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 part of wimlib, a library for working with WIM files.
12 * wimlib is free software; you can redistribute it and/or modify it under the
13 * terms of the GNU General Public License as published by the Free
14 * Software Foundation; either version 3 of the License, or (at your option)
17 * wimlib is distributed in the hope that it will be useful, but WITHOUT ANY
18 * WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
19 * A PARTICULAR PURPOSE. See the GNU General Public License for more
22 * You should have received a copy of the GNU General Public License
23 * along with wimlib; if not, see http://www.gnu.org/licenses/.
30 #include "wimlib/dentry.h"
31 #include "wimlib/encoding.h"
32 #include "wimlib/error.h"
33 #include "wimlib/file_io.h"
34 #include "wimlib/lookup_table.h"
35 #include "wimlib/metadata.h"
36 #include "wimlib/resource.h"
37 #include "wimlib/timestamp.h"
38 #include "wimlib/xml.h"
39 #include "wimlib/write.h"
41 #include <libxml/encoding.h>
42 #include <libxml/parser.h>
43 #include <libxml/tree.h>
44 #include <libxml/xmlwriter.h>
49 /* Structures used to form an in-memory representation of the XML data (other
50 * than the raw parse tree from libxml). */
52 struct windows_version {
64 tchar *installation_type;
65 tchar *pkeyconfigversion;
70 tchar *default_language;
73 bool windows_version_exists;
74 struct windows_version windows_version;
79 bool windows_info_exists;
85 u64 last_modification_time;
86 struct windows_info windows_info;
90 tchar *display_description;
94 /* Note: must update clone_image_info() if adding new fields here */
96 struct wim_lookup_table *lookup_table; /* temporary field */
99 /* A struct wim_info structure corresponds to the entire XML data for a WIM file. */
103 /* Array of `struct image_info's, one for each image in the WIM that is
104 * mentioned in the XML data. */
105 struct image_info *images;
108 struct xml_string_spec {
113 #define ELEM(STRING_NAME, MEMBER_NAME) \
114 {STRING_NAME, offsetof(struct image_info, MEMBER_NAME)}
115 static const struct xml_string_spec
116 image_info_xml_string_specs[] = {
118 ELEM("DESCRIPTION", description),
119 ELEM("DISPLAYNAME", display_name),
120 ELEM("DISPLAYDESCRIPTION", display_description),
121 ELEM("FLAGS", flags),
125 #define ELEM(STRING_NAME, MEMBER_NAME) \
126 {STRING_NAME, offsetof(struct windows_info, MEMBER_NAME)}
127 static const struct xml_string_spec
128 windows_info_xml_string_specs[] = {
129 ELEM("PRODUCTNAME", product_name),
130 ELEM("EDITIONID", edition_id),
131 ELEM("INSTALLATIONTYPE", installation_type),
133 ELEM("PRODUCTTYPE", product_type),
134 ELEM("PRODUCTSUITE", product_suite),
139 wim_info_get_total_bytes(const struct wim_info *info)
142 return info->total_bytes;
148 wim_info_get_image_hard_link_bytes(const struct wim_info *info, int image)
151 return info->images[image - 1].hard_link_bytes;
157 wim_info_get_image_total_bytes(const struct wim_info *info, int image)
160 return info->images[image - 1].total_bytes;
166 wim_info_get_num_images(const struct wim_info *info)
169 return info->num_images;
175 wim_info_set_wimboot(struct wim_info *info, int image, bool value)
177 info->images[image - 1].wimboot = value;
181 wim_info_get_wimboot(const struct wim_info *info, int image)
183 return info->images[image - 1].wimboot;
186 /* Architecture constants are from w64 mingw winnt.h */
187 #define PROCESSOR_ARCHITECTURE_INTEL 0
188 #define PROCESSOR_ARCHITECTURE_MIPS 1
189 #define PROCESSOR_ARCHITECTURE_ALPHA 2
190 #define PROCESSOR_ARCHITECTURE_PPC 3
191 #define PROCESSOR_ARCHITECTURE_SHX 4
192 #define PROCESSOR_ARCHITECTURE_ARM 5
193 #define PROCESSOR_ARCHITECTURE_IA64 6
194 #define PROCESSOR_ARCHITECTURE_ALPHA64 7
195 #define PROCESSOR_ARCHITECTURE_MSIL 8
196 #define PROCESSOR_ARCHITECTURE_AMD64 9
197 #define PROCESSOR_ARCHITECTURE_IA32_ON_WIN64 10
199 /* Returns a statically allocated string that is a string representation of the
200 * architecture number. */
205 case PROCESSOR_ARCHITECTURE_INTEL:
207 case PROCESSOR_ARCHITECTURE_MIPS:
209 case PROCESSOR_ARCHITECTURE_ARM:
211 case PROCESSOR_ARCHITECTURE_IA64:
213 case PROCESSOR_ARCHITECTURE_AMD64:
221 /* Iterate through the children of an xmlNode. */
222 #define for_node_child(parent, child) \
223 for (child = parent->children; child != NULL; child = child->next)
225 /* Utility functions for xmlNodes */
227 node_is_element(xmlNode *node)
229 return node->type == XML_ELEMENT_NODE;
233 node_is_text(xmlNode *node)
235 return node->type == XML_TEXT_NODE;
239 node_name_is(xmlNode *node, const char *name)
241 /* For now, both upper case and lower case element names are accepted. */
242 return strcasecmp((const char *)node->name, name) == 0;
246 node_get_number(const xmlNode *u64_node, int base)
249 for_node_child(u64_node, child)
250 if (node_is_text(child))
251 return strtoull(child->content, NULL, base);
255 /* Finds the text node that is a child of an element node and returns its
256 * content converted to a 64-bit unsigned integer. Returns 0 if no text node is
259 node_get_u64(const xmlNode *u64_node)
261 return node_get_number(u64_node, 10);
264 /* Like node_get_u64(), but expects a number in base 16. */
266 node_get_hex_u64(const xmlNode *u64_node)
268 return node_get_number(u64_node, 16);
272 node_get_string(const xmlNode *string_node, tchar **tstr_ret)
281 for_node_child(string_node, child) {
282 if (node_is_text(child) && child->content) {
283 ret = utf8_to_tstr_simple(child->content, &tstr);
293 /* Returns the timestamp from a time node. It has child elements <HIGHPART> and
294 * <LOWPART> that are then used to construct a 64-bit timestamp. */
296 node_get_timestamp(const xmlNode *time_node)
301 for_node_child(time_node, child) {
302 if (!node_is_element(child))
304 if (node_name_is(child, "HIGHPART"))
305 high_part = node_get_hex_u64(child);
306 else if (node_name_is(child, "LOWPART"))
307 low_part = node_get_hex_u64(child);
309 return (u64)low_part | ((u64)high_part << 32);
312 /* Used to sort an array of struct image_infos by their image indices. */
314 sort_by_index(const void *p1, const void *p2)
316 int index_1 = ((const struct image_info*)p1)->index;
317 int index_2 = ((const struct image_info*)p2)->index;
318 if (index_1 < index_2)
320 else if (index_1 > index_2)
327 /* Frees memory allocated inside a struct windows_info structure. */
329 destroy_windows_info(struct windows_info *windows_info)
331 FREE(windows_info->product_name);
332 FREE(windows_info->edition_id);
333 FREE(windows_info->installation_type);
334 FREE(windows_info->hal);
335 FREE(windows_info->product_type);
336 FREE(windows_info->product_suite);
337 FREE(windows_info->pkeyconfigversion);
338 for (size_t i = 0; i < windows_info->num_languages; i++)
339 FREE(windows_info->languages[i]);
340 FREE(windows_info->languages);
341 FREE(windows_info->default_language);
342 FREE(windows_info->system_root);
345 /* Frees memory allocated inside a struct image_info structure. */
347 destroy_image_info(struct image_info *image_info)
349 FREE(image_info->name);
350 FREE(image_info->description);
351 FREE(image_info->flags);
352 FREE(image_info->display_name);
353 FREE(image_info->display_description);
354 destroy_windows_info(&image_info->windows_info);
355 memset(image_info, 0, sizeof(struct image_info));
359 free_wim_info(struct wim_info *info)
363 for (int i = 0; i < info->num_images; i++)
364 destroy_image_info(&info->images[i]);
371 /* Reads the information from a <VERSION> element inside the <WINDOWS> element.
374 xml_read_windows_version(const xmlNode *version_node,
375 struct windows_version* windows_version)
378 for_node_child(version_node, child) {
379 if (!node_is_element(child))
381 if (node_name_is(child, "MAJOR"))
382 windows_version->major = node_get_u64(child);
383 else if (node_name_is(child, "MINOR"))
384 windows_version->minor = node_get_u64(child);
385 else if (node_name_is(child, "BUILD"))
386 windows_version->build = node_get_u64(child);
387 else if (node_name_is(child, "SPBUILD"))
388 windows_version->sp_build = node_get_u64(child);
389 else if (node_name_is(child, "SPLEVEL"))
390 windows_version->sp_level = node_get_u64(child);
394 /* Reads the information from a <LANGUAGE> element inside a <WINDOWS> element.
397 xml_read_languages(const xmlNode *languages_node,
398 tchar ***languages_ret,
399 size_t *num_languages_ret,
400 tchar **default_language_ret)
403 size_t num_languages = 0;
407 for_node_child(languages_node, child)
408 if (node_is_element(child) && node_name_is(child, "LANGUAGE"))
411 languages = CALLOC(num_languages, sizeof(languages[0]));
413 return WIMLIB_ERR_NOMEM;
415 *languages_ret = languages;
416 *num_languages_ret = num_languages;
419 for_node_child(languages_node, child) {
420 if (!node_is_element(child))
422 if (node_name_is(child, "LANGUAGE"))
423 ret = node_get_string(child, languages++);
424 else if (node_name_is(child, "DEFAULT"))
425 ret = node_get_string(child, default_language_ret);
432 /* Reads the information from a <WINDOWS> element inside an <IMAGE> element. */
434 xml_read_windows_info(const xmlNode *windows_node,
435 struct windows_info *windows_info)
440 for_node_child(windows_node, child) {
441 if (!node_is_element(child))
443 if (node_name_is(child, "ARCH")) {
444 windows_info->arch = node_get_u64(child);
445 } else if (node_name_is(child, "PRODUCTNAME")) {
446 ret = node_get_string(child,
447 &windows_info->product_name);
448 } else if (node_name_is(child, "EDITIONID")) {
449 ret = node_get_string(child,
450 &windows_info->edition_id);
451 } else if (node_name_is(child, "INSTALLATIONTYPE")) {
452 ret = node_get_string(child,
453 &windows_info->installation_type);
454 } else if (node_name_is(child, "PRODUCTTYPE")) {
455 ret = node_get_string(child,
456 &windows_info->product_type);
457 } else if (node_name_is(child, "PRODUCTSUITE")) {
458 ret = node_get_string(child,
459 &windows_info->product_suite);
460 } else if (node_name_is(child, "LANGUAGES")) {
461 ret = xml_read_languages(child,
462 &windows_info->languages,
463 &windows_info->num_languages,
464 &windows_info->default_language);
465 } else if (node_name_is(child, "VERSION")) {
466 xml_read_windows_version(child,
467 &windows_info->windows_version);
468 windows_info->windows_version_exists = true;
469 } else if (node_name_is(child, "SYSTEMROOT")) {
470 ret = node_get_string(child, &windows_info->system_root);
471 } else if (node_name_is(child, "HAL")) {
472 ret = node_get_string(child, &windows_info->hal);
473 } else if (node_name_is(child, "SERVICINGDATA")) {
476 for_node_child(child, grandchild) {
477 if (node_is_element(grandchild) &&
478 node_name_is(grandchild, "PKEYCONFIGVERSION"))
480 ret = node_get_string(grandchild,
481 &windows_info->pkeyconfigversion);
492 /* Reads the information from an <IMAGE> element. */
494 xml_read_image_info(xmlNode *image_node, struct image_info *image_info)
500 index_prop = xmlGetProp(image_node, "INDEX");
502 image_info->index = atoi(index_prop);
505 image_info->index = 1;
509 for_node_child(image_node, child) {
510 if (!node_is_element(child))
512 if (node_name_is(child, "DIRCOUNT"))
513 image_info->dir_count = node_get_u64(child);
514 else if (node_name_is(child, "FILECOUNT"))
515 image_info->file_count = node_get_u64(child);
516 else if (node_name_is(child, "TOTALBYTES"))
517 image_info->total_bytes = node_get_u64(child);
518 else if (node_name_is(child, "HARDLINKBYTES"))
519 image_info->hard_link_bytes = node_get_u64(child);
520 else if (node_name_is(child, "CREATIONTIME"))
521 image_info->creation_time = node_get_timestamp(child);
522 else if (node_name_is(child, "LASTMODIFICATIONTIME"))
523 image_info->last_modification_time = node_get_timestamp(child);
524 else if (node_name_is(child, "WINDOWS")) {
525 DEBUG("Found <WINDOWS> tag");
526 ret = xml_read_windows_info(child,
527 &image_info->windows_info);
528 image_info->windows_info_exists = true;
529 } else if (node_name_is(child, "NAME")) {
530 ret = node_get_string(child, &image_info->name);
531 } else if (node_name_is(child, "DESCRIPTION")) {
532 ret = node_get_string(child, &image_info->description);
533 } else if (node_name_is(child, "FLAGS")) {
534 ret = node_get_string(child, &image_info->flags);
535 } else if (node_name_is(child, "DISPLAYNAME")) {
536 ret = node_get_string(child, &image_info->display_name);
537 } else if (node_name_is(child, "DISPLAYDESCRIPTION")) {
538 ret = node_get_string(child, &image_info->display_description);
539 } else if (node_name_is(child, "WIMBOOT")) {
540 if (node_get_u64(child) == 1) {
541 image_info->wimboot = true;
547 if (!image_info->name) {
549 /*WARNING("Image with index %d has no name", image_info->index);*/
550 empty_name = MALLOC(sizeof(tchar));
552 return WIMLIB_ERR_NOMEM;
553 *empty_name = T('\0');
554 image_info->name = empty_name;
559 /* Reads the information from a <WIM> element, which should be the root element
560 * of the XML tree. */
562 xml_read_wim_info(const xmlNode *wim_node, struct wim_info **wim_info_ret)
564 struct wim_info *wim_info;
570 wim_info = CALLOC(1, sizeof(struct wim_info));
572 return WIMLIB_ERR_NOMEM;
574 /* Count how many images there are. */
576 for_node_child(wim_node, child) {
577 if (node_is_element(child) && node_name_is(child, "IMAGE")) {
578 if (unlikely(num_images == MAX_IMAGES)) {
579 ret = WIMLIB_ERR_IMAGE_COUNT;
586 if (num_images > 0) {
587 /* Allocate the array of struct image_infos and fill them in. */
588 wim_info->images = CALLOC(num_images, sizeof(wim_info->images[0]));
589 if (!wim_info->images) {
590 ret = WIMLIB_ERR_NOMEM;
593 wim_info->num_images = num_images;
595 for_node_child(wim_node, child) {
596 if (!node_is_element(child))
598 if (node_name_is(child, "IMAGE")) {
599 DEBUG("Found <IMAGE> tag");
600 ret = xml_read_image_info(child,
601 &wim_info->images[i]);
605 } else if (node_name_is(child, "TOTALBYTES")) {
606 wim_info->total_bytes = node_get_u64(child);
607 } else if (node_name_is(child, "ESD")) {
609 for_node_child(child, esdchild) {
610 if (node_is_element(esdchild) &&
611 node_name_is(esdchild, "ENCRYPTED"))
613 ret = WIMLIB_ERR_WIM_IS_ENCRYPTED;
620 /* Sort the array of image info by image index. */
621 qsort(wim_info->images, num_images,
622 sizeof(struct image_info), sort_by_index);
624 /* Make sure the image indices make sense */
625 for (i = 0; i < num_images; i++) {
626 if (wim_info->images[i].index != i + 1) {
627 ERROR("WIM images are not indexed [1...%d] "
628 "in XML data as expected",
630 ret = WIMLIB_ERR_IMAGE_COUNT;
636 *wim_info_ret = wim_info;
639 free_wim_info(wim_info);
643 /* Prints the information contained in a `struct windows_info'.
645 * Warning: any strings printed here are in UTF-8 encoding. If the locale
646 * character encoding is not UTF-8, the printed strings may be garbled. */
648 print_windows_info(const struct windows_info *windows_info)
650 const struct windows_version *windows_version;
652 tprintf(T("Architecture: %"TS"\n"),
653 get_arch(windows_info->arch));
655 if (windows_info->product_name) {
656 tprintf(T("Product Name: %"TS"\n"),
657 windows_info->product_name);
660 if (windows_info->edition_id) {
661 tprintf(T("Edition ID: %"TS"\n"),
662 windows_info->edition_id);
665 if (windows_info->installation_type) {
666 tprintf(T("Installation Type: %"TS"\n"),
667 windows_info->installation_type);
670 if (windows_info->hal) {
671 tprintf(T("HAL: %"TS"\n"),
675 if (windows_info->product_type) {
676 tprintf(T("Product Type: %"TS"\n"),
677 windows_info->product_type);
680 if (windows_info->product_suite) {
681 tprintf(T("Product Suite: %"TS"\n"),
682 windows_info->product_suite);
685 tprintf(T("Languages: "));
686 for (size_t i = 0; i < windows_info->num_languages; i++) {
688 tfputs(windows_info->languages[i], stdout);
692 if (windows_info->default_language) {
693 tprintf(T("Default Language: %"TS"\n"),
694 windows_info->default_language);
696 if (windows_info->system_root) {
697 tprintf(T("System Root: %"TS"\n"),
698 windows_info->system_root);
701 if (windows_info->windows_version_exists) {
702 windows_version = &windows_info->windows_version;
703 tprintf(T("Major Version: %"PRIu64"\n"),
704 windows_version->major);
705 tprintf(T("Minor Version: %"PRIu64"\n"),
706 windows_version->minor);
707 tprintf(T("Build: %"PRIu64"\n"),
708 windows_version->build);
709 tprintf(T("Service Pack Build: %"PRIu64"\n"),
710 windows_version->sp_build);
711 tprintf(T("Service Pack Level: %"PRIu64"\n"),
712 windows_version->sp_level);
717 xml_write_string(xmlTextWriter *writer, const char *name,
722 int rc = tstr_to_utf8_simple(tstr, &utf8_str);
725 rc = xmlTextWriterWriteElement(writer, name, utf8_str);
734 xml_write_strings_from_specs(xmlTextWriter *writer,
735 const void *struct_with_strings,
736 const struct xml_string_spec specs[],
739 for (size_t i = 0; i < num_specs; i++) {
740 int rc = xml_write_string(writer, specs[i].name,
741 *(const tchar * const *)
742 (struct_with_strings + specs[i].offset));
750 dup_strings_from_specs(const void *old_struct_with_strings,
751 void *new_struct_with_strings,
752 const struct xml_string_spec specs[],
755 for (size_t i = 0; i < num_specs; i++) {
756 const tchar *old_str = *(const tchar * const *)
757 ((const void*)old_struct_with_strings + specs[i].offset);
758 tchar **new_str_p = (tchar **)((void*)new_struct_with_strings + specs[i].offset);
760 *new_str_p = TSTRDUP(old_str);
762 return WIMLIB_ERR_NOMEM;
768 /* Writes the information contained in a `struct windows_version' to the XML
769 * document being written. This is the <VERSION> element inside the <WINDOWS>
772 xml_write_windows_version(xmlTextWriter *writer,
773 const struct windows_version *version)
776 rc = xmlTextWriterStartElement(writer, "VERSION");
780 rc = xmlTextWriterWriteFormatElement(writer, "MAJOR", "%"PRIu64,
785 rc = xmlTextWriterWriteFormatElement(writer, "MINOR", "%"PRIu64,
790 rc = xmlTextWriterWriteFormatElement(writer, "BUILD", "%"PRIu64,
795 rc = xmlTextWriterWriteFormatElement(writer, "SPBUILD", "%"PRIu64,
800 rc = xmlTextWriterWriteFormatElement(writer, "SPLEVEL", "%"PRIu64,
805 rc = xmlTextWriterEndElement(writer); /* </VERSION> */
812 /* Writes the information contained in a `struct windows_info' to the XML
813 * document being written. This is the <WINDOWS> element. */
815 xml_write_windows_info(xmlTextWriter *writer,
816 const struct windows_info *windows_info)
819 rc = xmlTextWriterStartElement(writer, "WINDOWS");
823 rc = xmlTextWriterWriteFormatElement(writer, "ARCH", "%"PRIu64,
828 rc = xml_write_strings_from_specs(writer,
830 windows_info_xml_string_specs,
831 ARRAY_LEN(windows_info_xml_string_specs));
835 if (windows_info->num_languages) {
836 rc = xmlTextWriterStartElement(writer, "LANGUAGES");
840 for (size_t i = 0; i < windows_info->num_languages; i++) {
841 rc = xml_write_string(writer, "LANGUAGE",
842 windows_info->languages[i]);
847 rc = xml_write_string(writer, "DEFAULT",
848 windows_info->default_language);
852 rc = xmlTextWriterEndElement(writer); /* </LANGUAGES> */
857 if (windows_info->pkeyconfigversion) {
858 rc = xmlTextWriterStartElement(writer, "SERVICINGDATA");
862 rc = xml_write_string(writer, "PKEYCONFIGVERSION",
863 windows_info->pkeyconfigversion);
867 rc = xmlTextWriterEndElement(writer);
872 if (windows_info->windows_version_exists) {
873 rc = xml_write_windows_version(writer, &windows_info->windows_version);
878 rc = xml_write_string(writer, "SYSTEMROOT", windows_info->system_root);
882 rc = xmlTextWriterEndElement(writer); /* </WINDOWS> */
889 /* Writes a time element to the XML document being constructed in memory. */
891 xml_write_time(xmlTextWriter *writer, const char *element_name, u64 time)
894 rc = xmlTextWriterStartElement(writer, element_name);
898 rc = xmlTextWriterWriteFormatElement(writer, "HIGHPART",
899 "0x%08"PRIX32, (u32)(time >> 32));
903 rc = xmlTextWriterWriteFormatElement(writer, "LOWPART",
904 "0x%08"PRIX32, (u32)time);
908 rc = xmlTextWriterEndElement(writer); /* </@element_name> */
914 /* Writes an <IMAGE> element to the XML document. */
916 xml_write_image_info(xmlTextWriter *writer, const struct image_info *image_info)
919 rc = xmlTextWriterStartElement(writer, "IMAGE");
923 rc = xmlTextWriterWriteFormatAttribute(writer, "INDEX", "%d",
928 rc = xmlTextWriterWriteFormatElement(writer, "DIRCOUNT", "%"PRIu64,
929 image_info->dir_count);
933 rc = xmlTextWriterWriteFormatElement(writer, "FILECOUNT", "%"PRIu64,
934 image_info->file_count);
938 rc = xmlTextWriterWriteFormatElement(writer, "TOTALBYTES", "%"PRIu64,
939 image_info->total_bytes);
943 rc = xmlTextWriterWriteFormatElement(writer, "HARDLINKBYTES", "%"PRIu64,
944 image_info->hard_link_bytes);
948 rc = xml_write_time(writer, "CREATIONTIME", image_info->creation_time);
952 rc = xml_write_time(writer, "LASTMODIFICATIONTIME",
953 image_info->last_modification_time);
957 if (image_info->windows_info_exists) {
958 rc = xml_write_windows_info(writer, &image_info->windows_info);
963 rc = xml_write_strings_from_specs(writer, image_info,
964 image_info_xml_string_specs,
965 ARRAY_LEN(image_info_xml_string_specs));
969 if (image_info->wimboot) {
970 rc = xmlTextWriterWriteFormatElement(writer, "WIMBOOT", "%d", 1);
975 rc = xmlTextWriterEndElement(writer); /* </IMAGE> */
984 /* Makes space for another image in the XML information and return a pointer to
986 static struct image_info *
987 add_image_info_struct(struct wim_info *wim_info)
989 struct image_info *images;
991 images = CALLOC(wim_info->num_images + 1, sizeof(struct image_info));
994 memcpy(images, wim_info->images,
995 wim_info->num_images * sizeof(struct image_info));
996 FREE(wim_info->images);
997 wim_info->images = images;
998 wim_info->num_images++;
999 return &images[wim_info->num_images - 1];
1003 clone_windows_info(const struct windows_info *old, struct windows_info *new)
1007 new->arch = old->arch;
1009 ret = dup_strings_from_specs(old, new, windows_info_xml_string_specs,
1010 ARRAY_LEN(windows_info_xml_string_specs));
1014 if (old->pkeyconfigversion) {
1015 new->pkeyconfigversion = TSTRDUP(old->pkeyconfigversion);
1016 if (new->pkeyconfigversion == NULL)
1017 return WIMLIB_ERR_NOMEM;
1020 if (old->languages) {
1021 new->languages = CALLOC(old->num_languages, sizeof(new->languages[0]));
1022 if (!new->languages)
1023 return WIMLIB_ERR_NOMEM;
1024 new->num_languages = old->num_languages;
1025 for (size_t i = 0; i < new->num_languages; i++) {
1026 if (!old->languages[i])
1028 new->languages[i] = TSTRDUP(old->languages[i]);
1029 if (!new->languages[i])
1030 return WIMLIB_ERR_NOMEM;
1033 if (old->default_language &&
1034 !(new->default_language = TSTRDUP(old->default_language)))
1035 return WIMLIB_ERR_NOMEM;
1036 if (old->system_root && !(new->system_root = TSTRDUP(old->system_root)))
1037 return WIMLIB_ERR_NOMEM;
1038 if (old->windows_version_exists) {
1039 new->windows_version_exists = true;
1040 memcpy(&new->windows_version, &old->windows_version,
1041 sizeof(old->windows_version));
1047 clone_image_info(const struct image_info *old, struct image_info *new)
1051 new->dir_count = old->dir_count;
1052 new->file_count = old->file_count;
1053 new->total_bytes = old->total_bytes;
1054 new->hard_link_bytes = old->hard_link_bytes;
1055 new->creation_time = old->creation_time;
1056 new->last_modification_time = old->last_modification_time;
1058 ret = dup_strings_from_specs(old, new,
1059 image_info_xml_string_specs,
1060 ARRAY_LEN(image_info_xml_string_specs));
1064 if (old->windows_info_exists) {
1065 new->windows_info_exists = true;
1066 ret = clone_windows_info(&old->windows_info,
1067 &new->windows_info);
1071 new->wimboot = old->wimboot;
1075 /* Copies the XML information for an image between WIM files.
1077 * @dest_image_name and @dest_image_description are ignored if they are NULL;
1078 * otherwise, they are used to override the image name and/or image description
1079 * from the XML data in the source WIM file.
1081 * On failure, WIMLIB_ERR_NOMEM is returned and no changes are made. Otherwise,
1082 * 0 is returned and the WIM information at *new_wim_info_p is modified.
1085 xml_export_image(const struct wim_info *old_wim_info,
1087 struct wim_info **new_wim_info_p,
1088 const tchar *dest_image_name,
1089 const tchar *dest_image_description)
1091 struct wim_info *new_wim_info;
1092 struct image_info *image_info;
1095 DEBUG("Copying XML data between WIM files for source image %d.", image);
1097 wimlib_assert(old_wim_info != NULL);
1098 wimlib_assert(image >= 1 && image <= old_wim_info->num_images);
1100 if (*new_wim_info_p) {
1101 new_wim_info = *new_wim_info_p;
1103 new_wim_info = CALLOC(1, sizeof(struct wim_info));
1108 image_info = add_image_info_struct(new_wim_info);
1112 ret = clone_image_info(&old_wim_info->images[image - 1], image_info);
1114 goto err_destroy_image_info;
1116 image_info->index = new_wim_info->num_images;
1118 if (dest_image_name) {
1119 FREE(image_info->name);
1120 image_info->name = TSTRDUP(dest_image_name);
1121 if (!image_info->name)
1122 goto err_destroy_image_info;
1124 if (dest_image_description) {
1125 FREE(image_info->description);
1126 image_info->description = TSTRDUP(dest_image_description);
1127 if (!image_info->description)
1128 goto err_destroy_image_info;
1130 *new_wim_info_p = new_wim_info;
1132 err_destroy_image_info:
1133 destroy_image_info(image_info);
1135 if (new_wim_info != *new_wim_info_p)
1136 free_wim_info(new_wim_info);
1137 return WIMLIB_ERR_NOMEM;
1140 /* Removes an image from the XML information. */
1142 xml_delete_image(struct wim_info **wim_info_p, int image)
1144 struct wim_info *wim_info;
1146 wim_info = *wim_info_p;
1147 wimlib_assert(image >= 1 && image <= wim_info->num_images);
1148 DEBUG("Deleting image %d from the XML data.", image);
1150 destroy_image_info(&wim_info->images[image - 1]);
1152 memmove(&wim_info->images[image - 1],
1153 &wim_info->images[image],
1154 (wim_info->num_images - image) * sizeof(struct image_info));
1156 if (--wim_info->num_images == 0) {
1157 free_wim_info(wim_info);
1160 for (int i = image - 1; i < wim_info->num_images; i++)
1161 wim_info->images[i].index--;
1166 xml_get_max_image_name_len(const WIMStruct *wim)
1169 for (u32 i = 0; i < wim->hdr.image_count; i++)
1170 max_len = max(max_len, tstrlen(wim->wim_info->images[i].name));
1175 xml_set_memory_allocator(void *(*malloc_func)(size_t),
1176 void (*free_func)(void *),
1177 void *(*realloc_func)(void *, size_t))
1179 xmlMemSetup(free_func, malloc_func, realloc_func, STRDUP);
1183 calculate_dentry_statistics(struct wim_dentry *dentry, void *arg)
1185 struct image_info *info = arg;
1186 const struct wim_inode *inode = dentry->d_inode;
1188 /* Update directory count and file count.
1190 * Each dentry counts as either a file or a directory, but not both.
1191 * The root directory is an exception: it is not counted at all.
1193 * Symbolic links and junction points (and presumably other reparse
1194 * points) count as regular files. This is despite the fact that
1195 * junction points have FILE_ATTRIBUTE_DIRECTORY set.
1198 if (!dentry_is_root(dentry)) {
1199 if (inode_is_directory(inode))
1206 * Update total bytes and hard link bytes.
1208 * We try to act the same as the MS implementation, even though there
1209 * are some inconsistencies/bugs in the way it operates.
1211 * If there are no alternate data streams in the image, the "total
1212 * bytes" is the sum of the size of the un-named data stream of each
1213 * inode times the link count of that inode. In other words, it would
1214 * be the total number of bytes of regular files you would have if you
1215 * extracted the full image without any hard-links. The "hard link
1216 * bytes" is equal to the "total bytes" minus the size of the un-named
1217 * data stream of each inode. In other words, the "hard link bytes"
1218 * counts the size of the un-named data stream for all the links to each
1219 * inode except the first one.
1221 * Reparse points and directories don't seem to be counted in either the
1222 * total bytes or the hard link bytes.
1224 * And now we get to the most confusing part, the alternate data
1225 * streams. They are not counted in the "total bytes". However, if the
1226 * link count of an inode with alternate data streams is 2 or greater,
1227 * the size of all the alternate data streams is included in the "hard
1228 * link bytes", and this size is multiplied by the link count (NOT one
1229 * less than the link count).
1231 if (!(inode->i_attributes & (FILE_ATTRIBUTE_DIRECTORY |
1232 FILE_ATTRIBUTE_REPARSE_POINT)))
1234 struct wim_lookup_table_entry *lte;
1236 lte = inode_unnamed_lte(inode, info->lookup_table);
1238 info->total_bytes += lte->size;
1239 if (!dentry_is_first_in_inode(dentry))
1240 info->hard_link_bytes += lte->size;
1243 if (inode->i_nlink >= 2 && dentry_is_first_in_inode(dentry)) {
1244 for (unsigned i = 0; i < inode->i_num_ads; i++) {
1245 if (inode->i_ads_entries[i].stream_name_nbytes) {
1246 lte = inode_stream_lte(inode, i + 1, info->lookup_table);
1248 info->hard_link_bytes += inode->i_nlink *
1259 * Calculate what to put in the <FILECOUNT>, <DIRCOUNT>, <TOTALBYTES>, and
1260 * <HARDLINKBYTES> elements of each <IMAGE>.
1262 * Please note there is no official documentation for exactly how this is done.
1263 * But, see calculate_dentry_statistics().
1266 xml_update_image_info(WIMStruct *wim, int image)
1268 struct image_info *image_info;
1270 DEBUG("Updating the image info for image %d", image);
1272 image_info = &wim->wim_info->images[image - 1];
1274 image_info->file_count = 0;
1275 image_info->dir_count = 0;
1276 image_info->total_bytes = 0;
1277 image_info->hard_link_bytes = 0;
1278 image_info->lookup_table = wim->lookup_table;
1280 for_dentry_in_tree(wim->image_metadata[image - 1]->root_dentry,
1281 calculate_dentry_statistics,
1283 image_info->last_modification_time = get_wim_timestamp();
1286 /* Adds an image to the XML information. */
1288 xml_add_image(WIMStruct *wim, const tchar *name)
1290 struct wim_info *wim_info;
1291 struct image_info *image_info;
1293 wimlib_assert(name != NULL);
1295 /* If this is the first image, allocate the struct wim_info. Otherwise
1296 * use the existing struct wim_info. */
1297 if (wim->wim_info) {
1298 wim_info = wim->wim_info;
1300 wim_info = CALLOC(1, sizeof(struct wim_info));
1302 return WIMLIB_ERR_NOMEM;
1305 image_info = add_image_info_struct(wim_info);
1307 goto out_free_wim_info;
1309 if (!(image_info->name = TSTRDUP(name)))
1310 goto out_destroy_image_info;
1312 wim->wim_info = wim_info;
1313 image_info->index = wim_info->num_images;
1314 image_info->creation_time = get_wim_timestamp();
1315 xml_update_image_info(wim, image_info->index);
1318 out_destroy_image_info:
1319 destroy_image_info(image_info);
1320 wim_info->num_images--;
1322 if (wim_info != wim->wim_info)
1324 return WIMLIB_ERR_NOMEM;
1327 /* Prints information about the specified image from struct wim_info structure.
1330 print_image_info(const struct wim_info *wim_info, int image)
1332 const struct image_info *image_info;
1336 wimlib_assert(image >= 1 && image <= wim_info->num_images);
1338 image_info = &wim_info->images[image - 1];
1340 tprintf(T("Index: %d\n"), image_info->index);
1341 tprintf(T("Name: %"TS"\n"), image_info->name);
1343 /* Always print the Description: part even if there is no
1345 if (image_info->description)
1346 desc = image_info->description;
1349 tprintf(T("Description: %"TS"\n"), desc);
1351 if (image_info->display_name) {
1352 tprintf(T("Display Name: %"TS"\n"),
1353 image_info->display_name);
1356 if (image_info->display_description) {
1357 tprintf(T("Display Description: %"TS"\n"),
1358 image_info->display_description);
1361 tprintf(T("Directory Count: %"PRIu64"\n"), image_info->dir_count);
1362 tprintf(T("File Count: %"PRIu64"\n"), image_info->file_count);
1363 tprintf(T("Total Bytes: %"PRIu64"\n"), image_info->total_bytes);
1364 tprintf(T("Hard Link Bytes: %"PRIu64"\n"), image_info->hard_link_bytes);
1366 wim_timestamp_to_str(image_info->creation_time, buf, sizeof(buf));
1367 tprintf(T("Creation Time: %"TS"\n"), buf);
1369 wim_timestamp_to_str(image_info->last_modification_time, buf, sizeof(buf));
1370 tprintf(T("Last Modification Time: %"TS"\n"), buf);
1371 if (image_info->windows_info_exists)
1372 print_windows_info(&image_info->windows_info);
1373 if (image_info->flags)
1374 tprintf(T("Flags: %"TS"\n"), image_info->flags);
1375 tprintf(T("WIMBoot compatible: %"TS"\n"),
1376 image_info->wimboot ? T("yes") : T("no"));
1381 libxml_global_init(void)
1384 xmlInitCharEncodingHandlers();
1388 libxml_global_cleanup(void)
1391 xmlCleanupCharEncodingHandlers();
1394 /* Reads the XML data from a WIM file. */
1396 read_wim_xml_data(WIMStruct *wim)
1405 ret = wimlib_get_xml_data(wim, &buf, &bufsize);
1410 doc = xmlReadMemory((const char *)xml_data, bufsize,
1411 NULL, "UTF-16LE", 0);
1413 ERROR("Failed to parse XML data");
1414 ret = WIMLIB_ERR_XML;
1415 goto out_free_xml_data;
1418 root = xmlDocGetRootElement(doc);
1419 if (!root || !node_is_element(root) || !node_name_is(root, "WIM")) {
1420 ERROR("WIM XML data is invalid");
1421 ret = WIMLIB_ERR_XML;
1425 ret = xml_read_wim_info(root, &wim->wim_info);
1434 /* Prepares an in-memory buffer containing the UTF-16LE XML data for a WIM file.
1436 * total_bytes is the number to write in <TOTALBYTES>, or
1437 * WIM_TOTALBYTES_USE_EXISTING to use the existing value in memory, or
1438 * WIM_TOTALBYTES_OMIT to omit <TOTALBYTES> entirely.
1441 prepare_wim_xml_data(WIMStruct *wim, int image, u64 total_bytes,
1442 u8 **xml_data_ret, size_t *xml_len_ret)
1444 xmlCharEncodingHandler *encoding_handler;
1446 xmlOutputBuffer *outbuf;
1447 xmlTextWriter *writer;
1450 const xmlChar *content;
1455 /* Open an xmlTextWriter that writes to an in-memory buffer using
1456 * UTF-16LE encoding. */
1458 encoding_handler = xmlGetCharEncodingHandler(XML_CHAR_ENCODING_UTF16LE);
1459 if (!encoding_handler) {
1460 ERROR("Failed to get XML character encoding handler for UTF-16LE");
1461 ret = WIMLIB_ERR_LIBXML_UTF16_HANDLER_NOT_AVAILABLE;
1465 buf = xmlBufferCreate();
1467 ERROR("Failed to create xmlBuffer");
1468 ret = WIMLIB_ERR_NOMEM;
1472 outbuf = xmlOutputBufferCreateBuffer(buf, encoding_handler);
1474 ERROR("Failed to allocate xmlOutputBuffer");
1475 ret = WIMLIB_ERR_NOMEM;
1476 goto out_buffer_free;
1479 writer = xmlNewTextWriter(outbuf);
1481 ERROR("Failed to allocate xmlTextWriter");
1482 ret = WIMLIB_ERR_NOMEM;
1483 goto out_output_buffer_close;
1486 /* Write the XML document. */
1488 ret = xmlTextWriterStartElement(writer, "WIM");
1490 goto out_write_error;
1492 /* The contents of the <TOTALBYTES> element in the XML data, under the
1493 * <WIM> element (not the <IMAGE> element), is for non-split WIMs the
1494 * size of the WIM file excluding the XML data and integrity table.
1495 * For split WIMs, <TOTALBYTES> takes into account the entire WIM, not
1496 * just the current part. */
1497 if (total_bytes != WIM_TOTALBYTES_OMIT) {
1498 if (total_bytes == WIM_TOTALBYTES_USE_EXISTING) {
1500 total_bytes = wim->wim_info->total_bytes;
1504 ret = xmlTextWriterWriteFormatElement(writer, "TOTALBYTES",
1505 "%"PRIu64, total_bytes);
1507 goto out_write_error;
1510 if (image == WIMLIB_ALL_IMAGES) {
1512 last = wim->hdr.image_count;
1518 for (int i = first; i <= last; i++) {
1519 ret = xml_write_image_info(writer, &wim->wim_info->images[i - 1]);
1522 goto out_write_error;
1523 goto out_free_text_writer;
1527 ret = xmlTextWriterEndElement(writer);
1529 goto out_write_error;
1531 ret = xmlTextWriterEndDocument(writer);
1533 goto out_write_error;
1535 ret = xmlTextWriterFlush(writer);
1537 goto out_write_error;
1539 /* Retrieve the buffer into which the document was written. */
1541 content = xmlBufferContent(buf);
1542 len = xmlBufferLength(buf);
1544 /* Copy the data into a new buffer, and prefix it with the UTF-16LE BOM
1545 * (byte order mark), which is required by MS's software to understand
1549 xml_data = MALLOC(xml_len);
1551 ret = WIMLIB_ERR_NOMEM;
1552 goto out_free_text_writer;
1556 memcpy(&xml_data[2], content, len);
1558 /* Clean up libxml objects and return success. */
1559 *xml_data_ret = xml_data;
1560 *xml_len_ret = xml_len;
1562 out_free_text_writer:
1563 /* xmlFreeTextWriter will free the attached xmlOutputBuffer. */
1564 xmlFreeTextWriter(writer);
1565 goto out_buffer_free;
1566 out_output_buffer_close:
1567 xmlOutputBufferClose(outbuf);
1571 DEBUG("ret=%d", ret);
1575 ERROR("Error writing XML data");
1576 ret = WIMLIB_ERR_WRITE;
1577 goto out_free_text_writer;
1580 /* Writes the XML data to a WIM file. */
1582 write_wim_xml_data(WIMStruct *wim, int image, u64 total_bytes,
1583 struct wim_reshdr *out_reshdr,
1584 int write_resource_flags)
1590 DEBUG("Writing WIM XML data (image=%d, offset=%"PRIu64")",
1591 image, wim->out_fd.offset);
1593 ret = prepare_wim_xml_data(wim, image, total_bytes,
1594 &xml_data, &xml_len);
1598 /* Write the XML data uncompressed. Although wimlib can handle
1599 * compressed XML data, MS software cannot. */
1600 ret = write_wim_resource_from_buffer(xml_data,
1602 WIM_RESHDR_FLAG_METADATA,
1604 WIMLIB_COMPRESSION_TYPE_NONE,
1608 write_resource_flags);
1610 DEBUG("ret=%d", ret);
1614 /* API function documented in wimlib.h */
1615 WIMLIBAPI const tchar *
1616 wimlib_get_image_name(const WIMStruct *wim, int image)
1618 if (image < 1 || image > wim->hdr.image_count)
1620 return wim->wim_info->images[image - 1].name;
1623 /* API function documented in wimlib.h */
1624 WIMLIBAPI const tchar *
1625 wimlib_get_image_description(const WIMStruct *wim, int image)
1627 if (image < 1 || image > wim->hdr.image_count)
1629 return wim->wim_info->images[image - 1].description;
1632 /* API function documented in wimlib.h */
1634 wimlib_image_name_in_use(const WIMStruct *wim, const tchar *name)
1636 if (!name || !*name)
1638 for (int i = 1; i <= wim->hdr.image_count; i++)
1639 if (!tstrcmp(wim->wim_info->images[i - 1].name, name))
1645 /* API function documented in wimlib.h */
1647 wimlib_get_xml_data(WIMStruct *wim, void **buf_ret, size_t *bufsize_ret)
1649 const struct wim_reshdr *xml_reshdr;
1651 if (wim->filename == NULL && filedes_is_seekable(&wim->in_fd))
1652 return WIMLIB_ERR_INVALID_PARAM;
1654 if (buf_ret == NULL || bufsize_ret == NULL)
1655 return WIMLIB_ERR_INVALID_PARAM;
1657 xml_reshdr = &wim->hdr.xml_data_reshdr;
1659 DEBUG("Reading XML data.");
1660 *bufsize_ret = xml_reshdr->uncompressed_size;
1661 return wim_reshdr_to_data(xml_reshdr, wim, buf_ret);
1665 wimlib_extract_xml_data(WIMStruct *wim, FILE *fp)
1671 ret = wimlib_get_xml_data(wim, &buf, &bufsize);
1675 if (fwrite(buf, 1, bufsize, fp) != bufsize) {
1676 ERROR_WITH_ERRNO("Failed to extract XML data");
1677 ret = WIMLIB_ERR_WRITE;
1683 /* API function documented in wimlib.h */
1685 wimlib_set_image_name(WIMStruct *wim, int image, const tchar *name)
1691 DEBUG("Setting the name of image %d to %"TS, image, name);
1693 ret = can_modify_wim(wim);
1700 if (image < 1 || image > wim->hdr.image_count) {
1701 ERROR("%d is not a valid image", image);
1702 return WIMLIB_ERR_INVALID_IMAGE;
1705 for (i = 1; i <= wim->hdr.image_count; i++) {
1708 if (!tstrcmp(wim->wim_info->images[i - 1].name, name)) {
1709 ERROR("The name \"%"TS"\" is already in use in the WIM!",
1711 return WIMLIB_ERR_IMAGE_NAME_COLLISION;
1717 return WIMLIB_ERR_NOMEM;
1719 FREE(wim->wim_info->images[image - 1].name);
1720 wim->wim_info->images[image - 1].name = p;
1725 do_set_image_info_str(WIMStruct *wim, int image, const tchar *tstr,
1729 tchar **dest_tstr_p;
1732 ret = can_modify_wim(wim);
1736 if (image < 1 || image > wim->hdr.image_count) {
1737 ERROR("%d is not a valid image", image);
1738 return WIMLIB_ERR_INVALID_IMAGE;
1741 tstr_copy = TSTRDUP(tstr);
1743 return WIMLIB_ERR_NOMEM;
1747 dest_tstr_p = (tchar**)((void*)&wim->wim_info->images[image - 1] + offset);
1750 *dest_tstr_p = tstr_copy;
1754 /* API function documented in wimlib.h */
1756 wimlib_set_image_descripton(WIMStruct *wim, int image,
1757 const tchar *description)
1759 return do_set_image_info_str(wim, image, description,
1760 offsetof(struct image_info, description));
1763 /* API function documented in wimlib.h */
1765 wimlib_set_image_flags(WIMStruct *wim, int image, const tchar *flags)
1767 return do_set_image_info_str(wim, image, flags,
1768 offsetof(struct image_info, flags));