4 * Deals with the XML information in WIM files. Uses the C library libxml2.
8 * Copyright (C) 2012 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/.
26 #include "wimlib_internal.h"
29 #include "timestamp.h"
30 #include "lookup_table.h"
33 #include <libxml/parser.h>
34 #include <libxml/tree.h>
35 #include <libxml/xmlwriter.h>
37 /* Structures used to form an in-memory representation of the XML data (other
38 * than the raw parse tree from libxml). */
40 struct windows_version {
52 char *installation_type;
57 char *default_language;
60 bool windows_version_exists;
61 struct windows_version windows_version;
71 u64 last_modification_time;
72 bool windows_info_exists;
73 struct windows_info windows_info;
77 char *display_description;
80 struct lookup_table *lookup_table;
85 /* Returns a statically allocated string that is a string representation of the
86 * architecture number. */
87 static const char *get_arch(int arch)
97 /* XXX Are there other arch values? */
99 snprintf(buf, sizeof(buf), "%d (unknown)", arch);
105 /* Iterate through the children of an xmlNode. */
106 #define for_node_child(parent, child) \
107 for (child = parent->children; child != NULL; child = child->next)
109 /* Utility functions for xmlNodes */
110 static inline bool node_is_element(xmlNode *node)
112 return node->type == XML_ELEMENT_NODE;
115 static inline bool node_is_text(xmlNode *node)
117 return node->type == XML_TEXT_NODE;
120 static inline bool node_is_attribute(xmlNode *node)
122 return node->type == XML_ATTRIBUTE_NODE;
125 static inline bool node_name_is(xmlNode *node, const char *name)
127 /* For now, both upper case and lower case element names are accepted. */
128 return strcasecmp((const char *)node->name, name) == 0;
131 /* Finds the text node that is a child of an element node and returns its
132 * content converted to a 64-bit unsigned integer. Returns 0 if no text node is
134 static u64 node_get_u64(const xmlNode *u64_node)
137 for_node_child(u64_node, child)
138 if (node_is_text(child))
139 return strtoull((const char *)child->content, NULL, 10);
143 /* Like node_get_u64(), but expects a number in base 16. */
144 static u64 node_get_hex_u64(const xmlNode *u64_node)
147 for_node_child(u64_node, child)
148 if (node_is_text(child))
149 return strtoull(child->content, NULL, 16);
153 static int node_get_string(const xmlNode *string_node, char **str)
158 for_node_child(string_node, child) {
159 if (node_is_text(child) && child->content) {
160 p = STRDUP(child->content);
162 ERROR("Out of memory");
163 return WIMLIB_ERR_NOMEM;
172 /* Returns the timestamp from a time node. It has child elements <HIGHPART> and
173 * <LOWPART> that are then used to construct a 64-bit timestamp. */
174 static u64 node_get_timestamp(const xmlNode *time_node)
179 for_node_child(time_node, child) {
180 if (!node_is_element(child))
182 if (node_name_is(child, "HIGHPART"))
183 high_part = node_get_hex_u64(child);
184 else if (node_name_is(child, "LOWPART"))
185 low_part = node_get_hex_u64(child);
187 return (u64)low_part | ((u64)high_part << 32);
190 /* Used to sort an array of struct image_infos by their image indices. */
191 static int sort_by_index(const void *p1, const void *p2)
193 u64 index_1 = ((struct image_info*)p1)->index;
194 u64 index_2 = ((struct image_info*)p1)->index;
195 if (index_1 < index_2)
197 else if (index_1 > index_2)
204 /* Frees memory allocated inside a struct windows_info structure. */
205 static void destroy_windows_info(struct windows_info *windows_info)
209 FREE(windows_info->product_name);
210 FREE(windows_info->edition_id);
211 FREE(windows_info->installation_type);
212 FREE(windows_info->product_type);
213 for (i = 0; i < windows_info->num_languages; i++)
214 FREE(windows_info->languages[i]);
215 FREE(windows_info->languages);
216 FREE(windows_info->system_root);
219 /* Frees memory allocated inside a struct image_info structure. */
220 static void destroy_image_info(struct image_info *image_info)
222 FREE(image_info->name);
223 FREE(image_info->description);
224 FREE(image_info->flags);
225 FREE(image_info->display_name);
226 FREE(image_info->display_description);
227 destroy_windows_info(&image_info->windows_info);
228 memset(image_info, 0, sizeof(struct image_info));
231 void free_wim_info(struct wim_info *info)
236 for (i = 0; i < info->num_images; i++)
237 destroy_image_info(&info->images[i]);
244 /* Reads the information from a <VERSION> element inside the <WINDOWS> element.
246 static void xml_read_windows_version(const xmlNode *version_node,
247 struct windows_version* windows_version)
250 for_node_child(version_node, child) {
251 if (!node_is_element(child))
253 if (node_name_is(child, "MAJOR"))
254 windows_version->major = node_get_u64(child);
255 else if (node_name_is(child, "MINOR"))
256 windows_version->minor = node_get_u64(child);
257 else if (node_name_is(child, "BUILD"))
258 windows_version->build = node_get_u64(child);
259 else if (node_name_is(child, "SPBUILD"))
260 windows_version->sp_build = node_get_u64(child);
261 else if (node_name_is(child, "SPLEVEL"))
262 windows_version->sp_level = node_get_u64(child);
266 /* Reads the information from a <LANGUAGE> element inside a <WINDOWS> element.
268 static int xml_read_languages(const xmlNode *languages_node,
269 char ***languages_ret,
270 u64 *num_languages_ret,
271 char **default_language_ret)
280 for_node_child(languages_node, child)
281 if (node_is_element(child) && node_name_is(child, "LANGUAGE"))
284 languages = CALLOC(num_languages, sizeof(char*));
286 ERROR("Out of memory");
287 return WIMLIB_ERR_NOMEM;
290 *languages_ret = languages;
291 *num_languages_ret = num_languages;
295 for_node_child(languages_node, child) {
296 if (!node_is_element(child))
298 if (node_name_is(child, "LANGUAGE"))
299 ret = node_get_string(child, &languages[i++]);
300 else if (node_name_is(child, "DEFAULT"))
301 ret = node_get_string(child, default_language_ret);
308 /* Reads the information from a <WINDOWS> element inside an <IMAGE> element. */
309 static int xml_read_windows_info(const xmlNode *windows_node,
310 struct windows_info *windows_info)
315 for_node_child(windows_node, child) {
316 if (!node_is_element(child))
318 if (node_name_is(child, "ARCH")) {
319 windows_info->arch = node_get_u64(child);
320 } else if (node_name_is(child, "PRODUCTNAME")) {
321 ret = node_get_string(child,
322 &windows_info->product_name);
323 } else if (node_name_is(child, "EDITIONID")) {
324 ret = node_get_string(child,
325 &windows_info->edition_id);
326 } else if (node_name_is(child, "INSTALLATIONTYPE")) {
327 ret = node_get_string(child,
328 &windows_info->installation_type);
329 } else if (node_name_is(child, "PRODUCTTYPE")) {
330 ret = node_get_string(child,
331 &windows_info->product_type);
332 } else if (node_name_is(child, "PRODUCTSUITE")) {
333 ret = node_get_string(child,
334 &windows_info->product_suite);
335 } else if (node_name_is(child, "LANGUAGES")) {
336 ret = xml_read_languages(child,
337 &windows_info->languages,
338 &windows_info->num_languages,
339 &windows_info->default_language);
340 } else if (node_name_is(child, "VERSION")) {
341 xml_read_windows_version(child,
342 &windows_info->windows_version);
343 windows_info->windows_version_exists = true;
344 } else if (node_name_is(child, "SYSTEMROOT")) {
345 ret = node_get_string(child, &windows_info->system_root);
346 } else if (node_name_is(child, "HAL")) {
347 ret = node_get_string(child, &windows_info->hal);
356 /* Reads the information from an <IMAGE> element. */
357 static int xml_read_image_info(xmlNode *image_node,
358 struct image_info *image_info)
364 index_prop = xmlGetProp(image_node, "INDEX");
367 image_info->index = strtoul(index_prop, &tmp, 10);
370 image_info->index = 0;
374 for_node_child(image_node, child) {
375 if (!node_is_element(child))
377 if (node_name_is(child, "DIRCOUNT"))
378 image_info->dir_count = node_get_u64(child);
379 else if (node_name_is(child, "FILECOUNT"))
380 image_info->file_count = node_get_u64(child);
381 else if (node_name_is(child, "TOTALBYTES"))
382 image_info->total_bytes = node_get_u64(child);
383 else if (node_name_is(child, "HARDLINKBYTES"))
384 image_info->hard_link_bytes = node_get_u64(child);
385 else if (node_name_is(child, "CREATIONTIME"))
386 image_info->creation_time = node_get_timestamp(child);
387 else if (node_name_is(child, "LASTMODIFICATIONTIME"))
388 image_info->last_modification_time = node_get_timestamp(child);
389 else if (node_name_is(child, "WINDOWS")) {
390 DEBUG("Found <WINDOWS> tag");
391 ret = xml_read_windows_info(child,
392 &image_info->windows_info);
393 image_info->windows_info_exists = true;
394 } else if (node_name_is(child, "NAME")) {
395 ret = node_get_string(child, &image_info->name);
396 } else if (node_name_is(child, "DESCRIPTION")) {
397 ret = node_get_string(child, &image_info->description);
398 } else if (node_name_is(child, "FLAGS")) {
399 ret = node_get_string(child, &image_info->flags);
400 } else if (node_name_is(child, "DISPLAYNAME")) {
401 ret = node_get_string(child, &image_info->display_name);
402 } else if (node_name_is(child, "DISPLAYDESCRIPTION")) {
403 ret = node_get_string(child, &image_info->display_description);
408 if (!image_info->name) {
409 WARNING("Image with index %"PRIu64" has no name",
411 image_info->name = MALLOC(1);
412 if (!image_info->name) {
413 ERROR("Out of memory");
414 return WIMLIB_ERR_NOMEM;
416 image_info->name[0] = '\0';
421 /* Reads the information from a <WIM> element, which should be the root element
422 * of the XML tree. */
423 static int xml_read_wim_info(const xmlNode *wim_node,
424 struct wim_info **wim_info_ret)
426 struct wim_info *wim_info;
430 struct image_info *cur_image_info;
432 wim_info = CALLOC(1, sizeof(struct wim_info));
434 ERROR("Out of memory");
435 return WIMLIB_ERR_NOMEM;
438 /* Count how many images there are. */
440 for_node_child(wim_node, child)
441 if (node_is_element(child) && node_name_is(child, "IMAGE"))
447 /* Allocate the array of struct image_infos and fill them in. */
448 wim_info->images = CALLOC(num_images, sizeof(wim_info->images[0]));
449 if (!wim_info->images) {
450 ret = WIMLIB_ERR_NOMEM;
451 ERROR("Out of memory!");
454 wim_info->num_images = num_images;
455 cur_image_info = wim_info->images;
456 for_node_child(wim_node, child) {
457 if (!node_is_element(child))
459 if (node_name_is(child, "IMAGE")) {
460 DEBUG("Found <IMAGE> tag");
461 ret = xml_read_image_info(child, cur_image_info++);
464 } else if (node_name_is(child, "TOTALBYTES")) {
465 wim_info->total_bytes = node_get_u64(child);
469 /* Sort the array of struct image_infos by image index. */
470 qsort(wim_info->images, wim_info->num_images,
471 sizeof(struct image_info), sort_by_index);
473 *wim_info_ret = wim_info;
476 free_wim_info(wim_info);
480 /* Prints the information contained in a struct windows_info structure. */
481 static void print_windows_info(const struct windows_info *windows_info)
484 const struct windows_version *windows_version;
486 printf("Architecture: %s\n", get_arch(windows_info->arch));
487 printf("Product Name: %s\n", windows_info->product_name);
488 printf("Edition ID: %s\n", windows_info->edition_id);
489 printf("Installation Type: %s\n", windows_info->installation_type);
490 if (windows_info->hal)
491 printf("HAL: %s\n", windows_info->hal);
492 printf("Product Type: %s\n", windows_info->product_type);
493 if (windows_info->product_suite)
494 printf("Product Suite: %s\n", windows_info->product_suite);
495 printf("Languages: ");
496 for (i = 0; i < windows_info->num_languages; i++) {
497 fputs(windows_info->languages[i], stdout);
501 printf("Default Language: %s\n", windows_info->default_language);
502 printf("System Root: %s\n", windows_info->system_root);
503 if (windows_info->windows_version_exists) {
504 windows_version = &windows_info->windows_version;
505 printf("Major Version: %"PRIu64"\n",
506 windows_version->major);
507 printf("Minor Version: %"PRIu64"\n",
508 windows_version->minor);
509 printf("Build: %"PRIu64"\n",
510 windows_version->build);
511 printf("Service Pack Build: %"PRIu64"\n",
512 windows_version->sp_build);
513 printf("Service Pack Level: %"PRIu64"\n",
514 windows_version->sp_level);
519 /* Writes the information contained in a struct windows_version structure to the XML
520 * document being constructed in memory. This is the <VERSION> element inside
521 * the <WINDOWS> element. */
522 static int xml_write_windows_version(xmlTextWriter *writer,
523 const struct windows_version *version)
526 rc = xmlTextWriterStartElement(writer, "VERSION");
530 rc = xmlTextWriterWriteFormatElement(writer, "MAJOR", "%"PRIu64,
535 rc = xmlTextWriterWriteFormatElement(writer, "MINOR", "%"PRIu64,
540 rc = xmlTextWriterWriteFormatElement(writer, "BUILD", "%"PRIu64,
545 rc = xmlTextWriterWriteFormatElement(writer, "SPBUILD", "%"PRIu64,
550 rc = xmlTextWriterWriteFormatElement(writer, "SPLEVEL", "%"PRIu64,
555 return xmlTextWriterEndElement(writer); /* </VERSION> */
558 /* Writes the information contained in a struct windows_info structure to the XML
559 * document being constructed in memory. This is the <WINDOWS> element. */
560 static int xml_write_windows_info(xmlTextWriter *writer,
561 const struct windows_info *windows_info)
564 rc = xmlTextWriterStartElement(writer, "WINDOWS");
569 rc = xmlTextWriterWriteFormatElement(writer, "ARCH", "%"PRIu64,
574 if (windows_info->product_name) {
575 rc = xmlTextWriterWriteElement(writer, "PRODUCTNAME",
576 windows_info->product_name);
581 if (windows_info->edition_id) {
582 rc = xmlTextWriterWriteElement(writer, "EDITIONID",
583 windows_info->edition_id);
588 if (windows_info->installation_type) {
589 rc = xmlTextWriterWriteElement(writer, "INSTALLATIONTYPE",
590 windows_info->installation_type);
595 if (windows_info->hal) {
596 rc = xmlTextWriterWriteElement(writer, "HAL",
602 if (windows_info->system_root) {
603 rc = xmlTextWriterWriteElement(writer, "SYSTEMROOT",
604 windows_info->system_root);
609 if (windows_info->product_type) {
610 rc = xmlTextWriterWriteElement(writer, "PRODUCTTYPE",
611 windows_info->product_type);
616 if (windows_info->product_suite) {
617 rc = xmlTextWriterWriteElement(writer, "PRODUCTSUITE",
618 windows_info->product_suite);
623 if (windows_info->num_languages) {
624 rc = xmlTextWriterStartElement(writer, "LANGUAGES");
628 for (int i = 0; i < windows_info->num_languages; i++) {
629 rc = xmlTextWriterWriteElement(writer, "LANGUAGE",
630 windows_info->languages[i]);
634 rc = xmlTextWriterWriteElement(writer, "DEFAULT",
635 windows_info->default_language);
639 rc = xmlTextWriterEndElement(writer); /* </LANGUAGES> */
644 if (windows_info->windows_version_exists) {
645 rc = xml_write_windows_version(writer, &windows_info->windows_version);
650 return xmlTextWriterEndElement(writer); /* </WINDOWS> */
653 /* Writes a time element to the XML document being constructed in memory. */
654 static int xml_write_time(xmlTextWriter *writer, const char *element_name,
658 rc = xmlTextWriterStartElement(writer, element_name);
662 rc = xmlTextWriterWriteFormatElement(writer, "HIGHPART",
663 "0x%"PRIX32, (u32)(time >> 32));
667 rc = xmlTextWriterWriteFormatElement(writer, "LOWPART",
668 "0x%"PRIX32, (u32)time);
672 rc = xmlTextWriterEndElement(writer); /* </@element_name> */
679 /* Writes an <IMAGE> element to the XML document. */
680 static int xml_write_image_info(xmlTextWriter *writer,
681 const struct image_info *image_info)
684 rc = xmlTextWriterStartElement(writer, "IMAGE");
688 rc = xmlTextWriterWriteFormatAttribute(writer, "INDEX", "%"PRIu64,
693 rc = xmlTextWriterWriteFormatElement(writer, "DIRCOUNT", "%"PRIu64,
694 image_info->dir_count);
698 rc = xmlTextWriterWriteFormatElement(writer, "FILECOUNT", "%"PRIu64,
699 image_info->file_count);
703 rc = xmlTextWriterWriteFormatElement(writer, "TOTALBYTES", "%"PRIu64,
704 image_info->total_bytes);
708 rc = xmlTextWriterWriteFormatElement(writer, "HARDLINKBYTES", "%"PRIu64,
709 image_info->hard_link_bytes);
713 rc = xml_write_time(writer, "CREATIONTIME",
714 image_info->creation_time);
718 rc = xml_write_time(writer, "LASTMODIFICATIONTIME",
719 image_info->last_modification_time);
723 if (image_info->windows_info_exists) {
724 rc = xml_write_windows_info(writer, &image_info->windows_info);
728 DEBUG("<WINDOWS> tag does not exist.");
731 if (image_info->name) {
732 rc = xmlTextWriterWriteElement(writer, "NAME", image_info->name);
736 if (image_info->description) {
737 rc = xmlTextWriterWriteElement(writer, "DESCRIPTION",
738 image_info->description);
742 if (image_info->display_name) {
743 rc = xmlTextWriterWriteElement(writer, "DISPLAYNAME",
744 image_info->display_name);
748 if (image_info->display_description) {
749 rc = xmlTextWriterWriteElement(writer, "DISPLAYDESCRIPTION",
750 image_info->display_description);
755 if (image_info->flags) {
756 rc = xmlTextWriterWriteElement(writer, "FLAGS",
762 return xmlTextWriterEndElement(writer); /* </IMAGE> */
767 /* Makes space for another image in the XML information and return a pointer to
769 static struct image_info *add_image_info_struct(struct wim_info *wim_info)
771 struct image_info *images;
773 images = CALLOC(wim_info->num_images + 1, sizeof(struct image_info));
776 memcpy(images, wim_info->images,
777 wim_info->num_images * sizeof(struct image_info));
778 FREE(wim_info->images);
779 wim_info->images = images;
780 wim_info->num_images++;
781 return &images[wim_info->num_images - 1];
784 static int clone_windows_info(const struct windows_info *old,
785 struct windows_info *new)
789 if (old->product_name && !(new->product_name = STRDUP(old->product_name)))
790 return WIMLIB_ERR_NOMEM;
791 if (old->edition_id && !(new->edition_id = STRDUP(old->edition_id)))
792 return WIMLIB_ERR_NOMEM;
793 if (old->installation_type && !(new->installation_type =
794 STRDUP(old->installation_type)))
795 return WIMLIB_ERR_NOMEM;
796 if (old->hal && !(new->hal = STRDUP(old->hal)))
797 return WIMLIB_ERR_NOMEM;
798 if (old->product_type && !(new->product_type = STRDUP(old->product_type)))
799 return WIMLIB_ERR_NOMEM;
800 if (old->product_suite && !(new->product_suite = STRDUP(old->product_suite)))
801 return WIMLIB_ERR_NOMEM;
803 if (old->languages) {
804 new->languages = CALLOC(old->num_languages, sizeof(char*));
806 return WIMLIB_ERR_NOMEM;
807 new->num_languages = old->num_languages;
808 for (i = 0; i < new->num_languages; i++) {
809 if (!old->languages[i])
811 new->languages[i] = STRDUP(old->languages[i]);
812 if (!new->languages[i])
813 return WIMLIB_ERR_NOMEM;
816 if (old->default_language &&
817 !(new->default_language = STRDUP(old->default_language)))
818 return WIMLIB_ERR_NOMEM;
819 if (old->system_root && !(new->system_root = STRDUP(old->system_root)))
820 return WIMLIB_ERR_NOMEM;
821 if (old->windows_version_exists) {
822 new->windows_version_exists = true;
823 memcpy(&new->windows_version, &old->windows_version,
824 sizeof(old->windows_version));
829 static int clone_image_info(const struct image_info *old, struct image_info *new)
831 new->dir_count = old->dir_count;
832 new->file_count = old->file_count;
833 new->total_bytes = old->total_bytes;
834 new->hard_link_bytes = old->hard_link_bytes;
835 new->creation_time = old->creation_time;
836 new->last_modification_time = old->last_modification_time;
838 if (!(new->name = STRDUP(old->name)))
839 return WIMLIB_ERR_NOMEM;
841 if (old->description)
842 if (!(new->description = STRDUP(old->description)))
843 return WIMLIB_ERR_NOMEM;
845 if (old->display_name)
846 if (!(new->display_name = STRDUP(old->display_name)))
847 return WIMLIB_ERR_NOMEM;
849 if (old->display_description)
850 if (!(new->display_description = STRDUP(old->display_description)))
851 return WIMLIB_ERR_NOMEM;
854 if (!(new->flags = STRDUP(old->flags)))
855 return WIMLIB_ERR_NOMEM;
857 if (old->windows_info_exists) {
858 new->windows_info_exists = true;
859 return clone_windows_info(&old->windows_info,
865 /* Copies the XML information for an image between WIM files.
867 * @dest_image_name and @dest_image_description are ignored if they are NULL;
868 * otherwise, they are used to override the image name and/or image description
869 * from the XML data in the source WIM file. */
870 int xml_export_image(const struct wim_info *old_wim_info,
872 struct wim_info **new_wim_info_p,
873 const char *dest_image_name,
874 const char *dest_image_description)
876 struct wim_info *new_wim_info;
877 struct image_info *image_info;
880 DEBUG("Copying XML data between WIM files for source image %d.", image);
882 wimlib_assert(image >= 1 && image <= old_wim_info->num_images);
884 if (*new_wim_info_p) {
885 new_wim_info = *new_wim_info_p;
887 new_wim_info = CALLOC(1, sizeof(struct wim_info));
892 image_info = add_image_info_struct(new_wim_info);
896 ret = clone_image_info(&old_wim_info->images[image - 1], image_info);
900 image_info->index = new_wim_info->num_images;
902 if (dest_image_name) {
903 FREE(image_info->name);
904 image_info->name = STRDUP(dest_image_name);
905 if (!image_info->name)
908 if (dest_image_description) {
909 FREE(image_info->description);
910 image_info->description = STRDUP(dest_image_description);
911 if (!image_info->description)
914 *new_wim_info_p = new_wim_info;
917 ERROR("Out of memory");
918 free_wim_info(new_wim_info);
919 return WIMLIB_ERR_NOMEM;
922 /* Removes an image from the XML information. */
923 void xml_delete_image(struct wim_info **wim_info_p, int image)
925 struct wim_info *wim_info;
928 DEBUG("Deleting image %d from the XML data.", image);
930 wim_info = *wim_info_p;
932 wimlib_assert(wim_info);
933 wimlib_assert(image >= 1 && image <= wim_info->num_images);
935 destroy_image_info(&wim_info->images[image - 1]);
937 for (i = image - 1; i < wim_info->num_images - 1; i++) {
938 memcpy(&wim_info->images[i], &wim_info->images[i + 1],
939 sizeof(struct image_info));
940 wim_info->images[i].index--;
943 if (--wim_info->num_images == 0) {
944 free_wim_info(wim_info);
949 size_t xml_get_max_image_name_len(const WIMStruct *w)
953 uint num_images = w->wim_info->num_images;
954 for (i = 0; i < num_images; i++)
955 len = max(len, strlen(w->wim_info->images[i].name));
959 #ifdef ENABLE_CUSTOM_MEMORY_ALLOCATOR
960 void xml_set_memory_allocator(void *(*malloc_func)(size_t),
961 void (*free_func)(void *),
962 void *(*realloc_func)(void *, size_t))
964 xmlMemSetup(free_func, malloc_func, realloc_func, STRDUP);
968 static int calculate_dentry_statistics(struct dentry *dentry, void *arg)
970 struct image_info *info = arg;
971 struct lookup_table *lookup_table = info->lookup_table;
972 const struct inode *inode = dentry->d_inode;
973 struct lookup_table_entry *lte;
975 /* Update directory count and file count.
977 * Each dentry counts as either a file or a directory, but not both.
978 * The root directory is an exception: it is not counted.
980 * Symbolic links and junction points (and presumably other reparse
981 * points) count as regular files. This is despite the fact that
982 * junction points have FILE_ATTRIBUTE_DIRECTORY set.
984 if (dentry_is_root(dentry))
987 if (inode_is_directory(inode))
993 * Update total bytes and hard link bytes.
995 * Unfortunately there are some inconsistencies/bugs in the way this is
998 * If there are no alternate data streams in the image, the "total
999 * bytes" is the sum of the size of the un-named data stream of each
1000 * inode times the link count of that inode. In other words, it would
1001 * be the total number of bytes of regular files you would have if you
1002 * extracted the full image without any hard-links. The "hard link
1003 * bytes" is equal to the "total bytes" minus the size of the un-named
1004 * data stream of each inode. In other words, the "hard link bytes"
1005 * counts the size of the un-named data stream for all the links to each
1006 * inode except the first one.
1008 * Reparse points and directories don't seem to be counted in either the
1009 * total bytes or the hard link bytes.
1011 * And now we get to the most confusing part, the alternate data
1012 * streams. They are not counted in the "total bytes". However, if the
1013 * link count of an inode with alternate data streams is 2 or greater,
1014 * the size of all the alternate data streams is included in the "hard
1015 * link bytes", and this size is multiplied by the link count (NOT one
1016 * less than the link count).
1018 lte = inode_unnamed_lte(inode, info->lookup_table);
1020 info->total_bytes += wim_resource_size(lte);
1021 if (!dentry_is_first_in_inode(dentry))
1022 info->hard_link_bytes += wim_resource_size(lte);
1025 if (inode->link_count >= 2 && dentry_is_first_in_inode(dentry)) {
1026 for (unsigned i = 0; i < inode->num_ads; i++) {
1027 if (inode->ads_entries[i].stream_name_len) {
1028 lte = inode_stream_lte(inode, i + 1, lookup_table);
1030 info->hard_link_bytes += inode->link_count *
1031 wim_resource_size(lte);
1039 void xml_update_image_info(WIMStruct *w, int image)
1041 struct image_info *image_info;
1042 struct dentry *root;
1045 DEBUG("Updating the image info for image %d", image);
1047 image_info = &w->wim_info->images[image - 1];
1049 image_info->file_count = 0;
1050 image_info->dir_count = 0;
1051 image_info->total_bytes = 0;
1052 image_info->hard_link_bytes = 0;
1054 flags_save = image_info->flags;
1055 image_info->lookup_table = w->lookup_table;
1057 for_dentry_in_tree(w->image_metadata[image - 1].root_dentry,
1058 calculate_dentry_statistics,
1061 image_info->lookup_table = NULL;
1062 image_info->flags = flags_save;
1063 image_info->last_modification_time = get_wim_timestamp();
1066 /* Adds an image to the XML information. */
1067 int xml_add_image(WIMStruct *w, const char *name)
1069 struct wim_info *wim_info;
1070 struct image_info *image_info;
1072 wimlib_assert(name);
1074 DEBUG("Adding image: name = %s", name);
1076 /* If this is the first image, allocate the struct wim_info. Otherwise
1077 * use the existing struct wim_info. */
1079 wim_info = w->wim_info;
1081 DEBUG("Allocing struct wim_info with 1 image");
1082 wim_info = CALLOC(1, sizeof(struct wim_info));
1084 ERROR("Could not allocate WIM information struct--- "
1086 return WIMLIB_ERR_NOMEM;
1090 image_info = add_image_info_struct(wim_info);
1092 goto out_free_wim_info;
1094 if (!(image_info->name = STRDUP(name)))
1095 goto out_destroy_image_info;
1097 w->wim_info = wim_info;
1098 image_info->index = wim_info->num_images;
1099 image_info->creation_time = get_wim_timestamp();
1100 xml_update_image_info(w, image_info->index);
1103 out_destroy_image_info:
1104 destroy_image_info(image_info);
1107 wim_info->num_images--;
1110 ERROR("Out of memory");
1111 return WIMLIB_ERR_NOMEM;
1114 /* Prints information about the specified image from struct wim_info structure.
1115 * @image may be WIM_ALL_IMAGES. */
1116 void print_image_info(const struct wim_info *wim_info, int image)
1119 const struct image_info *image_info;
1123 if (image == WIM_ALL_IMAGES) {
1124 for (i = 1; i <= wim_info->num_images; i++)
1125 print_image_info(wim_info, i);
1130 image_info = &wim_info->images[image - 1];
1132 printf("Index: %"PRIu64"\n",
1134 printf("Name: %s\n",
1137 /* Always print the Description: part even if there is no
1139 if (image_info->description)
1140 desc = image_info->description;
1143 printf("Description: %s\n", desc);
1145 if (image_info->display_name)
1146 printf("Display Name: %s\n",
1147 image_info->display_name);
1149 if (image_info->display_description)
1150 printf("Display Description: %s\n",
1151 image_info->display_description);
1153 printf("Directory Count: %"PRIu64"\n",
1154 image_info->dir_count);
1155 printf("File Count: %"PRIu64"\n",
1156 image_info->file_count);
1157 printf("Total Bytes: %"PRIu64"\n",
1158 image_info->total_bytes);
1159 printf("Hard Link Bytes: %"PRIu64"\n",
1160 image_info->hard_link_bytes);
1162 time = wim_timestamp_to_unix(image_info->creation_time);
1163 p = asctime(gmtime(&time));
1164 *(strrchr(p, '\n')) = '\0';
1166 printf("Creation Time: %s UTC\n", p);
1168 time = wim_timestamp_to_unix(image_info->last_modification_time);
1169 p = asctime(gmtime(&time));
1170 *(strrchr(p, '\n')) = '\0';
1172 printf("Last Modification Time: %s UTC\n", p);
1173 if (image_info->windows_info_exists)
1174 print_windows_info(&image_info->windows_info);
1175 if (image_info->flags)
1176 printf("Flags: %s\n", image_info->flags);
1182 * Reads the XML data from a WIM file.
1184 int read_xml_data(FILE *fp, const struct resource_entry *res, u8 **xml_data_ret,
1185 struct wim_info **info_ret)
1192 DEBUG("XML data is %"PRIu64" bytes at offset %"PRIu64"",
1193 (u64)res->size, res->offset);
1195 if (resource_is_compressed(res)) {
1196 ERROR("XML data is supposed to be uncompressed");
1197 ret = WIMLIB_ERR_XML;
1198 goto out_cleanup_parser;
1200 if (res->size < 2) {
1201 ERROR("XML data must be at least 2 bytes");
1202 ret = WIMLIB_ERR_XML;
1203 goto out_cleanup_parser;
1206 xml_data = MALLOC(res->size + 2);
1208 ret = WIMLIB_ERR_NOMEM;
1209 goto out_cleanup_parser;
1211 ret = read_uncompressed_resource(fp, res->offset, res->size, xml_data);
1213 goto out_free_xml_data;
1215 xml_data[res->size] = 0;
1216 xml_data[res->size + 1] = 0;
1218 DEBUG("Parsing XML using libxml2 to create XML tree.");
1220 doc = xmlReadMemory(xml_data, res->size, "noname.xml", "UTF-16", 0);
1224 ERROR("Failed to parse XML data");
1225 ret = WIMLIB_ERR_XML;
1226 goto out_free_xml_data;
1229 DEBUG("Constructing WIM information structure from XML tree.");
1231 root = xmlDocGetRootElement(doc);
1233 ERROR("Empty XML document");
1234 ret = WIMLIB_ERR_XML;
1238 if (!node_is_element(root) || !node_name_is(root, "WIM")) {
1239 ERROR("Expected <WIM> for the root XML element (found <%s>)",
1241 ret = WIMLIB_ERR_XML;
1245 ret = xml_read_wim_info(root, info_ret);
1249 DEBUG("Freeing XML tree.");
1253 *xml_data_ret = xml_data;
1264 #define CHECK_RET ({ if (ret < 0) { \
1265 ERROR("Error writing XML data"); \
1266 ret = WIMLIB_ERR_WRITE; \
1271 * Writes XML data to a WIM file.
1273 * If @total_bytes is non-zero, it specifies what to write to the TOTALBYTES
1274 * element in the XML data. If zero, TOTALBYTES is given the default value of
1275 * the offset of the XML data.
1277 int write_xml_data(const struct wim_info *wim_info, int image, FILE *out,
1281 xmlTextWriter *writer;
1286 const xmlChar *content;
1289 size_t bytes_written;
1291 wimlib_assert(image == WIM_ALL_IMAGES ||
1292 (wim_info != NULL && image >= 1 &&
1293 image <= wim_info->num_images));
1295 /* The contents of the <TOTALBYTES> element in the XML data, under the
1296 * <WIM> element not the <IMAGE> element, is (for non-spit WIMs) the
1297 * size of the WIM file excluding the XML data and integrity table,
1298 * which is the current offset, since the XML data goes at the end of
1299 * the WIM file before the integrity table. */
1300 if (total_bytes == 0) {
1301 total_bytes = ftello(out);
1302 if (total_bytes == (u64)-1)
1303 return WIMLIB_ERR_WRITE;
1306 DEBUG("Creating XML buffer and text writer.");
1307 buf = xmlBufferCreate();
1309 ERROR("Failed to allocate XML buffer");
1310 ret = WIMLIB_ERR_NOMEM;
1313 writer = xmlNewTextWriterMemory(buf, 0);
1315 ERROR("Failed to allocate XML writer");
1316 ret = WIMLIB_ERR_NOMEM;
1321 /* M$'s WIM files do not have XML declarations, so do not write one.
1322 * I'm not sure how we can force the document to be written in UTF-16
1323 * without calling xmlTextWriterStartDocument(), though, so currently it
1324 * is composed in a buffer UTF-8, then converted to UTF-16. */
1326 ret = xmlTextWriterStartDocument(writer, NULL, "UTF-16", NULL);
1330 DEBUG("Writing <WIM> element");
1331 ret = xmlTextWriterStartElement(writer, "WIM");
1334 ret = xmlTextWriterWriteFormatElement(writer, "TOTALBYTES", "%"PRIu64,
1339 num_images = wim_info->num_images;
1342 DEBUG("Writing %u <IMAGE> elements", num_images);
1344 for (i = 1; i <= num_images; i++) {
1345 if (image != WIM_ALL_IMAGES && i != image)
1347 DEBUG("Writing <IMAGE> element for image %d", i);
1348 ret = xml_write_image_info(writer, &wim_info->images[i - 1]);
1352 ret = xmlTextWriterEndElement(writer);
1355 ret = xmlTextWriterEndDocument(writer);
1358 DEBUG("Done composing XML document. Now converting to UTF-16 and "
1359 "writing it to the output file.");
1361 content = xmlBufferContent(buf);
1362 len = xmlBufferLength(buf);
1364 utf16_str = utf8_to_utf16(content, len, &utf16_len);
1366 ret = WIMLIB_ERR_NOMEM;
1370 if ((putc(0xff, out)) == EOF || (putc(0xfe, out) == EOF) ||
1371 ((bytes_written = fwrite(utf16_str, 1, utf16_len, out))
1373 ERROR_WITH_ERRNO("Error writing XML data");
1374 ret = WIMLIB_ERR_WRITE;
1378 DEBUG("Cleaning up.");
1384 xmlFreeTextWriter(writer);
1391 /* Returns the name of the specified image. */
1392 WIMLIBAPI const char *wimlib_get_image_name(const WIMStruct *w, int image)
1394 DEBUG("Getting the name of image %d", image);
1395 if (image < 1 || image > w->hdr.image_count)
1398 return w->wim_info->images[image - 1].name;
1401 /* Returns the description of the specified image. */
1402 WIMLIBAPI const char *wimlib_get_image_description(const WIMStruct *w,
1405 DEBUG("Getting the description of image %d", image);
1406 if (image < 1 || image > w->hdr.image_count)
1409 return w->wim_info->images[image - 1].description;
1412 /* Determines if an image name is already used by some image in the WIM. */
1413 WIMLIBAPI bool wimlib_image_name_in_use(const WIMStruct *w, const char *name)
1417 DEBUG("Checking to see if the image name `%s' is already in use", name);
1418 if (!name || !w->wim_info)
1420 for (i = 1; i <= w->wim_info->num_images; i++)
1421 if (strcmp(w->wim_info->images[i - 1].name, name) == 0)
1427 WIMLIBAPI int wimlib_extract_xml_data(WIMStruct *w, FILE *fp)
1429 DEBUG("Extracting the XML data.");
1430 if (fwrite(w->xml_data, 1, w->hdr.xml_res_entry.size, fp) !=
1431 w->hdr.xml_res_entry.size) {
1432 ERROR_WITH_ERRNO("Failed to extract XML data");
1433 return WIMLIB_ERR_WRITE;
1438 /* Sets the name of an image in the WIM. */
1439 WIMLIBAPI int wimlib_set_image_name(WIMStruct *w, int image, const char *name)
1444 DEBUG("Setting the name of image %d to %s", image, name);
1447 return WIMLIB_ERR_INVALID_PARAM;
1449 if (!name || !*name) {
1450 ERROR("Must specify a non-empty string for the image name");
1451 return WIMLIB_ERR_INVALID_PARAM;
1453 if (image < 1 || image > w->hdr.image_count) {
1454 ERROR("%d is not a valid image", image);
1455 return WIMLIB_ERR_INVALID_IMAGE;
1458 for (i = 1; i <= w->hdr.image_count; i++) {
1461 if (strcmp(w->wim_info->images[i - 1].name, name) == 0) {
1462 ERROR("The name `%s' is already used for image %d",
1464 return WIMLIB_ERR_IMAGE_NAME_COLLISION;
1470 ERROR("Out of memory");
1471 return WIMLIB_ERR_NOMEM;
1473 FREE(w->wim_info->images[image - 1].name);
1474 w->wim_info->images[image - 1].name = p;
1478 /* Sets the description of an image in the WIM. */
1479 WIMLIBAPI int wimlib_set_image_descripton(WIMStruct *w, int image,
1480 const char *description)
1484 DEBUG("Setting the description of image %d to %s", image, description);
1487 return WIMLIB_ERR_INVALID_PARAM;
1489 if (image < 1 || image > w->hdr.image_count) {
1490 ERROR("%d is not a valid image", image);
1491 return WIMLIB_ERR_INVALID_IMAGE;
1494 p = STRDUP(description);
1496 ERROR("Out of memory");
1497 return WIMLIB_ERR_NOMEM;
1502 FREE(w->wim_info->images[image - 1].description);
1503 w->wim_info->images[image - 1].description = p;
1507 WIMLIBAPI int wimlib_set_image_flags(WIMStruct *w, int image,
1512 DEBUG("Setting the flags of image %d to %s", image, flags);
1514 if (image < 1 || image > w->hdr.image_count) {
1515 ERROR("%d is not a valid image", image);
1516 return WIMLIB_ERR_INVALID_IMAGE;
1521 ERROR("Out of memory");
1522 return WIMLIB_ERR_NOMEM;
1527 FREE(w->wim_info->images[image - 1].flags);
1528 w->wim_info->images[image - 1].flags = p;