4 * Deals with the XML information in WIM files. Uses the C library libxml2.
8 * Copyright (C) 2012, 2013, 2015 Eric Biggers
10 * This file is free software; you can redistribute it and/or modify it under
11 * the terms of the GNU Lesser General Public License as published by the Free
12 * Software Foundation; either version 3 of the License, or (at your option) any
15 * This file is distributed in the hope that it will be useful, but WITHOUT
16 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
17 * FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more
20 * You should have received a copy of the GNU Lesser General Public License
21 * along with this file; if not, see http://www.gnu.org/licenses/.
28 #include <libxml/parser.h>
29 #include <libxml/tree.h>
30 #include <libxml/xmlsave.h>
33 #include "wimlib/blob_table.h"
34 #include "wimlib/dentry.h"
35 #include "wimlib/encoding.h"
36 #include "wimlib/error.h"
37 #include "wimlib/file_io.h"
38 #include "wimlib/metadata.h"
39 #include "wimlib/resource.h"
40 #include "wimlib/timestamp.h"
41 #include "wimlib/xml.h"
42 #include "wimlib/write.h"
45 * A wrapper around a WIM file's XML document. The XML document contains
46 * metadata about each image in the WIM file as well as metadata about the WIM
51 /* The parsed XML document as a libxml2 document tree */
54 /* The root element of the document. This is a cached value, equal to
55 * xmlDocGetRootElement(doc). */
58 /* A malloc()ed array containing a pointer to the IMAGE element for each
59 * WIM image. The image with 1-based index 'i' is at index 'i - 1' in
60 * this array. Note: these pointers are cached values, since they could
61 * also be found by searching the document. */
64 /* The number of WIM images (the length of 'images') */
67 /* Temporary memory for UTF-8 => 'tchar' string translations. When an
68 * API function needs to return a 'tchar' string, it uses one of these
69 * array slots to hold the string and returns a pointer to it. */
71 size_t next_string_idx;
75 /*----------------------------------------------------------------------------*
76 * Internal functions *
77 *----------------------------------------------------------------------------*/
79 /* Iterate through the children of an xmlNode. */
80 #define node_for_each_child(parent, child) \
81 for (child = (parent)->children; child != NULL; child = child->next)
83 /* Is the specified node an element of the specified name? */
85 node_is_element(const xmlNode *node, const xmlChar *name)
87 return node->type == XML_ELEMENT_NODE && xmlStrEqual(node->name, name);
90 /* Retrieve a pointer to the UTF-8 text contents of the specified node, or NULL
91 * if the node has no text contents. This assumes the simple case where the
92 * node has a single TEXT child node. */
93 static const xmlChar *
94 node_get_text(const xmlNode *node)
100 node_for_each_child(node, child)
101 if (child->type == XML_TEXT_NODE && child->content)
102 return child->content;
106 /* Retrieve an unsigned integer from the contents of the specified node,
107 * decoding it using the specified base. If the node has no contents or does
108 * not contain a valid number, returns 0. */
110 node_get_number(const xmlNode *node, int base)
112 const xmlChar *str = node_get_text(node);
114 unsigned long long v;
118 v = strtoull(str, &end, base);
119 if ((xmlChar *)end == str || *end || v >= UINT64_MAX)
124 /* Retrieve the timestamp from a time node. This node should have child
125 * elements HIGHPART and LOWPART; these elements will be used to construct a
126 * Windows-style timestamp. */
128 node_get_timestamp(const xmlNode *node)
135 node_for_each_child(node, child) {
136 if (node_is_element(child, "HIGHPART"))
137 timestamp |= node_get_number(child, 16) << 32;
138 else if (node_is_element(child, "LOWPART"))
139 timestamp |= node_get_number(child, 16);
145 tstr_get_utf8(const tchar *tstr, const xmlChar **utf8_ret)
147 if (wimlib_mbs_is_utf8) {
148 *utf8_ret = (xmlChar *)tstr;
151 return tstr_to_utf8_simple(tstr, (char **)utf8_ret);
155 tstr_put_utf8(const xmlChar *utf8)
157 if (!wimlib_mbs_is_utf8)
161 /* Retrieve the text contents of an XML element as a 'tchar' string. If not
162 * found or if the text could not be translated, returns NULL. */
164 node_get_ttext(struct wim_xml_info *info, xmlNode *node)
169 text = node_get_text(node);
171 if (!text || wimlib_mbs_is_utf8)
172 return (const tchar *)text;
174 ttext_p = &info->strings[info->next_string_idx];
175 if (info->num_strings >= ARRAY_LEN(info->strings)) {
179 if (utf8_to_tstr_simple(text, ttext_p))
181 if (info->num_strings < ARRAY_LEN(info->strings))
183 info->next_string_idx++;
184 info->next_string_idx %= ARRAY_LEN(info->strings);
188 /* Unlink the specified node from its parent, then free it (recursively). */
190 unlink_and_free_tree(xmlNode *node)
196 /* Unlink and free (recursively) all children of the specified node. */
198 unlink_and_free_children(xmlNode *node)
202 while ((child = node->last) != NULL)
203 unlink_and_free_tree(child);
206 /* Add the new child element 'replacement' to 'parent', replacing any same-named
207 * element that may already exist. */
209 node_replace_child_element(xmlNode *parent, xmlNode *replacement)
213 node_for_each_child(parent, child) {
214 if (node_is_element(child, replacement->name)) {
215 xmlReplaceNode(child, replacement);
221 xmlAddChild(parent, replacement);
224 /* Set the text contents of the specified element to the specified string,
225 * replacing the existing contents (if any). The string is "raw" and is
226 * permitted to contain characters that have special meaning in XML. */
228 node_set_text(xmlNode *node, const xmlChar *text)
230 xmlNode *text_node = xmlNewText(text);
232 return WIMLIB_ERR_NOMEM;
233 unlink_and_free_children(node);
234 xmlAddChild(node, text_node);
238 /* Like 'node_set_text()', but takes in a 'tchar' string. */
240 node_set_ttext(xmlNode *node, const tchar *ttext)
245 ret = tstr_get_utf8(ttext, &text);
248 ret = node_set_text(node, text);
253 /* Create a new element containing text and optionally link it into a tree. */
255 new_element_with_text(xmlNode *parent, const xmlChar *name, const xmlChar *text)
259 node = xmlNewNode(NULL, name);
263 if (node_set_text(node, text)) {
269 xmlAddChild(parent, node);
273 /* Create a new element containing text and optionally link it into a tree. */
275 new_element_with_ttext(xmlNode *parent, const xmlChar *name, const tchar *ttext,
282 ret = tstr_get_utf8(ttext, &text);
285 node = new_element_with_text(parent, name, text);
288 return WIMLIB_ERR_NOMEM;
294 /* Create a new timestamp element and optionally link it into a tree. */
296 new_element_with_timestamp(xmlNode *parent, const xmlChar *name, u64 timestamp)
301 node = xmlNewNode(NULL, name);
305 sprintf(buf, "0x%08"PRIX32, (u32)(timestamp >> 32));
306 if (!new_element_with_text(node, "HIGHPART", buf))
309 sprintf(buf, "0x%08"PRIX32, (u32)timestamp);
310 if (!new_element_with_text(node, "LOWPART", buf))
314 xmlAddChild(parent, node);
322 /* Create a new number element and optionally link it into a tree. */
324 new_element_with_u64(xmlNode *parent, const xmlChar *name, u64 value)
328 sprintf(buf, "%"PRIu64, value);
329 return new_element_with_text(parent, name, buf);
332 /* Allocate a 'struct wim_xml_info'. The caller is responsible for initializing
333 * the document and the images array. */
334 static struct wim_xml_info *
335 alloc_wim_xml_info(void)
337 struct wim_xml_info *info = MALLOC(sizeof(*info));
339 info->next_string_idx = 0;
340 info->num_strings = 0;
346 do_xml_path_walk(xmlNode *node, const xmlChar *path, bool create,
347 xmlNode **result_ret)
349 size_t n = strlen(path) + 1;
359 /* Copy the path to a temporary buffer. */
360 memcpy(buf, path, n);
365 if (strchr(p, '[')) /* reserved for future use */
373 /* We have another path component. */
375 /* Parse the element name. */
377 while (*p != '/' && *p != '\0')
379 if (p == name) /* empty name? */
384 /* Look for a matching child. */
385 node_for_each_child(node, child)
386 if (node_is_element(child, name))
389 /* No child matched the path. If create=false, the lookup
390 * failed. If create=true, create the needed element. */
393 child = xmlNewChild(node, NULL, name, NULL);
395 return WIMLIB_ERR_NOMEM;
397 /* Continue to the next path component, if there is one. */
406 ERROR("The XML path \"%s\" has invalid syntax.", path);
407 return WIMLIB_ERR_INVALID_PARAM;
410 /* Retrieve the XML element, if any, at the specified 'path'. This supports a
411 * simple filesystem-like syntax. If the element was found, returns a pointer
412 * to it; otherwise returns NULL. */
414 xml_get_node_by_path(xmlNode *root, const xmlChar *path)
417 do_xml_path_walk(root, path, false, &node);
421 /* Similar to xml_get_node_by_path(), but creates the element and any requisite
422 * ancestor elements as needed. If successful, 0 is returned and *node_ret is
423 * set to a pointer to the resulting element. If unsuccessful, an error code is
424 * returned and *node_ret is set to NULL. */
426 xml_ensure_node_by_path(xmlNode *root, const xmlChar *path, xmlNode **node_ret)
428 return do_xml_path_walk(root, path, true, node_ret);
432 xml_get_number_by_path(xmlNode *root, const xmlChar *path)
434 return node_get_number(xml_get_node_by_path(root, path), 10);
438 xml_get_timestamp_by_path(xmlNode *root, const xmlChar *path)
440 return node_get_timestamp(xml_get_node_by_path(root, path));
443 static const xmlChar *
444 xml_get_text_by_path(xmlNode *root, const xmlChar *path)
446 return node_get_text(xml_get_node_by_path(root, path));
450 xml_get_ttext_by_path(struct wim_xml_info *info, xmlNode *root,
453 return node_get_ttext(info, xml_get_node_by_path(root, path));
456 /* Creates/replaces (if ttext is not NULL and not empty) or removes (if ttext is
457 * NULL or empty) an element containing text. */
459 xml_set_ttext_by_path(xmlNode *root, const xmlChar *path, const tchar *ttext)
464 if (ttext && *ttext) {
465 /* Create or replace */
466 ret = xml_ensure_node_by_path(root, path, &node);
469 return node_set_ttext(node, ttext);
472 node = xml_get_node_by_path(root, path);
474 unlink_and_free_tree(node);
479 /* Sets a string property for the specified WIM image. */
481 set_image_property(WIMStruct *wim, int image, const xmlChar *name,
484 struct wim_xml_info *info = wim->xml_info;
486 if (image < 1 || image > info->image_count)
487 return WIMLIB_ERR_INVALID_IMAGE;
489 return xml_set_ttext_by_path(info->images[image - 1], name, value);
492 /* Gets a string property for the specified WIM image as a 'tchar' string.
493 * Returns a pointer to the property value if found; NULL if the image doesn't
494 * exist; or 'default_value' if the property doesn't exist in the image or if
495 * the property value could not be translated to a 'tchar' string. */
497 get_image_property(const WIMStruct *wim, int image, const xmlChar *name,
498 const tchar *default_value)
500 struct wim_xml_info *info = wim->xml_info;
503 if (image < 1 || image > info->image_count)
506 value = xml_get_ttext_by_path(info, info->images[image - 1], name);
507 return value ? value : default_value;
510 /* Unlink and return the node which represents the INDEX attribute of the
511 * specified IMAGE element. */
513 unlink_index_attribute(xmlNode *image_node)
515 xmlAttr *attr = xmlHasProp(image_node, "INDEX");
516 xmlUnlinkNode((xmlNode *)attr);
520 /* Compute the total uncompressed size of the streams of the specified inode. */
522 inode_sum_stream_sizes(const struct wim_inode *inode,
523 const struct blob_table *blob_table)
527 for (unsigned i = 0; i < inode->i_num_streams; i++) {
528 const struct blob_descriptor *blob;
530 blob = stream_blob(&inode->i_streams[i], blob_table);
532 total_size += blob->size;
538 append_image_node(struct wim_xml_info *info, xmlNode *image_node)
543 /* Limit exceeded? */
544 if (unlikely(info->image_count >= MAX_IMAGES))
545 return WIMLIB_ERR_IMAGE_COUNT;
547 /* Add the INDEX attribute. */
548 sprintf(buf, "%d", info->image_count + 1);
549 if (!xmlNewProp(image_node, "INDEX", buf))
550 return WIMLIB_ERR_NOMEM;
552 /* Append the IMAGE element to the 'images' array. */
553 images = REALLOC(info->images,
554 (info->image_count + 1) * sizeof(info->images[0]));
555 if (unlikely(!images))
556 return WIMLIB_ERR_NOMEM;
557 info->images = images;
558 images[info->image_count++] = image_node;
560 /* Add the IMAGE element to the document. */
561 xmlAddChild(info->root, image_node);
565 /*----------------------------------------------------------------------------*
566 * Functions for internal library use *
567 *----------------------------------------------------------------------------*/
569 /* Allocate an empty 'struct wim_xml_info', containing no images. */
570 struct wim_xml_info *
571 xml_new_info_struct(void)
573 struct wim_xml_info *info;
575 info = alloc_wim_xml_info();
579 info->doc = xmlNewDoc("1.0");
583 info->root = xmlNewNode(NULL, "WIM");
586 xmlDocSetRootElement(info->doc, info->root);
589 info->image_count = 0;
593 xmlFreeDoc(info->doc);
600 /* Free a 'struct wim_xml_info'. */
602 xml_free_info_struct(struct wim_xml_info *info)
605 xmlFreeDoc(info->doc);
607 for (size_t i = 0; i < info->num_strings; i++)
608 FREE(info->strings[i]);
613 /* Retrieve the number of images for which there exist IMAGE elements in the XML
616 xml_get_image_count(const struct wim_xml_info *info)
618 return info->image_count;
621 /* Retrieve the TOTALBYTES value for the WIM file, or 0 if this value is
624 xml_get_total_bytes(const struct wim_xml_info *info)
626 return xml_get_number_by_path(info->root, "TOTALBYTES");
629 /* Retrieve the TOTALBYTES value for the specified image, or 0 if this value is
632 xml_get_image_total_bytes(const struct wim_xml_info *info, int image)
634 return xml_get_number_by_path(info->images[image - 1], "TOTALBYTES");
637 /* Retrieve the HARDLINKBYTES value for the specified image, or 0 if this value
640 xml_get_image_hard_link_bytes(const struct wim_xml_info *info, int image)
642 return xml_get_number_by_path(info->images[image - 1], "HARDLINKBYTES");
645 /* Retrieve the WIMBOOT value for the specified image, or false if this value is
648 xml_get_wimboot(const struct wim_xml_info *info, int image)
650 return xml_get_number_by_path(info->images[image - 1], "WIMBOOT");
653 /* Retrieve the Windows build number for the specified image, or 0 if this
654 * information is not available. */
656 xml_get_windows_build_number(const struct wim_xml_info *info, int image)
658 return xml_get_number_by_path(info->images[image - 1],
659 "WINDOWS/VERSION/BUILD");
662 /* Set the WIMBOOT value for the specified image. */
664 xml_set_wimboot(struct wim_xml_info *info, int image)
666 return xml_set_ttext_by_path(info->images[image - 1], "WIMBOOT", T("1"));
670 * Update the DIRCOUNT, FILECOUNT, TOTALBYTES, HARDLINKBYTES, and
671 * LASTMODIFICATIONTIME elements for the specified WIM image.
673 * Note: since these stats are likely to be used for display purposes only, we
674 * no longer attempt to duplicate WIMGAPI's weird bugs when calculating them.
677 xml_update_image_info(WIMStruct *wim, int image)
679 const struct wim_image_metadata *imd = wim->image_metadata[image - 1];
680 xmlNode *image_node = wim->xml_info->images[image - 1];
681 const struct wim_inode *inode;
685 u64 hard_link_bytes = 0;
687 xmlNode *dircount_node;
688 xmlNode *filecount_node;
689 xmlNode *totalbytes_node;
690 xmlNode *hardlinkbytes_node;
691 xmlNode *lastmodificationtime_node;
693 image_for_each_inode(inode, imd) {
694 if (inode_is_directory(inode))
695 dir_count += inode->i_nlink;
697 file_count += inode->i_nlink;
698 size = inode_sum_stream_sizes(inode, wim->blob_table);
699 total_bytes += size * inode->i_nlink;
700 hard_link_bytes += size * (inode->i_nlink - 1);
703 dircount_node = new_element_with_u64(NULL, "DIRCOUNT", dir_count);
704 filecount_node = new_element_with_u64(NULL, "FILECOUNT", file_count);
705 totalbytes_node = new_element_with_u64(NULL, "TOTALBYTES", total_bytes);
706 hardlinkbytes_node = new_element_with_u64(NULL, "HARDLINKBYTES",
708 lastmodificationtime_node =
709 new_element_with_timestamp(NULL, "LASTMODIFICATIONTIME",
710 now_as_wim_timestamp());
712 if (unlikely(!dircount_node || !filecount_node || !totalbytes_node ||
713 !hardlinkbytes_node || !lastmodificationtime_node)) {
714 xmlFreeNode(dircount_node);
715 xmlFreeNode(filecount_node);
716 xmlFreeNode(totalbytes_node);
717 xmlFreeNode(hardlinkbytes_node);
718 xmlFreeNode(lastmodificationtime_node);
719 return WIMLIB_ERR_NOMEM;
722 node_replace_child_element(image_node, dircount_node);
723 node_replace_child_element(image_node, filecount_node);
724 node_replace_child_element(image_node, totalbytes_node);
725 node_replace_child_element(image_node, hardlinkbytes_node);
726 node_replace_child_element(image_node, lastmodificationtime_node);
730 /* Add an image to the XML information. */
732 xml_add_image(struct wim_xml_info *info, const tchar *name)
734 const u64 now = now_as_wim_timestamp();
738 ret = WIMLIB_ERR_NOMEM;
739 image_node = xmlNewNode(NULL, "IMAGE");
744 ret = new_element_with_ttext(image_node, "NAME", name, NULL);
748 ret = WIMLIB_ERR_NOMEM;
749 if (!new_element_with_u64(image_node, "DIRCOUNT", 0))
751 if (!new_element_with_u64(image_node, "FILECOUNT", 0))
753 if (!new_element_with_u64(image_node, "TOTALBYTES", 0))
755 if (!new_element_with_u64(image_node, "HARDLINKBYTES", 0))
757 if (!new_element_with_timestamp(image_node, "CREATIONTIME", now))
759 if (!new_element_with_timestamp(image_node, "LASTMODIFICATIONTIME", now))
761 ret = append_image_node(info, image_node);
767 xmlFreeNode(image_node);
772 * Make a copy of the XML information for the image with index @src_image in the
773 * @src_info XML document and append it to the @dest_info XML document.
775 * In the process, change the image's name and description to the values
776 * specified by @dest_image_name and @dest_image_description. Either or both
777 * may be NULL, which indicates that the corresponding element will not be
778 * included in the destination image.
781 xml_export_image(const struct wim_xml_info *src_info, int src_image,
782 struct wim_xml_info *dest_info, const tchar *dest_image_name,
783 const tchar *dest_image_description, bool wimboot)
788 ret = WIMLIB_ERR_NOMEM;
789 dest_node = xmlDocCopyNode(src_info->images[src_image - 1],
794 ret = xml_set_ttext_by_path(dest_node, "NAME", dest_image_name);
798 ret = xml_set_ttext_by_path(dest_node, "DESCRIPTION",
799 dest_image_description);
804 ret = xml_set_ttext_by_path(dest_node, "WIMBOOT", T("1"));
809 xmlFreeProp(unlink_index_attribute(dest_node));
811 ret = append_image_node(dest_info, dest_node);
817 xmlFreeNode(dest_node);
821 /* Remove the specified image from the XML document. */
823 xml_delete_image(struct wim_xml_info *info, int image)
826 xmlAttr *index_attr, *next_index_attr;
828 /* Free the IMAGE element for the deleted image. Then, shift all
829 * higher-indexed IMAGE elements down by 1, in the process re-assigning
830 * their INDEX attributes. */
832 next_image = info->images[image - 1];
833 next_index_attr = unlink_index_attribute(next_image);
834 unlink_and_free_tree(next_image);
836 while (image < info->image_count) {
837 index_attr = next_index_attr;
838 next_image = info->images[image];
839 next_index_attr = unlink_index_attribute(next_image);
840 xmlAddChild(next_image, (xmlNode *)index_attr);
841 info->images[image - 1] = next_image;
845 xmlFreeProp(next_index_attr);
849 /* Architecture constants are from w64 mingw winnt.h */
850 #define PROCESSOR_ARCHITECTURE_INTEL 0
851 #define PROCESSOR_ARCHITECTURE_MIPS 1
852 #define PROCESSOR_ARCHITECTURE_ALPHA 2
853 #define PROCESSOR_ARCHITECTURE_PPC 3
854 #define PROCESSOR_ARCHITECTURE_SHX 4
855 #define PROCESSOR_ARCHITECTURE_ARM 5
856 #define PROCESSOR_ARCHITECTURE_IA64 6
857 #define PROCESSOR_ARCHITECTURE_ALPHA64 7
858 #define PROCESSOR_ARCHITECTURE_MSIL 8
859 #define PROCESSOR_ARCHITECTURE_AMD64 9
860 #define PROCESSOR_ARCHITECTURE_IA32_ON_WIN64 10
863 describe_arch(u64 arch)
865 static const tchar * const descriptions[] = {
866 [PROCESSOR_ARCHITECTURE_INTEL] = T("x86"),
867 [PROCESSOR_ARCHITECTURE_MIPS] = T("MIPS"),
868 [PROCESSOR_ARCHITECTURE_ARM] = T("ARM"),
869 [PROCESSOR_ARCHITECTURE_IA64] = T("ia64"),
870 [PROCESSOR_ARCHITECTURE_AMD64] = T("x86_64"),
873 if (arch < ARRAY_LEN(descriptions) && descriptions[arch] != NULL)
874 return descriptions[arch];
879 /* Print information from the WINDOWS element, if present. */
881 print_windows_info(struct wim_xml_info *info, xmlNode *image_node)
883 xmlNode *windows_node;
885 xmlNode *version_node;
888 windows_node = xml_get_node_by_path(image_node, "WINDOWS");
892 tprintf(T("Architecture: %"TS"\n"),
893 describe_arch(xml_get_number_by_path(windows_node, "ARCH")));
895 text = xml_get_ttext_by_path(info, windows_node, "PRODUCTNAME");
897 tprintf(T("Product Name: %"TS"\n"), text);
899 text = xml_get_ttext_by_path(info, windows_node, "EDITIONID");
901 tprintf(T("Edition ID: %"TS"\n"), text);
903 text = xml_get_ttext_by_path(info, windows_node, "INSTALLATIONTYPE");
905 tprintf(T("Installation Type: %"TS"\n"), text);
907 text = xml_get_ttext_by_path(info, windows_node, "HAL");
909 tprintf(T("HAL: %"TS"\n"), text);
911 text = xml_get_ttext_by_path(info, windows_node, "PRODUCTTYPE");
913 tprintf(T("Product Type: %"TS"\n"), text);
915 text = xml_get_ttext_by_path(info, windows_node, "PRODUCTSUITE");
917 tprintf(T("Product Suite: %"TS"\n"), text);
919 langs_node = xml_get_node_by_path(windows_node, "LANGUAGES");
923 tprintf(T("Languages: "));
924 node_for_each_child(langs_node, lang_node) {
925 if (!node_is_element(lang_node, "LANGUAGE"))
927 text = node_get_ttext(info, lang_node);
930 tprintf(T("%"TS" "), text);
934 text = xml_get_ttext_by_path(info, langs_node, "DEFAULT");
936 tprintf(T("Default Language: %"TS"\n"), text);
939 text = xml_get_ttext_by_path(info, windows_node, "SYSTEMROOT");
941 tprintf(T("System Root: %"TS"\n"), text);
943 version_node = xml_get_node_by_path(windows_node, "VERSION");
945 tprintf(T("Major Version: %"PRIu64"\n"),
946 xml_get_number_by_path(version_node, "MAJOR"));
947 tprintf(T("Minor Version: %"PRIu64"\n"),
948 xml_get_number_by_path(version_node, "MINOR"));
949 tprintf(T("Build: %"PRIu64"\n"),
950 xml_get_number_by_path(version_node, "BUILD"));
951 tprintf(T("Service Pack Build: %"PRIu64"\n"),
952 xml_get_number_by_path(version_node, "SPBUILD"));
953 tprintf(T("Service Pack Level: %"PRIu64"\n"),
954 xml_get_number_by_path(version_node, "SPLEVEL"));
958 /* Prints information about the specified image. */
960 xml_print_image_info(struct wim_xml_info *info, int image)
962 xmlNode * const image_node = info->images[image - 1];
966 tprintf(T("Index: %d\n"), image);
968 /* Always print the Name and Description, even if the corresponding XML
969 * elements are not present. */
970 text = xml_get_ttext_by_path(info, image_node, "NAME");
971 tprintf(T("Name: %"TS"\n"), text ? text : T(""));
972 text = xml_get_ttext_by_path(info, image_node, "DESCRIPTION");
973 tprintf(T("Description: %"TS"\n"), text ? text : T(""));
975 text = xml_get_ttext_by_path(info, image_node, "DISPLAYNAME");
977 tprintf(T("Display Name: %"TS"\n"), text);
979 text = xml_get_ttext_by_path(info, image_node, "DISPLAYDESCRIPTION");
981 tprintf(T("Display Description: %"TS"\n"), text);
983 tprintf(T("Directory Count: %"PRIu64"\n"),
984 xml_get_number_by_path(image_node, "DIRCOUNT"));
986 tprintf(T("File Count: %"PRIu64"\n"),
987 xml_get_number_by_path(image_node, "FILECOUNT"));
989 tprintf(T("Total Bytes: %"PRIu64"\n"),
990 xml_get_number_by_path(image_node, "TOTALBYTES"));
992 tprintf(T("Hard Link Bytes: %"PRIu64"\n"),
993 xml_get_number_by_path(image_node, "HARDLINKBYTES"));
995 wim_timestamp_to_str(xml_get_timestamp_by_path(image_node,
997 timebuf, ARRAY_LEN(timebuf));
998 tprintf(T("Creation Time: %"TS"\n"), timebuf);
1000 wim_timestamp_to_str(xml_get_timestamp_by_path(image_node,
1001 "LASTMODIFICATIONTIME"),
1002 timebuf, ARRAY_LEN(timebuf));
1003 tprintf(T("Last Modification Time: %"TS"\n"), timebuf);
1005 print_windows_info(info, image_node);
1007 text = xml_get_ttext_by_path(info, image_node, "FLAGS");
1009 tprintf(T("Flags: %"TS"\n"), text);
1011 tprintf(T("WIMBoot compatible: %"TS"\n"),
1012 xml_get_number_by_path(image_node, "WIMBOOT") ?
1013 T("yes") : T("no"));
1018 /*----------------------------------------------------------------------------*
1019 * Reading and writing the XML data *
1020 *----------------------------------------------------------------------------*/
1023 image_node_get_index(const xmlNode *node)
1025 u64 v = node_get_number((const xmlNode *)xmlHasProp(node, "INDEX"), 10);
1026 return min(v, INT_MAX);
1029 /* Prepare the 'images' array from the XML document tree. */
1031 setup_images(struct wim_xml_info *info, xmlNode *root)
1038 info->images = NULL;
1039 info->image_count = 0;
1041 node_for_each_child(root, child) {
1042 if (!node_is_element(child, "IMAGE"))
1044 index = image_node_get_index(child);
1045 if (unlikely(index < 1 || info->image_count >= MAX_IMAGES))
1047 max_index = max(max_index, index);
1048 info->image_count++;
1050 if (unlikely(max_index != info->image_count))
1052 ret = WIMLIB_ERR_NOMEM;
1053 info->images = CALLOC(info->image_count, sizeof(info->images[0]));
1054 if (unlikely(!info->images))
1056 node_for_each_child(root, child) {
1057 if (!node_is_element(child, "IMAGE"))
1059 index = image_node_get_index(child);
1060 if (unlikely(info->images[index - 1]))
1062 info->images[index - 1] = child;
1067 ERROR("The WIM file's XML document does not contain exactly one IMAGE "
1068 "element per image!");
1069 ret = WIMLIB_ERR_XML;
1075 /* Reads the XML data from a WIM file. */
1077 read_wim_xml_data(WIMStruct *wim)
1079 struct wim_xml_info *info;
1086 /* Allocate the 'struct wim_xml_info'. */
1087 ret = WIMLIB_ERR_NOMEM;
1088 info = alloc_wim_xml_info();
1092 /* Read the raw UTF-16LE bytes. */
1093 ret = wimlib_get_xml_data(wim, &buf, &bufsize);
1097 /* Parse the document with libxml2, creating the document tree. */
1098 doc = xmlReadMemory(buf, bufsize, NULL, "UTF-16LE", XML_PARSE_NONET);
1102 ERROR("Unable to parse the WIM file's XML document!");
1103 ret = WIMLIB_ERR_XML;
1107 /* Verify the root element. */
1108 root = xmlDocGetRootElement(doc);
1109 if (!node_is_element(root, "WIM")) {
1110 ERROR("The WIM file's XML document has an unexpected format!");
1111 ret = WIMLIB_ERR_XML;
1115 /* Verify the WIM file is not encrypted. */
1116 if (xml_get_node_by_path(root, "ESD/ENCRYPTED")) {
1117 ret = WIMLIB_ERR_WIM_IS_ENCRYPTED;
1121 /* Validate the image elements and set up the images[] array. */
1122 ret = setup_images(info, root);
1126 /* Save the document and return. */
1129 wim->xml_info = info;
1140 /* Swap the INDEX attributes of two IMAGE elements. */
1142 swap_index_attributes(xmlNode *image_node_1, xmlNode *image_node_2)
1144 xmlAttr *attr_1, *attr_2;
1146 if (image_node_1 != image_node_2) {
1147 attr_1 = unlink_index_attribute(image_node_1);
1148 attr_2 = unlink_index_attribute(image_node_2);
1149 xmlAddChild(image_node_1, (xmlNode *)attr_2);
1150 xmlAddChild(image_node_2, (xmlNode *)attr_1);
1155 prepare_document_for_write(struct wim_xml_info *info, int image, u64 total_bytes,
1156 xmlNode **orig_totalbytes_node_ret)
1158 xmlNode *totalbytes_node = NULL;
1160 /* Allocate the new TOTALBYTES element if needed. */
1161 if (total_bytes != WIM_TOTALBYTES_USE_EXISTING &&
1162 total_bytes != WIM_TOTALBYTES_OMIT) {
1163 totalbytes_node = new_element_with_u64(NULL, "TOTALBYTES",
1165 if (!totalbytes_node)
1166 return WIMLIB_ERR_NOMEM;
1169 /* Adjust the IMAGE elements if needed. */
1170 if (image != WIMLIB_ALL_IMAGES) {
1171 /* We're writing a single image only. Temporarily unlink all
1172 * other IMAGE elements from the document. */
1173 for (int i = 0; i < info->image_count; i++)
1175 xmlUnlinkNode(info->images[i]);
1177 /* Temporarily set the INDEX attribute of the needed IMAGE
1179 swap_index_attributes(info->images[0], info->images[image - 1]);
1182 /* Adjust (add, change, or remove) the TOTALBYTES element if needed. */
1183 *orig_totalbytes_node_ret = NULL;
1184 if (total_bytes != WIM_TOTALBYTES_USE_EXISTING) {
1185 /* Unlink the previous TOTALBYTES element, if any. */
1186 *orig_totalbytes_node_ret = xml_get_node_by_path(info->root,
1188 if (*orig_totalbytes_node_ret)
1189 xmlUnlinkNode(*orig_totalbytes_node_ret);
1191 /* Link in the new TOTALBYTES element, if any. */
1192 if (totalbytes_node)
1193 xmlAddChild(info->root, totalbytes_node);
1199 restore_document_after_write(struct wim_xml_info *info, int image,
1200 xmlNode *orig_totalbytes_node)
1202 /* Restore the IMAGE elements if needed. */
1203 if (image != WIMLIB_ALL_IMAGES) {
1204 /* We wrote a single image only. Re-link all other IMAGE
1205 * elements to the document. */
1206 for (int i = 0; i < info->image_count; i++)
1208 xmlAddChild(info->root, info->images[i]);
1210 /* Restore the original INDEX attributes. */
1211 swap_index_attributes(info->images[0], info->images[image - 1]);
1214 /* Restore the original TOTALBYTES element if needed. */
1215 if (orig_totalbytes_node)
1216 node_replace_child_element(info->root, orig_totalbytes_node);
1220 * Writes the XML data to a WIM file.
1222 * 'image' specifies the image(s) to include in the XML data. Normally it is
1223 * WIMLIB_ALL_IMAGES, but it can also be a 1-based image index.
1225 * 'total_bytes' is the number to use in the top-level TOTALBYTES element, or
1226 * WIM_TOTALBYTES_USE_EXISTING to use the existing value from the XML document
1227 * (if any), or WIM_TOTALBYTES_OMIT to omit the TOTALBYTES element entirely.
1230 write_wim_xml_data(WIMStruct *wim, int image, u64 total_bytes,
1231 struct wim_reshdr *out_reshdr, int write_resource_flags)
1233 struct wim_xml_info *info = wim->xml_info;
1237 xmlNode *orig_totalbytes_node;
1238 xmlSaveCtxt *save_ctx;
1240 /* Make any needed temporary changes to the document. */
1241 ret = prepare_document_for_write(info, image, total_bytes,
1242 &orig_totalbytes_node);
1246 /* Create an in-memory buffer to hold the encoded document. */
1247 ret = WIMLIB_ERR_NOMEM;
1248 buffer = xmlBufferCreate();
1250 goto out_restore_document;
1252 /* Encode the document in UTF-16LE, with a byte order mark, and with no
1253 * XML declaration. Some other WIM software requires all of these
1254 * characteristics. */
1255 ret = WIMLIB_ERR_NOMEM;
1256 if (xmlBufferCat(buffer, "\xff\xfe"))
1257 goto out_free_buffer;
1258 save_ctx = xmlSaveToBuffer(buffer, "UTF-16LE", XML_SAVE_NO_DECL);
1260 goto out_free_buffer;
1261 ret = xmlSaveDoc(save_ctx, info->doc);
1262 ret2 = xmlSaveClose(save_ctx);
1263 if (ret < 0 || ret2 < 0) {
1264 ERROR("Unable to serialize the WIM file's XML document!");
1265 ret = WIMLIB_ERR_NOMEM;
1266 goto out_free_buffer;
1269 /* Write the XML data uncompressed. Although wimlib can handle
1270 * compressed XML data, some other WIM software cannot. */
1271 ret = write_wim_resource_from_buffer(xmlBufferContent(buffer),
1272 xmlBufferLength(buffer),
1275 WIMLIB_COMPRESSION_TYPE_NONE,
1279 write_resource_flags);
1281 xmlBufferFree(buffer);
1282 out_restore_document:
1283 /* Revert any temporary changes we made to the document. */
1284 restore_document_after_write(info, image, orig_totalbytes_node);
1289 /*----------------------------------------------------------------------------*
1290 * Global setup functions *
1291 *----------------------------------------------------------------------------*/
1294 xml_global_init(void)
1300 xml_global_cleanup(void)
1306 xml_set_memory_allocator(void *(*malloc_func)(size_t),
1307 void (*free_func)(void *),
1308 void *(*realloc_func)(void *, size_t))
1310 xmlMemSetup(free_func, malloc_func, realloc_func, wimlib_strdup);
1313 /*----------------------------------------------------------------------------*
1314 * Library API functions *
1315 *----------------------------------------------------------------------------*/
1318 wimlib_get_xml_data(WIMStruct *wim, void **buf_ret, size_t *bufsize_ret)
1320 const struct wim_reshdr *xml_reshdr;
1322 if (wim->filename == NULL && filedes_is_seekable(&wim->in_fd))
1323 return WIMLIB_ERR_NO_FILENAME;
1325 if (buf_ret == NULL || bufsize_ret == NULL)
1326 return WIMLIB_ERR_INVALID_PARAM;
1328 xml_reshdr = &wim->hdr.xml_data_reshdr;
1330 *bufsize_ret = xml_reshdr->uncompressed_size;
1331 return wim_reshdr_to_data(xml_reshdr, wim, buf_ret);
1335 wimlib_extract_xml_data(WIMStruct *wim, FILE *fp)
1341 ret = wimlib_get_xml_data(wim, &buf, &bufsize);
1345 if (fwrite(buf, 1, bufsize, fp) != bufsize) {
1346 ERROR_WITH_ERRNO("Failed to extract XML data");
1347 ret = WIMLIB_ERR_WRITE;
1354 image_name_in_use(const WIMStruct *wim, const tchar *name, int excluded_image)
1356 const struct wim_xml_info *info = wim->xml_info;
1357 const xmlChar *name_utf8;
1360 /* Any number of images can have "no name". */
1361 if (!name || !*name)
1364 /* Check for images that have the specified name. */
1365 if (tstr_get_utf8(name, &name_utf8))
1367 for (int i = 0; i < info->image_count && !found; i++) {
1368 if (i + 1 == excluded_image)
1370 found = xmlStrEqual(name_utf8, xml_get_text_by_path(
1371 info->images[i], "NAME"));
1373 tstr_put_utf8(name_utf8);
1378 wimlib_image_name_in_use(const WIMStruct *wim, const tchar *name)
1380 return image_name_in_use(wim, name, WIMLIB_NO_IMAGE);
1383 WIMLIBAPI const tchar *
1384 wimlib_get_image_name(const WIMStruct *wim, int image)
1386 return get_image_property(wim, image, "NAME", T(""));
1389 WIMLIBAPI const tchar *
1390 wimlib_get_image_description(const WIMStruct *wim, int image)
1392 return get_image_property(wim, image, "DESCRIPTION", NULL);
1395 WIMLIBAPI const tchar *
1396 wimlib_get_image_property(const WIMStruct *wim, int image,
1397 const tchar *property_name)
1399 const xmlChar *name;
1402 if (!property_name || !*property_name)
1404 if (tstr_get_utf8(property_name, &name))
1406 value = get_image_property(wim, image, name, NULL);
1407 tstr_put_utf8(name);
1412 wimlib_set_image_name(WIMStruct *wim, int image, const tchar *name)
1414 if (image_name_in_use(wim, name, image))
1415 return WIMLIB_ERR_IMAGE_NAME_COLLISION;
1417 return set_image_property(wim, image, "NAME", name);
1421 wimlib_set_image_descripton(WIMStruct *wim, int image, const tchar *description)
1423 return set_image_property(wim, image, "DESCRIPTION", description);
1427 wimlib_set_image_flags(WIMStruct *wim, int image, const tchar *flags)
1429 return set_image_property(wim, image, "FLAGS", flags);
1433 wimlib_set_image_property(WIMStruct *wim, int image, const tchar *property_name,
1434 const tchar *property_value)
1436 const xmlChar *name;
1439 if (!property_name || !*property_name)
1440 return WIMLIB_ERR_INVALID_PARAM;
1442 ret = tstr_get_utf8(property_name, &name);
1445 ret = set_image_property(wim, image, name, property_value);
1446 tstr_put_utf8(name);