS"\n"),
652 windows_info->installation_type);
655 if (windows_info->hal) {
656 tprintf(T("HAL: %"TS"\n"),
660 if (windows_info->product_type) {
661 tprintf(T("Product Type: %"TS"\n"),
662 windows_info->product_type);
665 if (windows_info->product_suite) {
666 tprintf(T("Product Suite: %"TS"\n"),
667 windows_info->product_suite);
670 tprintf(T("Languages: "));
671 for (size_t i = 0; i < windows_info->num_languages; i++) {
673 tfputs(windows_info->languages[i], stdout);
677 if (windows_info->default_language) {
678 tprintf(T("Default Language: %"TS"\n"),
679 windows_info->default_language);
681 if (windows_info->system_root) {
682 tprintf(T("System Root: %"TS"\n"),
683 windows_info->system_root);
686 if (windows_info->windows_version_exists) {
687 windows_version = &windows_info->windows_version;
688 tprintf(T("Major Version: %"PRIu64"\n"),
689 windows_version->major);
690 tprintf(T("Minor Version: %"PRIu64"\n"),
691 windows_version->minor);
692 tprintf(T("Build: %"PRIu64"\n"),
693 windows_version->build);
694 tprintf(T("Service Pack Build: %"PRIu64"\n"),
695 windows_version->sp_build);
696 tprintf(T("Service Pack Level: %"PRIu64"\n"),
697 windows_version->sp_level);
702 xml_write_string(xmlTextWriter *writer, const char *name,
707 int rc = tstr_to_utf8_simple(tstr, &utf8_str);
710 rc = xmlTextWriterWriteElement(writer, name, utf8_str);
719 xml_write_strings_from_specs(xmlTextWriter *writer,
720 const void *struct_with_strings,
721 const struct xml_string_spec specs[],
724 for (size_t i = 0; i < num_specs; i++) {
725 int rc = xml_write_string(writer, specs[i].name,
726 *(const tchar * const *)
727 (struct_with_strings + specs[i].offset));
735 dup_strings_from_specs(const void *old_struct_with_strings,
736 void *new_struct_with_strings,
737 const struct xml_string_spec specs[],
740 for (size_t i = 0; i < num_specs; i++) {
741 const tchar *old_str = *(const tchar * const *)
742 ((const void*)old_struct_with_strings + specs[i].offset);
743 tchar **new_str_p = (tchar **)((void*)new_struct_with_strings + specs[i].offset);
745 *new_str_p = TSTRDUP(old_str);
747 return WIMLIB_ERR_NOMEM;
753 /* Writes the information contained in a `struct windows_version' to the XML
754 * document being written. This is the <VERSION> element inside the <WINDOWS>
757 xml_write_windows_version(xmlTextWriter *writer,
758 const struct windows_version *version)
761 rc = xmlTextWriterStartElement(writer, "VERSION");
765 rc = xmlTextWriterWriteFormatElement(writer, "MAJOR", "%"PRIu64,
770 rc = xmlTextWriterWriteFormatElement(writer, "MINOR", "%"PRIu64,
775 rc = xmlTextWriterWriteFormatElement(writer, "BUILD", "%"PRIu64,
780 rc = xmlTextWriterWriteFormatElement(writer, "SPBUILD", "%"PRIu64,
785 rc = xmlTextWriterWriteFormatElement(writer, "SPLEVEL", "%"PRIu64,
790 rc = xmlTextWriterEndElement(writer); /* </VERSION> */
797 /* Writes the information contained in a `struct windows_info' to the XML
798 * document being written. This is the <WINDOWS> element. */
800 xml_write_windows_info(xmlTextWriter *writer,
801 const struct windows_info *windows_info)
804 rc = xmlTextWriterStartElement(writer, "WINDOWS");
808 rc = xmlTextWriterWriteFormatElement(writer, "ARCH", "%"PRIu64,
813 rc = xml_write_strings_from_specs(writer,
815 windows_info_xml_string_specs,
816 ARRAY_LEN(windows_info_xml_string_specs));
820 if (windows_info->num_languages) {
821 rc = xmlTextWriterStartElement(writer, "LANGUAGES");
825 for (size_t i = 0; i < windows_info->num_languages; i++) {
826 rc = xml_write_string(writer, "LANGUAGE",
827 windows_info->languages[i]);
832 rc = xml_write_string(writer, "DEFAULT",
833 windows_info->default_language);
837 rc = xmlTextWriterEndElement(writer); /* </LANGUAGES> */
842 if (windows_info->pkeyconfigversion) {
843 rc = xmlTextWriterStartElement(writer, "SERVICINGDATA");
847 rc = xml_write_string(writer, "PKEYCONFIGVERSION",
848 windows_info->pkeyconfigversion);
852 rc = xmlTextWriterEndElement(writer);
857 if (windows_info->windows_version_exists) {
858 rc = xml_write_windows_version(writer, &windows_info->windows_version);
863 rc = xml_write_string(writer, "SYSTEMROOT", windows_info->system_root);
867 rc = xmlTextWriterEndElement(writer); /* </WINDOWS> */
874 /* Writes a time element to the XML document being constructed in memory. */
876 xml_write_time(xmlTextWriter *writer, const char *element_name, u64 time)
879 rc = xmlTextWriterStartElement(writer, element_name);
883 rc = xmlTextWriterWriteFormatElement(writer, "HIGHPART",
884 "0x%08"PRIX32, (u32)(time >> 32));
888 rc = xmlTextWriterWriteFormatElement(writer, "LOWPART",
889 "0x%08"PRIX32, (u32)time);
893 rc = xmlTextWriterEndElement(writer); /* </@element_name> */
899 /* Writes an <IMAGE> element to the XML document. */
901 xml_write_image_info(xmlTextWriter *writer, const struct image_info *image_info)
904 rc = xmlTextWriterStartElement(writer, "IMAGE");
908 rc = xmlTextWriterWriteFormatAttribute(writer, "INDEX", "%d",
913 rc = xmlTextWriterWriteFormatElement(writer, "DIRCOUNT", "%"PRIu64,
914 image_info->dir_count);
918 rc = xmlTextWriterWriteFormatElement(writer, "FILECOUNT", "%"PRIu64,
919 image_info->file_count);
923 rc = xmlTextWriterWriteFormatElement(writer, "TOTALBYTES", "%"PRIu64,
924 image_info->total_bytes);
928 rc = xmlTextWriterWriteFormatElement(writer, "HARDLINKBYTES", "%"PRIu64,
929 image_info->hard_link_bytes);
933 rc = xml_write_time(writer, "CREATIONTIME", image_info->creation_time);
937 rc = xml_write_time(writer, "LASTMODIFICATIONTIME",
938 image_info->last_modification_time);
942 if (image_info->windows_info_exists) {
943 rc = xml_write_windows_info(writer, &image_info->windows_info);
948 rc = xml_write_strings_from_specs(writer, image_info,
949 image_info_xml_string_specs,
950 ARRAY_LEN(image_info_xml_string_specs));
954 if (image_info->wimboot) {
955 rc = xmlTextWriterWriteFormatElement(writer, "WIMBOOT", "%d", 1);
960 rc = xmlTextWriterEndElement(writer); /* </IMAGE> */
969 /* Makes space for another image in the XML information and return a pointer to
971 static struct image_info *
972 add_image_info_struct(struct wim_info *wim_info)
974 struct image_info *images;
976 images = CALLOC(wim_info->num_images + 1, sizeof(struct image_info));
979 memcpy(images, wim_info->images,
980 wim_info->num_images * sizeof(struct image_info));
981 FREE(wim_info->images);
982 wim_info->images = images;
983 wim_info->num_images++;
984 return &images[wim_info->num_images - 1];
988 clone_windows_info(const struct windows_info *old, struct windows_info *new)
992 new->arch = old->arch;
994 ret = dup_strings_from_specs(old, new, windows_info_xml_string_specs,
995 ARRAY_LEN(windows_info_xml_string_specs));
999 if (old->pkeyconfigversion) {
1000 new->pkeyconfigversion = TSTRDUP(old->pkeyconfigversion);
1001 if (new->pkeyconfigversion == NULL)
1002 return WIMLIB_ERR_NOMEM;
1005 if (old->languages) {
1006 new->languages = CALLOC(old->num_languages, sizeof(new->languages[0]));
1007 if (!new->languages)
1008 return WIMLIB_ERR_NOMEM;
1009 new->num_languages = old->num_languages;
1010 for (size_t i = 0; i < new->num_languages; i++) {
1011 if (!old->languages[i])
1013 new->languages[i] = TSTRDUP(old->languages[i]);
1014 if (!new->languages[i])
1015 return WIMLIB_ERR_NOMEM;
1018 if (old->default_language &&
1019 !(new->default_language = TSTRDUP(old->default_language)))
1020 return WIMLIB_ERR_NOMEM;
1021 if (old->system_root && !(new->system_root = TSTRDUP(old->system_root)))
1022 return WIMLIB_ERR_NOMEM;
1023 if (old->windows_version_exists) {
1024 new->windows_version_exists = true;
1025 memcpy(&new->windows_version, &old->windows_version,
1026 sizeof(old->windows_version));
1032 clone_image_info(const struct image_info *old, struct image_info *new)
1036 new->dir_count = old->dir_count;
1037 new->file_count = old->file_count;
1038 new->total_bytes = old->total_bytes;
1039 new->hard_link_bytes = old->hard_link_bytes;
1040 new->creation_time = old->creation_time;
1041 new->last_modification_time = old->last_modification_time;
1043 ret = dup_strings_from_specs(old, new,
1044 image_info_xml_string_specs,
1045 ARRAY_LEN(image_info_xml_string_specs));
1049 if (old->windows_info_exists) {
1050 new->windows_info_exists = true;
1051 ret = clone_windows_info(&old->windows_info,
1052 &new->windows_info);
1056 new->wimboot = old->wimboot;
1060 /* Copies the XML information for an image between WIM files.
1062 * @dest_image_name and @dest_image_description are ignored if they are NULL;
1063 * otherwise, they are used to override the image name and/or image description
1064 * from the XML data in the source WIM file.
1066 * On failure, WIMLIB_ERR_NOMEM is returned and no changes are made. Otherwise,
1067 * 0 is returned and the WIM information at *new_wim_info_p is modified.
1070 xml_export_image(const struct wim_info *old_wim_info,
1072 struct wim_info **new_wim_info_p,
1073 const tchar *dest_image_name,
1074 const tchar *dest_image_description)
1076 struct wim_info *new_wim_info;
1077 struct image_info *image_info;
1080 DEBUG("Copying XML data between WIM files for source image %d.", image);
1082 wimlib_assert(old_wim_info != NULL);
1083 wimlib_assert(image >= 1 && image <= old_wim_info->num_images);
1085 if (*new_wim_info_p) {
1086 new_wim_info = *new_wim_info_p;
1088 new_wim_info = CALLOC(1, sizeof(struct wim_info));
1093 image_info = add_image_info_struct(new_wim_info);
1097 ret = clone_image_info(&old_wim_info->images[image - 1], image_info);
1099 goto err_destroy_image_info;
1101 image_info->index = new_wim_info->num_images;
1103 if (dest_image_name) {
1104 FREE(image_info->name);
1105 image_info->name = TSTRDUP(dest_image_name);
1106 if (!image_info->name)
1107 goto err_destroy_image_info;
1109 if (dest_image_description) {
1110 FREE(image_info->description);
1111 image_info->description = TSTRDUP(dest_image_description);
1112 if (!image_info->description)
1113 goto err_destroy_image_info;
1115 *new_wim_info_p = new_wim_info;
1117 err_destroy_image_info:
1118 destroy_image_info(image_info);
1120 if (new_wim_info != *new_wim_info_p)
1121 free_wim_info(new_wim_info);
1122 return WIMLIB_ERR_NOMEM;
1125 /* Removes an image from the XML information. */
1127 xml_delete_image(struct wim_info **wim_info_p, int image)
1129 struct wim_info *wim_info;
1131 wim_info = *wim_info_p;
1132 wimlib_assert(image >= 1 && image <= wim_info->num_images);
1133 DEBUG("Deleting image %d from the XML data.", image);
1135 destroy_image_info(&wim_info->images[image - 1]);
1137 memmove(&wim_info->images[image - 1],
1138 &wim_info->images[image],
1139 (wim_info->num_images - image) * sizeof(struct image_info));
1141 if (--wim_info->num_images == 0) {
1142 free_wim_info(wim_info);
1145 for (int i = image - 1; i < wim_info->num_images; i++)
1146 wim_info->images[i].index--;
1151 xml_get_max_image_name_len(const WIMStruct *wim)
1154 for (u32 i = 0; i < wim->hdr.image_count; i++)
1155 max_len = max(max_len, tstrlen(wim->wim_info->images[i].name));
1160 xml_set_memory_allocator(void *(*malloc_func)(size_t),
1161 void (*free_func)(void *),
1162 void *(*realloc_func)(void *, size_t))
1164 xmlMemSetup(free_func, malloc_func, realloc_func, STRDUP);
1168 calculate_dentry_statistics(struct wim_dentry *dentry, void *arg)
1170 struct image_info *info = arg;
1171 const struct wim_inode *inode = dentry->d_inode;
1173 /* Update directory count and file count.
1175 * Each dentry counts as either a file or a directory, but not both.
1176 * The root directory is an exception: it is not counted at all.
1178 * Symbolic links and junction points (and presumably other reparse
1179 * points) count as regular files. This is despite the fact that
1180 * junction points have FILE_ATTRIBUTE_DIRECTORY set.
1183 if (!dentry_is_root(dentry)) {
1184 if (inode_is_directory(inode))
1191 * Update total bytes and hard link bytes.
1193 * We try to act the same as the MS implementation, even though there
1194 * are some inconsistencies/bugs in the way it operates.
1196 * If there are no alternate data streams in the image, the "total
1197 * bytes" is the sum of the size of the un-named data stream of each
1198 * inode times the link count of that inode. In other words, it would
1199 * be the total number of bytes of regular files you would have if you
1200 * extracted the full image without any hard-links. The "hard link
1201 * bytes" is equal to the "total bytes" minus the size of the un-named
1202 * data stream of each inode. In other words, the "hard link bytes"
1203 * counts the size of the un-named data stream for all the links to each
1204 * inode except the first one.
1206 * Reparse points and directories don't seem to be counted in either the
1207 * total bytes or the hard link bytes.
1209 * And now we get to the most confusing part, the alternate data
1210 * streams. They are not counted in the "total bytes". However, if the
1211 * link count of an inode with alternate data streams is 2 or greater,
1212 * the size of all the alternate data streams is included in the "hard
1213 * link bytes", and this size is multiplied by the link count (NOT one
1214 * less than the link count).
1216 if (!(inode->i_attributes & (FILE_ATTRIBUTE_DIRECTORY |
1217 FILE_ATTRIBUTE_REPARSE_POINT)))
1219 struct wim_lookup_table_entry *lte;
1221 lte = inode_unnamed_lte(inode, info->lookup_table);
1223 info->total_bytes += lte->size;
1224 if (!dentry_is_first_in_inode(dentry))
1225 info->hard_link_bytes += lte->size;
1228 if (inode->i_nlink >= 2 && dentry_is_first_in_inode(dentry)) {
1229 for (unsigned i = 0; i < inode->i_num_ads; i++) {
1230 if (inode->i_ads_entries[i].stream_name_nbytes) {
1231 lte = inode_stream_lte(inode, i + 1, info->lookup_table);
1233 info->hard_link_bytes += inode->i_nlink *
1244 * Calculate what to put in the <FILECOUNT>, <DIRCOUNT>, <TOTALBYTES>, and
1245 * <HARDLINKBYTES> elements of the specified WIM image.
1248 xml_update_image_info(WIMStruct *wim, int image)
1250 struct image_info *image_info;
1252 DEBUG("Updating the image info for image %d", image);
1254 image_info = &wim->wim_info->images[image - 1];
1256 image_info->file_count = 0;
1257 image_info->dir_count = 0;
1258 image_info->total_bytes = 0;
1259 image_info->hard_link_bytes = 0;
1260 image_info->lookup_table = wim->lookup_table;
1262 for_dentry_in_tree(wim->image_metadata[image - 1]->root_dentry,
1263 calculate_dentry_statistics,
1265 image_info->last_modification_time = now_as_wim_timestamp();
1268 /* Adds an image to the XML information. */
1270 xml_add_image(WIMStruct *wim, const tchar *name)
1272 struct wim_info *wim_info;
1273 struct image_info *image_info;
1275 wimlib_assert(name != NULL);
1277 /* If this is the first image, allocate the struct wim_info. Otherwise
1278 * use the existing struct wim_info. */
1279 if (wim->wim_info) {
1280 wim_info = wim->wim_info;
1282 wim_info = CALLOC(1, sizeof(struct wim_info));
1284 return WIMLIB_ERR_NOMEM;
1287 image_info = add_image_info_struct(wim_info);
1289 goto out_free_wim_info;
1291 if (!(image_info->name = TSTRDUP(name)))
1292 goto out_destroy_image_info;
1294 wim->wim_info = wim_info;
1295 image_info->index = wim_info->num_images;
1296 image_info->creation_time = now_as_wim_timestamp();
1297 xml_update_image_info(wim, image_info->index);
1300 out_destroy_image_info:
1301 destroy_image_info(image_info);
1302 wim_info->num_images--;
1304 if (wim_info != wim->wim_info)
1306 return WIMLIB_ERR_NOMEM;
1309 /* Prints information about the specified image from struct wim_info structure.
1312 print_image_info(const struct wim_info *wim_info, int image)
1314 const struct image_info *image_info;
1318 wimlib_assert(image >= 1 && image <= wim_info->num_images);
1320 image_info = &wim_info->images[image - 1];
1322 tprintf(T("Index: %d\n"), image_info->index);
1323 tprintf(T("Name: %"TS"\n"), image_info->name);
1325 /* Always print the Description: part even if there is no
1327 if (image_info->description)
1328 desc = image_info->description;
1331 tprintf(T("Description: %"TS"\n"), desc);
1333 if (image_info->display_name) {
1334 tprintf(T("Display Name: %"TS"\n"),
1335 image_info->display_name);
1338 if (image_info->display_description) {
1339 tprintf(T("Display Description: %"TS"\n"),
1340 image_info->display_description);
1343 tprintf(T("Directory Count: %"PRIu64"\n"), image_info->dir_count);
1344 tprintf(T("File Count: %"PRIu64"\n"), image_info->file_count);
1345 tprintf(T("Total Bytes: %"PRIu64"\n"), image_info->total_bytes);
1346 tprintf(T("Hard Link Bytes: %"PRIu64"\n"), image_info->hard_link_bytes);
1348 wim_timestamp_to_str(image_info->creation_time, buf, sizeof(buf));
1349 tprintf(T("Creation Time: %"TS"\n"), buf);
1351 wim_timestamp_to_str(image_info->last_modification_time, buf, sizeof(buf));
1352 tprintf(T("Last Modification Time: %"TS"\n"), buf);
1353 if (image_info->windows_info_exists)
1354 print_windows_info(&image_info->windows_info);
1355 if (image_info->flags)
1356 tprintf(T("Flags: %"TS"\n"), image_info->flags);
1357 tprintf(T("WIMBoot compatible: %"TS"\n"),
1358 image_info->wimboot ? T("yes") : T("no"));
1363 libxml_global_init(void)
1366 xmlInitCharEncodingHandlers();
1370 libxml_global_cleanup(void)
1373 xmlCleanupCharEncodingHandlers();
1376 /* Reads the XML data from a WIM file. */
1378 read_wim_xml_data(WIMStruct *wim)
1387 ret = wimlib_get_xml_data(wim, &buf, &bufsize);
1392 doc = xmlReadMemory((const char *)xml_data, bufsize,
1393 NULL, "UTF-16LE", 0);
1395 ERROR("Failed to parse XML data");
1396 ret = WIMLIB_ERR_XML;
1397 goto out_free_xml_data;
1400 root = xmlDocGetRootElement(doc);
1401 if (!root || !node_is_element(root) || !node_name_is(root, "WIM")) {
1402 ERROR("WIM XML data is invalid");
1403 ret = WIMLIB_ERR_XML;
1407 ret = xml_read_wim_info(root, &wim->wim_info);
1416 /* Prepares an in-memory buffer containing the UTF-16LE XML data for a WIM file.
1418 * total_bytes is the number to write in <TOTALBYTES>, or
1419 * WIM_TOTALBYTES_USE_EXISTING to use the existing value in memory, or
1420 * WIM_TOTALBYTES_OMIT to omit <TOTALBYTES> entirely.
1423 prepare_wim_xml_data(WIMStruct *wim, int image, u64 total_bytes,
1424 u8 **xml_data_ret, size_t *xml_len_ret)
1426 xmlCharEncodingHandler *encoding_handler;
1428 xmlOutputBuffer *outbuf;
1429 xmlTextWriter *writer;
1432 const xmlChar *content;
1437 /* Open an xmlTextWriter that writes to an in-memory buffer using
1438 * UTF-16LE encoding. */
1440 encoding_handler = xmlGetCharEncodingHandler(XML_CHAR_ENCODING_UTF16LE);
1441 if (!encoding_handler) {
1442 ERROR("Failed to get XML character encoding handler for UTF-16LE");
1443 ret = WIMLIB_ERR_LIBXML_UTF16_HANDLER_NOT_AVAILABLE;
1447 buf = xmlBufferCreate();
1449 ERROR("Failed to create xmlBuffer");
1450 ret = WIMLIB_ERR_NOMEM;
1454 outbuf = xmlOutputBufferCreateBuffer(buf, encoding_handler);
1456 ERROR("Failed to allocate xmlOutputBuffer");
1457 ret = WIMLIB_ERR_NOMEM;
1458 goto out_buffer_free;
1461 writer = xmlNewTextWriter(outbuf);
1463 ERROR("Failed to allocate xmlTextWriter");
1464 ret = WIMLIB_ERR_NOMEM;
1465 goto out_output_buffer_close;
1468 /* Write the XML document. */
1470 ret = xmlTextWriterStartElement(writer, "WIM");
1472 goto out_write_error;
1474 /* The contents of the <TOTALBYTES> element in the XML data, under the
1475 * <WIM> element (not the <IMAGE> element), is for non-split WIMs the
1476 * size of the WIM file excluding the XML data and integrity table.
1477 * For split WIMs, <TOTALBYTES> takes into account the entire WIM, not
1478 * just the current part. */
1479 if (total_bytes != WIM_TOTALBYTES_OMIT) {
1480 if (total_bytes == WIM_TOTALBYTES_USE_EXISTING) {
1482 total_bytes = wim->wim_info->total_bytes;
1486 ret = xmlTextWriterWriteFormatElement(writer, "TOTALBYTES",
1487 "%"PRIu64, total_bytes);
1489 goto out_write_error;
1492 if (image == WIMLIB_ALL_IMAGES) {
1494 last = wim->hdr.image_count;
1500 for (int i = first; i <= last; i++) {
1501 ret = xml_write_image_info(writer, &wim->wim_info->images[i - 1]);
1504 goto out_write_error;
1505 goto out_free_text_writer;
1509 ret = xmlTextWriterEndElement(writer);
1511 goto out_write_error;
1513 ret = xmlTextWriterEndDocument(writer);
1515 goto out_write_error;
1517 ret = xmlTextWriterFlush(writer);
1519 goto out_write_error;
1521 /* Retrieve the buffer into which the document was written. */
1523 content = xmlBufferContent(buf);
1524 len = xmlBufferLength(buf);
1526 /* Copy the data into a new buffer, and prefix it with the UTF-16LE BOM
1527 * (byte order mark), which is required by MS's software to understand
1531 xml_data = MALLOC(xml_len);
1533 ret = WIMLIB_ERR_NOMEM;
1534 goto out_free_text_writer;
1538 memcpy(&xml_data[2], content, len);
1540 /* Clean up libxml objects and return success. */
1541 *xml_data_ret = xml_data;
1542 *xml_len_ret = xml_len;
1544 out_free_text_writer:
1545 /* xmlFreeTextWriter will free the attached xmlOutputBuffer. */
1546 xmlFreeTextWriter(writer);
1547 goto out_buffer_free;
1548 out_output_buffer_close:
1549 xmlOutputBufferClose(outbuf);
1553 DEBUG("ret=%d", ret);
1557 ERROR("Error writing XML data");
1558 ret = WIMLIB_ERR_WRITE;
1559 goto out_free_text_writer;
1562 /* Writes the XML data to a WIM file. */
1564 write_wim_xml_data(WIMStruct *wim, int image, u64 total_bytes,
1565 struct wim_reshdr *out_reshdr,
1566 int write_resource_flags)
1572 DEBUG("Writing WIM XML data (image=%d, offset=%"PRIu64")",
1573 image, wim->out_fd.offset);
1575 ret = prepare_wim_xml_data(wim, image, total_bytes,
1576 &xml_data, &xml_len);
1580 /* Write the XML data uncompressed. Although wimlib can handle
1581 * compressed XML data, MS software cannot. */
1582 ret = write_wim_resource_from_buffer(xml_data,
1584 WIM_RESHDR_FLAG_METADATA,
1586 WIMLIB_COMPRESSION_TYPE_NONE,
1590 write_resource_flags);
1592 DEBUG("ret=%d", ret);
1596 /* API function documented in wimlib.h */
1597 WIMLIBAPI const tchar *
1598 wimlib_get_image_name(const WIMStruct *wim, int image)
1600 if (image < 1 || image > wim->hdr.image_count)
1602 return wim->wim_info->images[image - 1].name;
1605 /* API function documented in wimlib.h */
1606 WIMLIBAPI const tchar *
1607 wimlib_get_image_description(const WIMStruct *wim, int image)
1609 if (image < 1 || image > wim->hdr.image_count)
1611 return wim->wim_info->images[image - 1].description;
1614 /* API function documented in wimlib.h */
1616 wimlib_image_name_in_use(const WIMStruct *wim, const tchar *name)
1618 if (!name || !*name)
1620 for (int i = 1; i <= wim->hdr.image_count; i++)
1621 if (!tstrcmp(wim->wim_info->images[i - 1].name, name))
1627 /* API function documented in wimlib.h */
1629 wimlib_get_xml_data(WIMStruct *wim, void **buf_ret, size_t *bufsize_ret)
1631 const struct wim_reshdr *xml_reshdr;
1633 if (wim->filename == NULL && filedes_is_seekable(&wim->in_fd))
1634 return WIMLIB_ERR_INVALID_PARAM;
1636 if (buf_ret == NULL || bufsize_ret == NULL)
1637 return WIMLIB_ERR_INVALID_PARAM;
1639 xml_reshdr = &wim->hdr.xml_data_reshdr;
1641 DEBUG("Reading XML data.");
1642 *bufsize_ret = xml_reshdr->uncompressed_size;
1643 return wim_reshdr_to_data(xml_reshdr, wim, buf_ret);
1647 wimlib_extract_xml_data(WIMStruct *wim, FILE *fp)
1653 ret = wimlib_get_xml_data(wim, &buf, &bufsize);
1657 if (fwrite(buf, 1, bufsize, fp) != bufsize) {
1658 ERROR_WITH_ERRNO("Failed to extract XML data");
1659 ret = WIMLIB_ERR_WRITE;
1665 /* API function documented in wimlib.h */
1667 wimlib_set_image_name(WIMStruct *wim, int image, const tchar *name)
1675 if (image < 1 || image > wim->hdr.image_count)
1676 return WIMLIB_ERR_INVALID_IMAGE;
1679 for (i = 1; i <= wim->hdr.image_count; i++) {
1682 if (!tstrcmp(wim->wim_info->images[i - 1].name, name))
1683 return WIMLIB_ERR_IMAGE_NAME_COLLISION;
1689 return WIMLIB_ERR_NOMEM;
1691 FREE(wim->wim_info->images[image - 1].name);
1692 wim->wim_info->images[image - 1].name = p;
1697 do_set_image_info_str(WIMStruct *wim, int image, const tchar *tstr,
1701 tchar **dest_tstr_p;
1703 if (image < 1 || image > wim->hdr.image_count) {
1704 ERROR("%d is not a valid image", image);
1705 return WIMLIB_ERR_INVALID_IMAGE;
1708 tstr_copy = TSTRDUP(tstr);
1710 return WIMLIB_ERR_NOMEM;
1714 dest_tstr_p = (tchar**)((void*)&wim->wim_info->images[image - 1] + offset);
1717 *dest_tstr_p = tstr_copy;
1721 /* API function documented in wimlib.h */
1723 wimlib_set_image_descripton(WIMStruct *wim, int image,
1724 const tchar *description)
1726 return do_set_image_info_str(wim, image, description,
1727 offsetof(struct image_info, description));
1730 /* API function documented in wimlib.h */
1732 wimlib_set_image_flags(WIMStruct *wim, int image, const tchar *flags)
1734 return do_set_image_info_str(wim, image, flags,
1735 offsetof(struct image_info, flags));