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)
279 for_node_child(string_node, child)
280 if (node_is_text(child) && child->content)
281 return utf8_to_tstr_simple(child->content, tstr_ret);
285 /* Returns the timestamp from a time node. It has child elements <HIGHPART> and
286 * <LOWPART> that are then used to construct a 64-bit timestamp. */
288 node_get_timestamp(const xmlNode *time_node)
293 for_node_child(time_node, child) {
294 if (!node_is_element(child))
296 if (node_name_is(child, "HIGHPART"))
297 high_part = node_get_hex_u64(child);
298 else if (node_name_is(child, "LOWPART"))
299 low_part = node_get_hex_u64(child);
301 return (u64)low_part | ((u64)high_part << 32);
304 /* Used to sort an array of struct image_infos by their image indices. */
306 sort_by_index(const void *p1, const void *p2)
308 int index_1 = ((const struct image_info*)p1)->index;
309 int index_2 = ((const struct image_info*)p2)->index;
310 if (index_1 < index_2)
312 else if (index_1 > index_2)
319 /* Frees memory allocated inside a struct windows_info structure. */
321 destroy_windows_info(struct windows_info *windows_info)
323 FREE(windows_info->product_name);
324 FREE(windows_info->edition_id);
325 FREE(windows_info->installation_type);
326 FREE(windows_info->hal);
327 FREE(windows_info->product_type);
328 FREE(windows_info->product_suite);
329 FREE(windows_info->pkeyconfigversion);
330 for (size_t i = 0; i < windows_info->num_languages; i++)
331 FREE(windows_info->languages[i]);
332 FREE(windows_info->languages);
333 FREE(windows_info->default_language);
334 FREE(windows_info->system_root);
337 /* Frees memory allocated inside a struct image_info structure. */
339 destroy_image_info(struct image_info *image_info)
341 FREE(image_info->name);
342 FREE(image_info->description);
343 FREE(image_info->flags);
344 FREE(image_info->display_name);
345 FREE(image_info->display_description);
346 destroy_windows_info(&image_info->windows_info);
347 memset(image_info, 0, sizeof(struct image_info));
351 free_wim_info(struct wim_info *info)
355 for (int i = 0; i < info->num_images; i++)
356 destroy_image_info(&info->images[i]);
363 /* Reads the information from a <VERSION> element inside the <WINDOWS> element.
366 xml_read_windows_version(const xmlNode *version_node,
367 struct windows_version* windows_version)
370 for_node_child(version_node, child) {
371 if (!node_is_element(child))
373 if (node_name_is(child, "MAJOR"))
374 windows_version->major = node_get_u64(child);
375 else if (node_name_is(child, "MINOR"))
376 windows_version->minor = node_get_u64(child);
377 else if (node_name_is(child, "BUILD"))
378 windows_version->build = node_get_u64(child);
379 else if (node_name_is(child, "SPBUILD"))
380 windows_version->sp_build = node_get_u64(child);
381 else if (node_name_is(child, "SPLEVEL"))
382 windows_version->sp_level = node_get_u64(child);
386 /* Reads the information from a <LANGUAGE> element inside a <WINDOWS> element.
389 xml_read_languages(const xmlNode *languages_node,
390 tchar ***languages_ret,
391 size_t *num_languages_ret,
392 tchar **default_language_ret)
395 size_t num_languages = 0;
399 for_node_child(languages_node, child)
400 if (node_is_element(child) && node_name_is(child, "LANGUAGE"))
403 languages = CALLOC(num_languages, sizeof(languages[0]));
405 return WIMLIB_ERR_NOMEM;
407 *languages_ret = languages;
408 *num_languages_ret = num_languages;
411 for_node_child(languages_node, child) {
412 if (!node_is_element(child))
414 if (node_name_is(child, "LANGUAGE"))
415 ret = node_get_string(child, languages++);
416 else if (node_name_is(child, "DEFAULT"))
417 ret = node_get_string(child, default_language_ret);
424 /* Reads the information from a <WINDOWS> element inside an <IMAGE> element. */
426 xml_read_windows_info(const xmlNode *windows_node,
427 struct windows_info *windows_info)
432 for_node_child(windows_node, child) {
433 if (!node_is_element(child))
435 if (node_name_is(child, "ARCH")) {
436 windows_info->arch = node_get_u64(child);
437 } else if (node_name_is(child, "PRODUCTNAME")) {
438 ret = node_get_string(child,
439 &windows_info->product_name);
440 } else if (node_name_is(child, "EDITIONID")) {
441 ret = node_get_string(child,
442 &windows_info->edition_id);
443 } else if (node_name_is(child, "INSTALLATIONTYPE")) {
444 ret = node_get_string(child,
445 &windows_info->installation_type);
446 } else if (node_name_is(child, "PRODUCTTYPE")) {
447 ret = node_get_string(child,
448 &windows_info->product_type);
449 } else if (node_name_is(child, "PRODUCTSUITE")) {
450 ret = node_get_string(child,
451 &windows_info->product_suite);
452 } else if (node_name_is(child, "LANGUAGES")) {
453 ret = xml_read_languages(child,
454 &windows_info->languages,
455 &windows_info->num_languages,
456 &windows_info->default_language);
457 } else if (node_name_is(child, "VERSION")) {
458 xml_read_windows_version(child,
459 &windows_info->windows_version);
460 windows_info->windows_version_exists = true;
461 } else if (node_name_is(child, "SYSTEMROOT")) {
462 ret = node_get_string(child, &windows_info->system_root);
463 } else if (node_name_is(child, "HAL")) {
464 ret = node_get_string(child, &windows_info->hal);
465 } else if (node_name_is(child, "SERVICINGDATA")) {
468 for_node_child(child, grandchild) {
469 if (node_is_element(grandchild) &&
470 node_name_is(grandchild, "PKEYCONFIGVERSION"))
472 ret = node_get_string(grandchild,
473 &windows_info->pkeyconfigversion);
484 /* Reads the information from an <IMAGE> element. */
486 xml_read_image_info(xmlNode *image_node, struct image_info *image_info)
492 index_prop = xmlGetProp(image_node, "INDEX");
494 image_info->index = atoi(index_prop);
497 image_info->index = 1;
501 for_node_child(image_node, child) {
502 if (!node_is_element(child))
504 if (node_name_is(child, "DIRCOUNT"))
505 image_info->dir_count = node_get_u64(child);
506 else if (node_name_is(child, "FILECOUNT"))
507 image_info->file_count = node_get_u64(child);
508 else if (node_name_is(child, "TOTALBYTES"))
509 image_info->total_bytes = node_get_u64(child);
510 else if (node_name_is(child, "HARDLINKBYTES"))
511 image_info->hard_link_bytes = node_get_u64(child);
512 else if (node_name_is(child, "CREATIONTIME"))
513 image_info->creation_time = node_get_timestamp(child);
514 else if (node_name_is(child, "LASTMODIFICATIONTIME"))
515 image_info->last_modification_time = node_get_timestamp(child);
516 else if (node_name_is(child, "WINDOWS")) {
517 ret = xml_read_windows_info(child,
518 &image_info->windows_info);
519 image_info->windows_info_exists = true;
520 } else if (node_name_is(child, "NAME")) {
521 ret = node_get_string(child, &image_info->name);
522 } else if (node_name_is(child, "DESCRIPTION")) {
523 ret = node_get_string(child, &image_info->description);
524 } else if (node_name_is(child, "FLAGS")) {
525 ret = node_get_string(child, &image_info->flags);
526 } else if (node_name_is(child, "DISPLAYNAME")) {
527 ret = node_get_string(child, &image_info->display_name);
528 } else if (node_name_is(child, "DISPLAYDESCRIPTION")) {
529 ret = node_get_string(child, &image_info->display_description);
530 } else if (node_name_is(child, "WIMBOOT")) {
531 if (node_get_u64(child) == 1) {
532 image_info->wimboot = true;
538 if (!image_info->name) {
540 empty_name = MALLOC(sizeof(tchar));
542 return WIMLIB_ERR_NOMEM;
543 *empty_name = T('\0');
544 image_info->name = empty_name;
549 /* Reads the information from a <WIM> element, which should be the root element
550 * of the XML tree. */
552 xml_read_wim_info(const xmlNode *wim_node, struct wim_info **wim_info_ret)
554 struct wim_info *wim_info;
560 wim_info = CALLOC(1, sizeof(struct wim_info));
562 return WIMLIB_ERR_NOMEM;
564 /* Count how many images there are. */
566 for_node_child(wim_node, child) {
567 if (node_is_element(child) && node_name_is(child, "IMAGE")) {
568 if (unlikely(num_images == MAX_IMAGES)) {
569 ret = WIMLIB_ERR_IMAGE_COUNT;
576 if (num_images > 0) {
577 /* Allocate the array of struct image_infos and fill them in. */
578 wim_info->images = CALLOC(num_images, sizeof(wim_info->images[0]));
579 if (!wim_info->images) {
580 ret = WIMLIB_ERR_NOMEM;
583 wim_info->num_images = num_images;
585 for_node_child(wim_node, child) {
586 if (!node_is_element(child))
588 if (node_name_is(child, "IMAGE")) {
589 ret = xml_read_image_info(child,
590 &wim_info->images[i]);
594 } else if (node_name_is(child, "TOTALBYTES")) {
595 wim_info->total_bytes = node_get_u64(child);
596 } else if (node_name_is(child, "ESD")) {
598 for_node_child(child, esdchild) {
599 if (node_is_element(esdchild) &&
600 node_name_is(esdchild, "ENCRYPTED"))
602 ret = WIMLIB_ERR_WIM_IS_ENCRYPTED;
609 /* Sort the array of image info by image index. */
610 qsort(wim_info->images, num_images,
611 sizeof(struct image_info), sort_by_index);
613 /* Make sure the image indices make sense */
614 for (i = 0; i < num_images; i++) {
615 if (wim_info->images[i].index != i + 1) {
616 ERROR("WIM images are not indexed [1...%d] "
617 "in XML data as expected",
619 ret = WIMLIB_ERR_IMAGE_COUNT;
625 *wim_info_ret = wim_info;
628 free_wim_info(wim_info);
632 /* Prints the information contained in a `struct windows_info'. */
634 print_windows_info(const struct windows_info *windows_info)
636 const struct windows_version *windows_version;
638 tprintf(T("Architecture: %"TS"\n"),
639 get_arch(windows_info->arch));
641 if (windows_info->product_name) {
642 tprintf(T("Product Name: %"TS"\n"),
643 windows_info->product_name);
646 if (windows_info->edition_id) {
647 tprintf(T("Edition ID: %"TS"\n"),
648 windows_info->edition_id);
651 if (windows_info->installation_type) {
652 tprintf(T("Installation Type: %"TS"\n"),
653 windows_info->installation_type);
656 if (windows_info->hal) {
657 tprintf(T("HAL: %"TS"\n"),
661 if (windows_info->product_type) {
662 tprintf(T("Product Type: %"TS"\n"),
663 windows_info->product_type);
666 if (windows_info->product_suite) {
667 tprintf(T("Product Suite: %"TS"\n"),
668 windows_info->product_suite);
671 tprintf(T("Languages: "));
672 for (size_t i = 0; i < windows_info->num_languages; i++) {
674 tfputs(windows_info->languages[i], stdout);
678 if (windows_info->default_language) {
679 tprintf(T("Default Language: %"TS"\n"),
680 windows_info->default_language);
682 if (windows_info->system_root) {
683 tprintf(T("System Root: %"TS"\n"),
684 windows_info->system_root);
687 if (windows_info->windows_version_exists) {
688 windows_version = &windows_info->windows_version;
689 tprintf(T("Major Version: %"PRIu64"\n"),
690 windows_version->major);
691 tprintf(T("Minor Version: %"PRIu64"\n"),
692 windows_version->minor);
693 tprintf(T("Build: %"PRIu64"\n"),
694 windows_version->build);
695 tprintf(T("Service Pack Build: %"PRIu64"\n"),
696 windows_version->sp_build);
697 tprintf(T("Service Pack Level: %"PRIu64"\n"),
698 windows_version->sp_level);
703 xml_write_string(xmlTextWriter *writer, const char *name,
708 int rc = tstr_to_utf8_simple(tstr, &utf8_str);
711 rc = xmlTextWriterWriteElement(writer, name, utf8_str);
720 xml_write_strings_from_specs(xmlTextWriter *writer,
721 const void *struct_with_strings,
722 const struct xml_string_spec specs[],
725 for (size_t i = 0; i < num_specs; i++) {
726 int rc = xml_write_string(writer, specs[i].name,
727 *(const tchar * const *)
728 (struct_with_strings + specs[i].offset));
736 dup_strings_from_specs(const void *old_struct_with_strings,
737 void *new_struct_with_strings,
738 const struct xml_string_spec specs[],
741 for (size_t i = 0; i < num_specs; i++) {
742 const tchar *old_str = *(const tchar * const *)
743 ((const void*)old_struct_with_strings + specs[i].offset);
744 tchar **new_str_p = (tchar **)((void*)new_struct_with_strings + specs[i].offset);
746 *new_str_p = TSTRDUP(old_str);
748 return WIMLIB_ERR_NOMEM;
754 /* Writes the information contained in a `struct windows_version' to the XML
755 * document being written. This is the <VERSION> element inside the <WINDOWS>
758 xml_write_windows_version(xmlTextWriter *writer,
759 const struct windows_version *version)
762 rc = xmlTextWriterStartElement(writer, "VERSION");
766 rc = xmlTextWriterWriteFormatElement(writer, "MAJOR", "%"PRIu64,
771 rc = xmlTextWriterWriteFormatElement(writer, "MINOR", "%"PRIu64,
776 rc = xmlTextWriterWriteFormatElement(writer, "BUILD", "%"PRIu64,
781 rc = xmlTextWriterWriteFormatElement(writer, "SPBUILD", "%"PRIu64,
786 rc = xmlTextWriterWriteFormatElement(writer, "SPLEVEL", "%"PRIu64,
791 rc = xmlTextWriterEndElement(writer); /* </VERSION> */
798 /* Writes the information contained in a `struct windows_info' to the XML
799 * document being written. This is the <WINDOWS> element. */
801 xml_write_windows_info(xmlTextWriter *writer,
802 const struct windows_info *windows_info)
805 rc = xmlTextWriterStartElement(writer, "WINDOWS");
809 rc = xmlTextWriterWriteFormatElement(writer, "ARCH", "%"PRIu64,
814 rc = xml_write_strings_from_specs(writer,
816 windows_info_xml_string_specs,
817 ARRAY_LEN(windows_info_xml_string_specs));
821 if (windows_info->num_languages) {
822 rc = xmlTextWriterStartElement(writer, "LANGUAGES");
826 for (size_t i = 0; i < windows_info->num_languages; i++) {
827 rc = xml_write_string(writer, "LANGUAGE",
828 windows_info->languages[i]);
833 rc = xml_write_string(writer, "DEFAULT",
834 windows_info->default_language);
838 rc = xmlTextWriterEndElement(writer); /* </LANGUAGES> */
843 if (windows_info->pkeyconfigversion) {
844 rc = xmlTextWriterStartElement(writer, "SERVICINGDATA");
848 rc = xml_write_string(writer, "PKEYCONFIGVERSION",
849 windows_info->pkeyconfigversion);
853 rc = xmlTextWriterEndElement(writer);
858 if (windows_info->windows_version_exists) {
859 rc = xml_write_windows_version(writer, &windows_info->windows_version);
864 rc = xml_write_string(writer, "SYSTEMROOT", windows_info->system_root);
868 rc = xmlTextWriterEndElement(writer); /* </WINDOWS> */
875 /* Writes a time element to the XML document being constructed in memory. */
877 xml_write_time(xmlTextWriter *writer, const char *element_name, u64 time)
880 rc = xmlTextWriterStartElement(writer, element_name);
884 rc = xmlTextWriterWriteFormatElement(writer, "HIGHPART",
885 "0x%08"PRIX32, (u32)(time >> 32));
889 rc = xmlTextWriterWriteFormatElement(writer, "LOWPART",
890 "0x%08"PRIX32, (u32)time);
894 rc = xmlTextWriterEndElement(writer); /* </@element_name> */
900 /* Writes an <IMAGE> element to the XML document. */
902 xml_write_image_info(xmlTextWriter *writer, const struct image_info *image_info)
905 rc = xmlTextWriterStartElement(writer, "IMAGE");
909 rc = xmlTextWriterWriteFormatAttribute(writer, "INDEX", "%d",
914 rc = xmlTextWriterWriteFormatElement(writer, "DIRCOUNT", "%"PRIu64,
915 image_info->dir_count);
919 rc = xmlTextWriterWriteFormatElement(writer, "FILECOUNT", "%"PRIu64,
920 image_info->file_count);
924 rc = xmlTextWriterWriteFormatElement(writer, "TOTALBYTES", "%"PRIu64,
925 image_info->total_bytes);
929 rc = xmlTextWriterWriteFormatElement(writer, "HARDLINKBYTES", "%"PRIu64,
930 image_info->hard_link_bytes);
934 rc = xml_write_time(writer, "CREATIONTIME", image_info->creation_time);
938 rc = xml_write_time(writer, "LASTMODIFICATIONTIME",
939 image_info->last_modification_time);
943 if (image_info->windows_info_exists) {
944 rc = xml_write_windows_info(writer, &image_info->windows_info);
949 rc = xml_write_strings_from_specs(writer, image_info,
950 image_info_xml_string_specs,
951 ARRAY_LEN(image_info_xml_string_specs));
955 if (image_info->wimboot) {
956 rc = xmlTextWriterWriteFormatElement(writer, "WIMBOOT", "%d", 1);
961 rc = xmlTextWriterEndElement(writer); /* </IMAGE> */
970 /* Makes space for another image in the XML information and return a pointer to
972 static struct image_info *
973 add_image_info_struct(struct wim_info *wim_info)
975 struct image_info *images;
977 images = CALLOC(wim_info->num_images + 1, sizeof(struct image_info));
980 memcpy(images, wim_info->images,
981 wim_info->num_images * sizeof(struct image_info));
982 FREE(wim_info->images);
983 wim_info->images = images;
984 wim_info->num_images++;
985 return &images[wim_info->num_images - 1];
989 clone_windows_info(const struct windows_info *old, struct windows_info *new)
993 new->arch = old->arch;
995 ret = dup_strings_from_specs(old, new, windows_info_xml_string_specs,
996 ARRAY_LEN(windows_info_xml_string_specs));
1000 if (old->pkeyconfigversion) {
1001 new->pkeyconfigversion = TSTRDUP(old->pkeyconfigversion);
1002 if (new->pkeyconfigversion == NULL)
1003 return WIMLIB_ERR_NOMEM;
1006 if (old->languages) {
1007 new->languages = CALLOC(old->num_languages, sizeof(new->languages[0]));
1008 if (!new->languages)
1009 return WIMLIB_ERR_NOMEM;
1010 new->num_languages = old->num_languages;
1011 for (size_t i = 0; i < new->num_languages; i++) {
1012 if (!old->languages[i])
1014 new->languages[i] = TSTRDUP(old->languages[i]);
1015 if (!new->languages[i])
1016 return WIMLIB_ERR_NOMEM;
1019 if (old->default_language &&
1020 !(new->default_language = TSTRDUP(old->default_language)))
1021 return WIMLIB_ERR_NOMEM;
1022 if (old->system_root && !(new->system_root = TSTRDUP(old->system_root)))
1023 return WIMLIB_ERR_NOMEM;
1024 if (old->windows_version_exists) {
1025 new->windows_version_exists = true;
1026 memcpy(&new->windows_version, &old->windows_version,
1027 sizeof(old->windows_version));
1033 clone_image_info(const struct image_info *old, struct image_info *new)
1037 new->dir_count = old->dir_count;
1038 new->file_count = old->file_count;
1039 new->total_bytes = old->total_bytes;
1040 new->hard_link_bytes = old->hard_link_bytes;
1041 new->creation_time = old->creation_time;
1042 new->last_modification_time = old->last_modification_time;
1044 ret = dup_strings_from_specs(old, new,
1045 image_info_xml_string_specs,
1046 ARRAY_LEN(image_info_xml_string_specs));
1050 if (old->windows_info_exists) {
1051 new->windows_info_exists = true;
1052 ret = clone_windows_info(&old->windows_info,
1053 &new->windows_info);
1057 new->wimboot = old->wimboot;
1061 /* Copies the XML information for an image between WIM files.
1063 * @dest_image_name and @dest_image_description are ignored if they are NULL;
1064 * otherwise, they are used to override the image name and/or image description
1065 * from the XML data in the source WIM file.
1067 * On failure, WIMLIB_ERR_NOMEM is returned and no changes are made. Otherwise,
1068 * 0 is returned and the WIM information at *new_wim_info_p is modified.
1071 xml_export_image(const struct wim_info *old_wim_info,
1073 struct wim_info **new_wim_info_p,
1074 const tchar *dest_image_name,
1075 const tchar *dest_image_description)
1077 struct wim_info *new_wim_info;
1078 struct image_info *image_info;
1081 DEBUG("Copying XML data between WIM files for source image %d.", image);
1083 wimlib_assert(old_wim_info != NULL);
1084 wimlib_assert(image >= 1 && image <= old_wim_info->num_images);
1086 if (*new_wim_info_p) {
1087 new_wim_info = *new_wim_info_p;
1089 new_wim_info = CALLOC(1, sizeof(struct wim_info));
1094 image_info = add_image_info_struct(new_wim_info);
1098 ret = clone_image_info(&old_wim_info->images[image - 1], image_info);
1100 goto err_destroy_image_info;
1102 image_info->index = new_wim_info->num_images;
1104 if (dest_image_name) {
1105 FREE(image_info->name);
1106 image_info->name = TSTRDUP(dest_image_name);
1107 if (!image_info->name)
1108 goto err_destroy_image_info;
1110 if (dest_image_description) {
1111 FREE(image_info->description);
1112 image_info->description = TSTRDUP(dest_image_description);
1113 if (!image_info->description)
1114 goto err_destroy_image_info;
1116 *new_wim_info_p = new_wim_info;
1118 err_destroy_image_info:
1119 destroy_image_info(image_info);
1121 if (new_wim_info != *new_wim_info_p)
1122 free_wim_info(new_wim_info);
1123 return WIMLIB_ERR_NOMEM;
1126 /* Removes an image from the XML information. */
1128 xml_delete_image(struct wim_info **wim_info_p, int image)
1130 struct wim_info *wim_info;
1132 wim_info = *wim_info_p;
1133 wimlib_assert(image >= 1 && image <= wim_info->num_images);
1134 DEBUG("Deleting image %d from the XML data.", image);
1136 destroy_image_info(&wim_info->images[image - 1]);
1138 memmove(&wim_info->images[image - 1],
1139 &wim_info->images[image],
1140 (wim_info->num_images - image) * sizeof(struct image_info));
1142 if (--wim_info->num_images == 0) {
1143 free_wim_info(wim_info);
1146 for (int i = image - 1; i < wim_info->num_images; i++)
1147 wim_info->images[i].index--;
1152 xml_get_max_image_name_len(const WIMStruct *wim)
1155 for (u32 i = 0; i < wim->hdr.image_count; i++)
1156 max_len = max(max_len, tstrlen(wim->wim_info->images[i].name));
1161 xml_set_memory_allocator(void *(*malloc_func)(size_t),
1162 void (*free_func)(void *),
1163 void *(*realloc_func)(void *, size_t))
1165 xmlMemSetup(free_func, malloc_func, realloc_func, STRDUP);
1169 calculate_dentry_statistics(struct wim_dentry *dentry, void *arg)
1171 struct image_info *info = arg;
1172 const struct wim_inode *inode = dentry->d_inode;
1174 /* Update directory count and file count.
1176 * Each dentry counts as either a file or a directory, but not both.
1177 * The root directory is an exception: it is not counted at all.
1179 * Symbolic links and junction points (and presumably other reparse
1180 * points) count as regular files. This is despite the fact that
1181 * junction points have FILE_ATTRIBUTE_DIRECTORY set.
1184 if (!dentry_is_root(dentry)) {
1185 if (inode_is_directory(inode))
1192 * Update total bytes and hard link bytes.
1194 * We try to act the same as the MS implementation, even though there
1195 * are some inconsistencies/bugs in the way it operates.
1197 * If there are no alternate data streams in the image, the "total
1198 * bytes" is the sum of the size of the un-named data stream of each
1199 * inode times the link count of that inode. In other words, it would
1200 * be the total number of bytes of regular files you would have if you
1201 * extracted the full image without any hard-links. The "hard link
1202 * bytes" is equal to the "total bytes" minus the size of the un-named
1203 * data stream of each inode. In other words, the "hard link bytes"
1204 * counts the size of the un-named data stream for all the links to each
1205 * inode except the first one.
1207 * Reparse points and directories don't seem to be counted in either the
1208 * total bytes or the hard link bytes.
1210 * And now we get to the most confusing part, the alternate data
1211 * streams. They are not counted in the "total bytes". However, if the
1212 * link count of an inode with alternate data streams is 2 or greater,
1213 * the size of all the alternate data streams is included in the "hard
1214 * link bytes", and this size is multiplied by the link count (NOT one
1215 * less than the link count).
1217 if (!(inode->i_attributes & (FILE_ATTRIBUTE_DIRECTORY |
1218 FILE_ATTRIBUTE_REPARSE_POINT)))
1220 struct wim_lookup_table_entry *lte;
1222 lte = inode_unnamed_lte(inode, info->lookup_table);
1224 info->total_bytes += lte->size;
1225 if (!dentry_is_first_in_inode(dentry))
1226 info->hard_link_bytes += lte->size;
1229 if (inode->i_nlink >= 2 && dentry_is_first_in_inode(dentry)) {
1230 for (unsigned i = 0; i < inode->i_num_ads; i++) {
1231 if (inode->i_ads_entries[i].stream_name_nbytes) {
1232 lte = inode_stream_lte(inode, i + 1, info->lookup_table);
1234 info->hard_link_bytes += inode->i_nlink *
1245 * Calculate what to put in the <FILECOUNT>, <DIRCOUNT>, <TOTALBYTES>, and
1246 * <HARDLINKBYTES> elements of each <IMAGE>.
1249 xml_update_image_info(WIMStruct *wim, int image)
1251 struct image_info *image_info;
1253 DEBUG("Updating the image info for image %d", image);
1255 image_info = &wim->wim_info->images[image - 1];
1257 image_info->file_count = 0;
1258 image_info->dir_count = 0;
1259 image_info->total_bytes = 0;
1260 image_info->hard_link_bytes = 0;
1261 image_info->lookup_table = wim->lookup_table;
1263 for_dentry_in_tree(wim->image_metadata[image - 1]->root_dentry,
1264 calculate_dentry_statistics,
1266 image_info->last_modification_time = get_wim_timestamp();
1269 /* Adds an image to the XML information. */
1271 xml_add_image(WIMStruct *wim, const tchar *name)
1273 struct wim_info *wim_info;
1274 struct image_info *image_info;
1276 wimlib_assert(name != NULL);
1278 /* If this is the first image, allocate the struct wim_info. Otherwise
1279 * use the existing struct wim_info. */
1280 if (wim->wim_info) {
1281 wim_info = wim->wim_info;
1283 wim_info = CALLOC(1, sizeof(struct wim_info));
1285 return WIMLIB_ERR_NOMEM;
1288 image_info = add_image_info_struct(wim_info);
1290 goto out_free_wim_info;
1292 if (!(image_info->name = TSTRDUP(name)))
1293 goto out_destroy_image_info;
1295 wim->wim_info = wim_info;
1296 image_info->index = wim_info->num_images;
1297 image_info->creation_time = get_wim_timestamp();
1298 xml_update_image_info(wim, image_info->index);
1301 out_destroy_image_info:
1302 destroy_image_info(image_info);
1303 wim_info->num_images--;
1305 if (wim_info != wim->wim_info)
1307 return WIMLIB_ERR_NOMEM;
1310 /* Prints information about the specified image from struct wim_info structure.
1313 print_image_info(const struct wim_info *wim_info, int image)
1315 const struct image_info *image_info;
1319 wimlib_assert(image >= 1 && image <= wim_info->num_images);
1321 image_info = &wim_info->images[image - 1];
1323 tprintf(T("Index: %d\n"), image_info->index);
1324 tprintf(T("Name: %"TS"\n"), image_info->name);
1326 /* Always print the Description: part even if there is no
1328 if (image_info->description)
1329 desc = image_info->description;
1332 tprintf(T("Description: %"TS"\n"), desc);
1334 if (image_info->display_name) {
1335 tprintf(T("Display Name: %"TS"\n"),
1336 image_info->display_name);
1339 if (image_info->display_description) {
1340 tprintf(T("Display Description: %"TS"\n"),
1341 image_info->display_description);
1344 tprintf(T("Directory Count: %"PRIu64"\n"), image_info->dir_count);
1345 tprintf(T("File Count: %"PRIu64"\n"), image_info->file_count);
1346 tprintf(T("Total Bytes: %"PRIu64"\n"), image_info->total_bytes);
1347 tprintf(T("Hard Link Bytes: %"PRIu64"\n"), image_info->hard_link_bytes);
1349 wim_timestamp_to_str(image_info->creation_time, buf, sizeof(buf));
1350 tprintf(T("Creation Time: %"TS"\n"), buf);
1352 wim_timestamp_to_str(image_info->last_modification_time, buf, sizeof(buf));
1353 tprintf(T("Last Modification Time: %"TS"\n"), buf);
1354 if (image_info->windows_info_exists)
1355 print_windows_info(&image_info->windows_info);
1356 if (image_info->flags)
1357 tprintf(T("Flags: %"TS"\n"), image_info->flags);
1358 tprintf(T("WIMBoot compatible: %"TS"\n"),
1359 image_info->wimboot ? T("yes") : T("no"));
1364 libxml_global_init(void)
1367 xmlInitCharEncodingHandlers();
1371 libxml_global_cleanup(void)
1374 xmlCleanupCharEncodingHandlers();
1377 /* Reads the XML data from a WIM file. */
1379 read_wim_xml_data(WIMStruct *wim)
1388 ret = wimlib_get_xml_data(wim, &buf, &bufsize);
1393 doc = xmlReadMemory((const char *)xml_data, bufsize,
1394 NULL, "UTF-16LE", 0);
1396 ERROR("Failed to parse XML data");
1397 ret = WIMLIB_ERR_XML;
1398 goto out_free_xml_data;
1401 root = xmlDocGetRootElement(doc);
1402 if (!root || !node_is_element(root) || !node_name_is(root, "WIM")) {
1403 ERROR("WIM XML data is invalid");
1404 ret = WIMLIB_ERR_XML;
1408 ret = xml_read_wim_info(root, &wim->wim_info);
1417 /* Prepares an in-memory buffer containing the UTF-16LE XML data for a WIM file.
1419 * total_bytes is the number to write in <TOTALBYTES>, or
1420 * WIM_TOTALBYTES_USE_EXISTING to use the existing value in memory, or
1421 * WIM_TOTALBYTES_OMIT to omit <TOTALBYTES> entirely.
1424 prepare_wim_xml_data(WIMStruct *wim, int image, u64 total_bytes,
1425 u8 **xml_data_ret, size_t *xml_len_ret)
1427 xmlCharEncodingHandler *encoding_handler;
1429 xmlOutputBuffer *outbuf;
1430 xmlTextWriter *writer;
1433 const xmlChar *content;
1438 /* Open an xmlTextWriter that writes to an in-memory buffer using
1439 * UTF-16LE encoding. */
1441 encoding_handler = xmlGetCharEncodingHandler(XML_CHAR_ENCODING_UTF16LE);
1442 if (!encoding_handler) {
1443 ERROR("Failed to get XML character encoding handler for UTF-16LE");
1444 ret = WIMLIB_ERR_LIBXML_UTF16_HANDLER_NOT_AVAILABLE;
1448 buf = xmlBufferCreate();
1450 ERROR("Failed to create xmlBuffer");
1451 ret = WIMLIB_ERR_NOMEM;
1455 outbuf = xmlOutputBufferCreateBuffer(buf, encoding_handler);
1457 ERROR("Failed to allocate xmlOutputBuffer");
1458 ret = WIMLIB_ERR_NOMEM;
1459 goto out_buffer_free;
1462 writer = xmlNewTextWriter(outbuf);
1464 ERROR("Failed to allocate xmlTextWriter");
1465 ret = WIMLIB_ERR_NOMEM;
1466 goto out_output_buffer_close;
1469 /* Write the XML document. */
1471 ret = xmlTextWriterStartElement(writer, "WIM");
1473 goto out_write_error;
1475 /* The contents of the <TOTALBYTES> element in the XML data, under the
1476 * <WIM> element (not the <IMAGE> element), is for non-split WIMs the
1477 * size of the WIM file excluding the XML data and integrity table.
1478 * For split WIMs, <TOTALBYTES> takes into account the entire WIM, not
1479 * just the current part. */
1480 if (total_bytes != WIM_TOTALBYTES_OMIT) {
1481 if (total_bytes == WIM_TOTALBYTES_USE_EXISTING) {
1483 total_bytes = wim->wim_info->total_bytes;
1487 ret = xmlTextWriterWriteFormatElement(writer, "TOTALBYTES",
1488 "%"PRIu64, total_bytes);
1490 goto out_write_error;
1493 if (image == WIMLIB_ALL_IMAGES) {
1495 last = wim->hdr.image_count;
1501 for (int i = first; i <= last; i++) {
1502 ret = xml_write_image_info(writer, &wim->wim_info->images[i - 1]);
1505 goto out_write_error;
1506 goto out_free_text_writer;
1510 ret = xmlTextWriterEndElement(writer);
1512 goto out_write_error;
1514 ret = xmlTextWriterEndDocument(writer);
1516 goto out_write_error;
1518 ret = xmlTextWriterFlush(writer);
1520 goto out_write_error;
1522 /* Retrieve the buffer into which the document was written. */
1524 content = xmlBufferContent(buf);
1525 len = xmlBufferLength(buf);
1527 /* Copy the data into a new buffer, and prefix it with the UTF-16LE BOM
1528 * (byte order mark), which is required by MS's software to understand
1532 xml_data = MALLOC(xml_len);
1534 ret = WIMLIB_ERR_NOMEM;
1535 goto out_free_text_writer;
1539 memcpy(&xml_data[2], content, len);
1541 /* Clean up libxml objects and return success. */
1542 *xml_data_ret = xml_data;
1543 *xml_len_ret = xml_len;
1545 out_free_text_writer:
1546 /* xmlFreeTextWriter will free the attached xmlOutputBuffer. */
1547 xmlFreeTextWriter(writer);
1548 goto out_buffer_free;
1549 out_output_buffer_close:
1550 xmlOutputBufferClose(outbuf);
1554 DEBUG("ret=%d", ret);
1558 ERROR("Error writing XML data");
1559 ret = WIMLIB_ERR_WRITE;
1560 goto out_free_text_writer;
1563 /* Writes the XML data to a WIM file. */
1565 write_wim_xml_data(WIMStruct *wim, int image, u64 total_bytes,
1566 struct wim_reshdr *out_reshdr,
1567 int write_resource_flags)
1573 DEBUG("Writing WIM XML data (image=%d, offset=%"PRIu64")",
1574 image, wim->out_fd.offset);
1576 ret = prepare_wim_xml_data(wim, image, total_bytes,
1577 &xml_data, &xml_len);
1581 /* Write the XML data uncompressed. Although wimlib can handle
1582 * compressed XML data, MS software cannot. */
1583 ret = write_wim_resource_from_buffer(xml_data,
1585 WIM_RESHDR_FLAG_METADATA,
1587 WIMLIB_COMPRESSION_TYPE_NONE,
1591 write_resource_flags);
1593 DEBUG("ret=%d", ret);
1597 /* API function documented in wimlib.h */
1598 WIMLIBAPI const tchar *
1599 wimlib_get_image_name(const WIMStruct *wim, int image)
1601 if (image < 1 || image > wim->hdr.image_count)
1603 return wim->wim_info->images[image - 1].name;
1606 /* API function documented in wimlib.h */
1607 WIMLIBAPI const tchar *
1608 wimlib_get_image_description(const WIMStruct *wim, int image)
1610 if (image < 1 || image > wim->hdr.image_count)
1612 return wim->wim_info->images[image - 1].description;
1615 /* API function documented in wimlib.h */
1617 wimlib_image_name_in_use(const WIMStruct *wim, const tchar *name)
1619 if (!name || !*name)
1621 for (int i = 1; i <= wim->hdr.image_count; i++)
1622 if (!tstrcmp(wim->wim_info->images[i - 1].name, name))
1628 /* API function documented in wimlib.h */
1630 wimlib_get_xml_data(WIMStruct *wim, void **buf_ret, size_t *bufsize_ret)
1632 const struct wim_reshdr *xml_reshdr;
1634 if (wim->filename == NULL && filedes_is_seekable(&wim->in_fd))
1635 return WIMLIB_ERR_INVALID_PARAM;
1637 if (buf_ret == NULL || bufsize_ret == NULL)
1638 return WIMLIB_ERR_INVALID_PARAM;
1640 xml_reshdr = &wim->hdr.xml_data_reshdr;
1642 DEBUG("Reading XML data.");
1643 *bufsize_ret = xml_reshdr->uncompressed_size;
1644 return wim_reshdr_to_data(xml_reshdr, wim, buf_ret);
1648 wimlib_extract_xml_data(WIMStruct *wim, FILE *fp)
1654 ret = wimlib_get_xml_data(wim, &buf, &bufsize);
1658 if (fwrite(buf, 1, bufsize, fp) != bufsize) {
1659 ERROR_WITH_ERRNO("Failed to extract XML data");
1660 ret = WIMLIB_ERR_WRITE;
1666 /* API function documented in wimlib.h */
1668 wimlib_set_image_name(WIMStruct *wim, int image, const tchar *name)
1674 DEBUG("Setting the name of image %d to %"TS, image, name);
1676 ret = can_modify_wim(wim);
1683 if (image < 1 || image > wim->hdr.image_count) {
1684 ERROR("%d is not a valid image", image);
1685 return WIMLIB_ERR_INVALID_IMAGE;
1688 for (i = 1; i <= wim->hdr.image_count; i++) {
1691 if (!tstrcmp(wim->wim_info->images[i - 1].name, name)) {
1692 ERROR("The name \"%"TS"\" is already in use in the WIM!",
1694 return WIMLIB_ERR_IMAGE_NAME_COLLISION;
1700 return WIMLIB_ERR_NOMEM;
1702 FREE(wim->wim_info->images[image - 1].name);
1703 wim->wim_info->images[image - 1].name = p;
1708 do_set_image_info_str(WIMStruct *wim, int image, const tchar *tstr,
1712 tchar **dest_tstr_p;
1715 ret = can_modify_wim(wim);
1719 if (image < 1 || image > wim->hdr.image_count) {
1720 ERROR("%d is not a valid image", image);
1721 return WIMLIB_ERR_INVALID_IMAGE;
1724 tstr_copy = TSTRDUP(tstr);
1726 return WIMLIB_ERR_NOMEM;
1730 dest_tstr_p = (tchar**)((void*)&wim->wim_info->images[image - 1] + offset);
1733 *dest_tstr_p = tstr_copy;
1737 /* API function documented in wimlib.h */
1739 wimlib_set_image_descripton(WIMStruct *wim, int image,
1740 const tchar *description)
1742 return do_set_image_info_str(wim, image, description,
1743 offsetof(struct image_info, description));
1746 /* API function documented in wimlib.h */
1748 wimlib_set_image_flags(WIMStruct *wim, int image, const tchar *flags)
1750 return do_set_image_info_str(wim, image, flags,
1751 offsetof(struct image_info, flags));