X-Git-Url: https://wimlib.net/git/?p=wimlib;a=blobdiff_plain;f=src%2Fxml.c;h=8280e8edfc496f014029d3abcb25560c5c89f100;hp=2ac8cea8c71e2a50f1a78bd4534c59eb009dfbb5;hb=52c44c6a506cdfae17fd5fd5c6611622d342c128;hpb=9e56d04309e3e6a896319225288f0c86bd36d34e diff --git a/src/xml.c b/src/xml.c index 2ac8cea8..8280e8ed 100644 --- a/src/xml.c +++ b/src/xml.c @@ -5,1530 +5,1444 @@ */ /* - * Copyright (C) 2012, 2013 Eric Biggers + * Copyright (C) 2012, 2013, 2015 Eric Biggers * - * This file is part of wimlib, a library for working with WIM files. + * This file is free software; you can redistribute it and/or modify it under + * the terms of the GNU Lesser General Public License as published by the Free + * Software Foundation; either version 3 of the License, or (at your option) any + * later version. * - * wimlib is free software; you can redistribute it and/or modify it under the - * terms of the GNU General Public License as published by the Free - * Software Foundation; either version 3 of the License, or (at your option) - * any later version. - * - * wimlib is distributed in the hope that it will be useful, but WITHOUT ANY - * WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR - * A PARTICULAR PURPOSE. See the GNU General Public License for more + * This file is distributed in the hope that it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS + * FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more * details. * - * You should have received a copy of the GNU General Public License - * along with wimlib; if not, see http://www.gnu.org/licenses/. + * You should have received a copy of the GNU Lesser General Public License + * along with this file; if not, see http://www.gnu.org/licenses/. */ -#include "dentry.h" -#include "lookup_table.h" -#include "timestamp.h" -#include "wimlib_internal.h" -#include "xml.h" +#ifdef HAVE_CONFIG_H +# include "config.h" +#endif -#include #include #include -#include -#include -#include - -/* Structures used to form an in-memory representation of the XML data (other - * than the raw parse tree from libxml). */ - -struct windows_version { - u64 major; - u64 minor; - u64 build; - u64 sp_build; - u64 sp_level; -}; +#include +#include -struct windows_info { - u64 arch; - char *product_name; - char *edition_id; - char *installation_type; - char *hal; - char *product_type; - char *product_suite; - char **languages; - char *default_language; - size_t num_languages; - char *system_root; - bool windows_version_exists; - struct windows_version windows_version; -}; +#include "wimlib/blob_table.h" +#include "wimlib/dentry.h" +#include "wimlib/encoding.h" +#include "wimlib/error.h" +#include "wimlib/file_io.h" +#include "wimlib/metadata.h" +#include "wimlib/resource.h" +#include "wimlib/timestamp.h" +#include "wimlib/xml.h" +#include "wimlib/write.h" -struct image_info { - int index; - bool windows_info_exists; - u64 dir_count; - u64 file_count; - u64 total_bytes; - u64 hard_link_bytes; - u64 creation_time; - u64 last_modification_time; - struct windows_info windows_info; - char *name; - char *description; - char *display_name; - char *display_description; - union { - char *flags; - struct wim_lookup_table *lookup_table; - }; -}; +/* + * A wrapper around a WIM file's XML document. The XML document contains + * metadata about each image in the WIM file as well as metadata about the WIM + * file itself. + */ +struct wim_xml_info { + /* The parsed XML document as a libxml2 document tree */ + xmlDocPtr doc; -/* Returns a statically allocated string that is a string representation of the - * architecture number. */ -static const char *get_arch(int arch) -{ - static char buf[20]; - switch (arch) { - case 0: - return "x86"; - case 6: - return "ia64"; - case 9: - return "x86_64"; - /* XXX Are there other arch values? */ - default: - snprintf(buf, sizeof(buf), "%d (unknown)", arch); - return buf; - } -} + /* The root element of the document. This is a cached value, equal to + * xmlDocGetRootElement(doc). */ + xmlNode *root; + /* A malloc()ed array containing a pointer to the IMAGE element for each + * WIM image. The image with 1-based index 'i' is at index 'i - 1' in + * this array. Note: these pointers are cached values, since they could + * also be found by searching the document. */ + xmlNode **images; + + /* The number of WIM images (the length of 'images') */ + int image_count; + + /* Temporary memory for UTF-8 => 'tchar' string translations. When an + * API function needs to return a 'tchar' string, it uses one of these + * array slots to hold the string and returns a pointer to it. */ + tchar *strings[128]; + size_t next_string_idx; + size_t num_strings; +}; -/* Iterate through the children of an xmlNode. */ -#define for_node_child(parent, child) \ - for (child = parent->children; child != NULL; child = child->next) +/*----------------------------------------------------------------------------* + * Internal functions * + *----------------------------------------------------------------------------*/ -/* Utility functions for xmlNodes */ -static inline bool node_is_element(xmlNode *node) -{ - return node->type == XML_ELEMENT_NODE; -} +/* Iterate through the children of an xmlNode. */ +#define node_for_each_child(parent, child) \ + for (child = (parent)->children; child != NULL; child = child->next) -static inline bool node_is_text(xmlNode *node) +/* Is the specified node an element of the specified name? */ +static bool +node_is_element(const xmlNode *node, const xmlChar *name) { - return node->type == XML_TEXT_NODE; + return node->type == XML_ELEMENT_NODE && xmlStrEqual(node->name, name); } -static inline bool node_name_is(xmlNode *node, const char *name) +/* Retrieve a pointer to the UTF-8 text contents of the specified node, or NULL + * if the node has no text contents. This assumes the simple case where the + * node has a single TEXT child node. */ +static const xmlChar * +node_get_text(const xmlNode *node) { - /* For now, both upper case and lower case element names are accepted. */ - return strcasecmp((const char *)node->name, name) == 0; -} + const xmlNode *child; -/* Finds the text node that is a child of an element node and returns its - * content converted to a 64-bit unsigned integer. Returns 0 if no text node is - * found. */ -static u64 node_get_u64(const xmlNode *u64_node) -{ - xmlNode *child; - for_node_child(u64_node, child) - if (node_is_text(child)) - return strtoull((const char *)child->content, NULL, 10); - return 0; + if (!node) + return NULL; + node_for_each_child(node, child) + if (child->type == XML_TEXT_NODE && child->content) + return child->content; + return NULL; } -/* Like node_get_u64(), but expects a number in base 16. */ -static u64 node_get_hex_u64(const xmlNode *u64_node) +/* Retrieve an unsigned integer from the contents of the specified node, + * decoding it using the specified base. If the node has no contents or does + * not contain a valid number, returns 0. */ +static u64 +node_get_number(const xmlNode *node, int base) { - xmlNode *child; - for_node_child(u64_node, child) - if (node_is_text(child)) - return strtoull(child->content, NULL, 16); - return 0; -} + const xmlChar *str = node_get_text(node); + char *end; + unsigned long long v; -static int node_get_string(const xmlNode *string_node, char **str) -{ - xmlNode *child; - char *p = NULL; - - for_node_child(string_node, child) { - if (node_is_text(child) && child->content) { - p = STRDUP(child->content); - if (!p) - return WIMLIB_ERR_NOMEM; - break; - } - } - *str = p; - return 0; + if (!str) + return 0; + v = strtoull(str, &end, base); + if ((xmlChar *)end == str || *end || v >= UINT64_MAX) + return 0; + return v; } -/* Returns the timestamp from a time node. It has child elements and - * that are then used to construct a 64-bit timestamp. */ -static u64 node_get_timestamp(const xmlNode *time_node) +/* Retrieve the timestamp from a time node. This node should have child + * elements HIGHPART and LOWPART; these elements will be used to construct a + * Windows-style timestamp. */ +static u64 +node_get_timestamp(const xmlNode *node) { - u32 high_part = 0; - u32 low_part = 0; + u64 timestamp = 0; xmlNode *child; - for_node_child(time_node, child) { - if (!node_is_element(child)) - continue; - if (node_name_is(child, "HIGHPART")) - high_part = node_get_hex_u64(child); - else if (node_name_is(child, "LOWPART")) - low_part = node_get_hex_u64(child); + + if (!node) + return 0; + node_for_each_child(node, child) { + if (node_is_element(child, "HIGHPART")) + timestamp |= node_get_number(child, 16) << 32; + else if (node_is_element(child, "LOWPART")) + timestamp |= node_get_number(child, 16); } - return (u64)low_part | ((u64)high_part << 32); + return timestamp; } -/* Used to sort an array of struct image_infos by their image indices. */ -static int sort_by_index(const void *p1, const void *p2) +static int +tstr_get_utf8(const tchar *tstr, const xmlChar **utf8_ret) { - int index_1 = ((const struct image_info*)p1)->index; - int index_2 = ((const struct image_info*)p2)->index; - if (index_1 < index_2) - return -1; - else if (index_1 > index_2) - return 1; - else + if (wimlib_mbs_is_utf8) { + *utf8_ret = (xmlChar *)tstr; return 0; + } + return tstr_to_utf8_simple(tstr, (char **)utf8_ret); } - -/* Frees memory allocated inside a struct windows_info structure. */ -static void destroy_windows_info(struct windows_info *windows_info) +static void +tstr_put_utf8(const xmlChar *utf8) { - FREE(windows_info->product_name); - FREE(windows_info->edition_id); - FREE(windows_info->installation_type); - FREE(windows_info->hal); - FREE(windows_info->product_type); - FREE(windows_info->product_suite); - for (size_t i = 0; i < windows_info->num_languages; i++) - FREE(windows_info->languages[i]); - FREE(windows_info->languages); - FREE(windows_info->default_language); - FREE(windows_info->system_root); + if (!wimlib_mbs_is_utf8) + FREE((void *)utf8); } -/* Frees memory allocated inside a struct image_info structure. */ -static void destroy_image_info(struct image_info *image_info) +/* Retrieve the text contents of an XML element as a 'tchar' string. If not + * found or if the text could not be translated, returns NULL. */ +static const tchar * +node_get_ttext(struct wim_xml_info *info, xmlNode *node) { - FREE(image_info->name); - FREE(image_info->description); - FREE(image_info->flags); - FREE(image_info->display_name); - FREE(image_info->display_description); - destroy_windows_info(&image_info->windows_info); - memset(image_info, 0, sizeof(struct image_info)); + const xmlChar *text; + tchar **ttext_p; + + text = node_get_text(node); + + if (!text || wimlib_mbs_is_utf8) + return (const tchar *)text; + + ttext_p = &info->strings[info->next_string_idx]; + if (info->num_strings >= ARRAY_LEN(info->strings)) { + FREE(*ttext_p); + *ttext_p = NULL; + } + if (utf8_to_tstr_simple(text, ttext_p)) + return NULL; + if (info->num_strings < ARRAY_LEN(info->strings)) + info->num_strings++; + info->next_string_idx++; + info->next_string_idx %= ARRAY_LEN(info->strings); + return *ttext_p; } -void free_wim_info(struct wim_info *info) +/* Unlink the specified node from its parent, then free it (recursively). */ +static void +unlink_and_free_tree(xmlNode *node) { - if (info) { - if (info->images) { - for (int i = 0; i < info->num_images; i++) - destroy_image_info(&info->images[i]); - FREE(info->images); - } - FREE(info); - } + xmlUnlinkNode(node); + xmlFreeNode(node); } -/* Reads the information from a element inside the element. - * */ -static void xml_read_windows_version(const xmlNode *version_node, - struct windows_version* windows_version) +/* Unlink and free (recursively) all children of the specified node. */ +static void +unlink_and_free_children(xmlNode *node) { xmlNode *child; - for_node_child(version_node, child) { - if (!node_is_element(child)) - continue; - if (node_name_is(child, "MAJOR")) - windows_version->major = node_get_u64(child); - else if (node_name_is(child, "MINOR")) - windows_version->minor = node_get_u64(child); - else if (node_name_is(child, "BUILD")) - windows_version->build = node_get_u64(child); - else if (node_name_is(child, "SPBUILD")) - windows_version->sp_build = node_get_u64(child); - else if (node_name_is(child, "SPLEVEL")) - windows_version->sp_level = node_get_u64(child); - } + + while ((child = node->last) != NULL) + unlink_and_free_tree(child); } -/* Reads the information from a element inside a element. - * */ -static int xml_read_languages(const xmlNode *languages_node, - char ***languages_ret, - size_t *num_languages_ret, - char **default_language_ret) +/* Add the new child element 'replacement' to 'parent', replacing any same-named + * element that may already exist. */ +static void +node_replace_child_element(xmlNode *parent, xmlNode *replacement) { xmlNode *child; - size_t num_languages = 0; - char **languages; - int ret; - - for_node_child(languages_node, child) - if (node_is_element(child) && node_name_is(child, "LANGUAGE")) - num_languages++; - languages = CALLOC(num_languages, sizeof(languages[0])); - if (!languages) - return WIMLIB_ERR_NOMEM; + node_for_each_child(parent, child) { + if (node_is_element(child, replacement->name)) { + xmlReplaceNode(child, replacement); + xmlFreeNode(child); + return; + } + } - *languages_ret = languages; - *num_languages_ret = num_languages; + xmlAddChild(parent, replacement); +} - ret = 0; - for_node_child(languages_node, child) { - if (!node_is_element(child)) - continue; - if (node_name_is(child, "LANGUAGE")) - ret = node_get_string(child, languages++); - else if (node_name_is(child, "DEFAULT")) - ret = node_get_string(child, default_language_ret); - if (ret != 0) - break; - } - return ret; +/* Set the text contents of the specified element to the specified string, + * replacing the existing contents (if any). The string is "raw" and is + * permitted to contain characters that have special meaning in XML. */ +static int +node_set_text(xmlNode *node, const xmlChar *text) +{ + xmlNode *text_node = xmlNewText(text); + if (!text_node) + return WIMLIB_ERR_NOMEM; + unlink_and_free_children(node); + xmlAddChild(node, text_node); + return 0; } -/* Reads the information from a element inside an element. */ -static int xml_read_windows_info(const xmlNode *windows_node, - struct windows_info *windows_info) +/* Like 'node_set_text()', but takes in a 'tchar' string. */ +static int +node_set_ttext(xmlNode *node, const tchar *ttext) { - xmlNode *child; - int ret = 0; + const xmlChar *text; + int ret; - for_node_child(windows_node, child) { - if (!node_is_element(child)) - continue; - if (node_name_is(child, "ARCH")) { - windows_info->arch = node_get_u64(child); - } else if (node_name_is(child, "PRODUCTNAME")) { - ret = node_get_string(child, - &windows_info->product_name); - } else if (node_name_is(child, "EDITIONID")) { - ret = node_get_string(child, - &windows_info->edition_id); - } else if (node_name_is(child, "INSTALLATIONTYPE")) { - ret = node_get_string(child, - &windows_info->installation_type); - } else if (node_name_is(child, "PRODUCTTYPE")) { - ret = node_get_string(child, - &windows_info->product_type); - } else if (node_name_is(child, "PRODUCTSUITE")) { - ret = node_get_string(child, - &windows_info->product_suite); - } else if (node_name_is(child, "LANGUAGES")) { - ret = xml_read_languages(child, - &windows_info->languages, - &windows_info->num_languages, - &windows_info->default_language); - } else if (node_name_is(child, "VERSION")) { - xml_read_windows_version(child, - &windows_info->windows_version); - windows_info->windows_version_exists = true; - } else if (node_name_is(child, "SYSTEMROOT")) { - ret = node_get_string(child, &windows_info->system_root); - } else if (node_name_is(child, "HAL")) { - ret = node_get_string(child, &windows_info->hal); - } - if (ret != 0) - return ret; - } + ret = tstr_get_utf8(ttext, &text); + if (ret) + return ret; + ret = node_set_text(node, text); + tstr_put_utf8(text); return ret; } -/* Reads the information from an element. */ -static int xml_read_image_info(xmlNode *image_node, - struct image_info *image_info) +/* Create a new element containing text and optionally link it into a tree. */ +static xmlNode * +new_element_with_text(xmlNode *parent, const xmlChar *name, const xmlChar *text) { - xmlNode *child; - xmlChar *index_prop; - int ret; + xmlNode *node; - index_prop = xmlGetProp(image_node, "INDEX"); - if (index_prop) { - image_info->index = atoi(index_prop); - FREE(index_prop); - } else { - image_info->index = 1; - } + node = xmlNewNode(NULL, name); + if (!node) + return NULL; - ret = 0; - for_node_child(image_node, child) { - if (!node_is_element(child)) - continue; - if (node_name_is(child, "DIRCOUNT")) - image_info->dir_count = node_get_u64(child); - else if (node_name_is(child, "FILECOUNT")) - image_info->file_count = node_get_u64(child); - else if (node_name_is(child, "TOTALBYTES")) - image_info->total_bytes = node_get_u64(child); - else if (node_name_is(child, "HARDLINKBYTES")) - image_info->hard_link_bytes = node_get_u64(child); - else if (node_name_is(child, "CREATIONTIME")) - image_info->creation_time = node_get_timestamp(child); - else if (node_name_is(child, "LASTMODIFICATIONTIME")) - image_info->last_modification_time = node_get_timestamp(child); - else if (node_name_is(child, "WINDOWS")) { - DEBUG("Found tag"); - ret = xml_read_windows_info(child, - &image_info->windows_info); - image_info->windows_info_exists = true; - } else if (node_name_is(child, "NAME")) { - ret = node_get_string(child, &image_info->name); - } else if (node_name_is(child, "DESCRIPTION")) { - ret = node_get_string(child, &image_info->description); - } else if (node_name_is(child, "FLAGS")) { - ret = node_get_string(child, &image_info->flags); - } else if (node_name_is(child, "DISPLAYNAME")) { - ret = node_get_string(child, &image_info->display_name); - } else if (node_name_is(child, "DISPLAYDESCRIPTION")) { - ret = node_get_string(child, &image_info->display_description); - } - if (ret != 0) - return ret; - } - if (!image_info->name) { - char *empty_name; - WARNING("Image with index %d has no name", image_info->index); - empty_name = MALLOC(1); - if (!empty_name) - return WIMLIB_ERR_NOMEM; - *empty_name = '\0'; - image_info->name = empty_name; + if (node_set_text(node, text)) { + xmlFreeNode(node); + return NULL; } - return ret; + + if (parent) + xmlAddChild(parent, node); + return node; } -/* Reads the information from a element, which should be the root element - * of the XML tree. */ -static int xml_read_wim_info(const xmlNode *wim_node, - struct wim_info **wim_info_ret) +/* Create a new element containing text and optionally link it into a tree. */ +static int +new_element_with_ttext(xmlNode *parent, const xmlChar *name, const tchar *ttext, + xmlNode **node_ret) { - struct wim_info *wim_info; - xmlNode *child; + const xmlChar *text; int ret; - int num_images; - int i; - - wim_info = CALLOC(1, sizeof(struct wim_info)); - if (!wim_info) + xmlNode *node; + + ret = tstr_get_utf8(ttext, &text); + if (ret) + return ret; + node = new_element_with_text(parent, name, text); + tstr_put_utf8(text); + if (!node) return WIMLIB_ERR_NOMEM; + if (node_ret) + *node_ret = node; + return 0; +} - /* Count how many images there are. */ - num_images = 0; - for_node_child(wim_node, child) { - if (node_is_element(child) && node_name_is(child, "IMAGE")) { - if (num_images == INT_MAX) { - return WIMLIB_ERR_IMAGE_COUNT; - } - num_images++; - } - } +/* Create a new timestamp element and optionally link it into a tree. */ +static xmlNode * +new_element_with_timestamp(xmlNode *parent, const xmlChar *name, u64 timestamp) +{ + xmlNode *node; + char buf[32]; - if (num_images > 0) { - /* Allocate the array of struct image_infos and fill them in. */ - wim_info->images = CALLOC(num_images, sizeof(wim_info->images[0])); - if (!wim_info->images) { - ret = WIMLIB_ERR_NOMEM; - goto err; - } - wim_info->num_images = num_images; - i = 0; - for_node_child(wim_node, child) { - if (!node_is_element(child)) - continue; - if (node_name_is(child, "IMAGE")) { - DEBUG("Found tag"); - ret = xml_read_image_info(child, - &wim_info->images[i]); - if (ret != 0) - goto err; - i++; - } else if (node_name_is(child, "TOTALBYTES")) { - wim_info->total_bytes = node_get_u64(child); - } - } + node = xmlNewNode(NULL, name); + if (!node) + goto err; - /* Sort the array of image info by image index. */ - qsort(wim_info->images, num_images, - sizeof(struct image_info), sort_by_index); - - /* Make sure the image indices make sense */ - for (i = 0; i < num_images; i++) { - if (wim_info->images[i].index != i + 1) { - ERROR("WIM images are not indexed [1...%d] " - "in XML data as expected", - num_images); - return WIMLIB_ERR_IMAGE_COUNT; - } - } + sprintf(buf, "0x%08"PRIX32, (u32)(timestamp >> 32)); + if (!new_element_with_text(node, "HIGHPART", buf)) + goto err; + + sprintf(buf, "0x%08"PRIX32, (u32)timestamp); + if (!new_element_with_text(node, "LOWPART", buf)) + goto err; + + if (parent) + xmlAddChild(parent, node); + return node; - } - *wim_info_ret = wim_info; - return 0; err: - free_wim_info(wim_info); - return ret; + xmlFreeNode(node); + return NULL; } -/* Prints the information contained in a `struct windows_info'. */ -static void print_windows_info(const struct windows_info *windows_info) +/* Create a new number element and optionally link it into a tree. */ +static xmlNode * +new_element_with_u64(xmlNode *parent, const xmlChar *name, u64 value) { - const struct windows_version *windows_version; + char buf[32]; - printf("Architecture: %s\n", get_arch(windows_info->arch)); + sprintf(buf, "%"PRIu64, value); + return new_element_with_text(parent, name, buf); +} - if (windows_info->product_name) - printf("Product Name: %s\n", - windows_info->product_name); +/* Allocate a 'struct wim_xml_info'. The caller is responsible for initializing + * the document and the images array. */ +static struct wim_xml_info * +alloc_wim_xml_info(void) +{ + struct wim_xml_info *info = MALLOC(sizeof(*info)); + if (info) { + info->next_string_idx = 0; + info->num_strings = 0; + } + return info; +} - if (windows_info->edition_id) - printf("Edition ID: %s\n", - windows_info->edition_id); +static int +do_xml_path_walk(xmlNode *node, const xmlChar *path, bool create, + xmlNode **result_ret) +{ + size_t n = strlen(path) + 1; + xmlChar buf[n]; + xmlChar *p; + xmlChar c; - if (windows_info->installation_type) - printf("Installation Type: %s\n", - windows_info->installation_type); + *result_ret = NULL; - if (windows_info->hal) - printf("HAL: %s\n", - windows_info->hal); + if (!node) + return 0; - if (windows_info->product_type) - printf("Product Type: %s\n", - windows_info->product_type); + /* Copy the path to a temporary buffer. */ + memcpy(buf, path, n); + p = buf; + + if (*p == '/') + goto bad_syntax; + if (strchr(p, '[')) /* reserved for future use */ + goto bad_syntax; + c = *p; + + while (c != '\0') { + const xmlChar *name; + xmlNode *child; + + /* We have another path component. */ + + /* Parse the element name. */ + name = p; + while (*p != '/' && *p != '\0') + p++; + if (p == name) /* empty name? */ + goto bad_syntax; + c = *p; + *p = '\0'; + + /* Look for a matching child. */ + node_for_each_child(node, child) + if (node_is_element(child, name)) + goto next_step; + + /* No child matched the path. If create=false, the lookup + * failed. If create=true, create the needed element. */ + if (!create) + return 0; + child = xmlNewChild(node, NULL, name, NULL); + if (!child) + return WIMLIB_ERR_NOMEM; + next_step: + /* Continue to the next path component, if there is one. */ + node = child; + p++; + } - if (windows_info->product_suite) - printf("Product Suite: %s\n", - windows_info->product_suite); + *result_ret = node; + return 0; - printf("Languages: "); - for (size_t i = 0; i < windows_info->num_languages; i++) { - fputs(windows_info->languages[i], stdout); - putchar(' '); - } - putchar('\n'); - if (windows_info->default_language) - printf("Default Language: %s\n", - windows_info->default_language); - if (windows_info->system_root) - printf("System Root: %s\n", - windows_info->system_root); - if (windows_info->windows_version_exists) { - windows_version = &windows_info->windows_version; - printf("Major Version: %"PRIu64"\n", - windows_version->major); - printf("Minor Version: %"PRIu64"\n", - windows_version->minor); - printf("Build: %"PRIu64"\n", - windows_version->build); - printf("Service Pack Build: %"PRIu64"\n", - windows_version->sp_build); - printf("Service Pack Level: %"PRIu64"\n", - windows_version->sp_level); - } +bad_syntax: + ERROR("The XML path \"%s\" has invalid syntax.", path); + return WIMLIB_ERR_INVALID_PARAM; } +/* Retrieve the XML element, if any, at the specified 'path'. This supports a + * simple filesystem-like syntax. If the element was found, returns a pointer + * to it; otherwise returns NULL. */ +static xmlNode * +xml_get_node_by_path(xmlNode *root, const xmlChar *path) +{ + xmlNode *node; + do_xml_path_walk(root, path, false, &node); + return node; +} -/* Writes the information contained in a `struct windows_version' to the XML - * document being written. This is the element inside the - * element. */ -static int xml_write_windows_version(xmlTextWriter *writer, - const struct windows_version *version) +/* Similar to xml_get_node_by_path(), but creates the element and any requisite + * ancestor elements as needed. If successful, 0 is returned and *node_ret is + * set to a pointer to the resulting element. If unsuccessful, an error code is + * returned and *node_ret is set to NULL. */ +static int +xml_ensure_node_by_path(xmlNode *root, const xmlChar *path, xmlNode **node_ret) { - int rc; - rc = xmlTextWriterStartElement(writer, "VERSION"); - if (rc < 0) - return rc; + return do_xml_path_walk(root, path, true, node_ret); +} - rc = xmlTextWriterWriteFormatElement(writer, "MAJOR", "%"PRIu64, - version->major); - if (rc < 0) - return rc; +static u64 +xml_get_number_by_path(xmlNode *root, const xmlChar *path) +{ + return node_get_number(xml_get_node_by_path(root, path), 10); +} - rc = xmlTextWriterWriteFormatElement(writer, "MINOR", "%"PRIu64, - version->minor); - if (rc < 0) - return rc; +static u64 +xml_get_timestamp_by_path(xmlNode *root, const xmlChar *path) +{ + return node_get_timestamp(xml_get_node_by_path(root, path)); +} - rc = xmlTextWriterWriteFormatElement(writer, "BUILD", "%"PRIu64, - version->build); - if (rc < 0) - return rc; +static const xmlChar * +xml_get_text_by_path(xmlNode *root, const xmlChar *path) +{ + return node_get_text(xml_get_node_by_path(root, path)); +} - rc = xmlTextWriterWriteFormatElement(writer, "SPBUILD", "%"PRIu64, - version->sp_build); - if (rc < 0) - return rc; +static const tchar * +xml_get_ttext_by_path(struct wim_xml_info *info, xmlNode *root, + const xmlChar *path) +{ + return node_get_ttext(info, xml_get_node_by_path(root, path)); +} - rc = xmlTextWriterWriteFormatElement(writer, "SPLEVEL", "%"PRIu64, - version->sp_level); - if (rc < 0) - return rc; +/* Creates/replaces (if ttext is not NULL and not empty) or removes (if ttext is + * NULL or empty) an element containing text. */ +static int +xml_set_ttext_by_path(xmlNode *root, const xmlChar *path, const tchar *ttext) +{ + int ret; + xmlNode *node; - return xmlTextWriterEndElement(writer); /* */ + if (ttext && *ttext) { + /* Create or replace */ + ret = xml_ensure_node_by_path(root, path, &node); + if (ret) + return ret; + return node_set_ttext(node, ttext); + } else { + /* Remove */ + node = xml_get_node_by_path(root, path); + if (node) + unlink_and_free_tree(node); + return 0; + } } -/* Writes the information contained in a `struct windows_info' to the XML - * document being written. This is the element. */ -static int xml_write_windows_info(xmlTextWriter *writer, - const struct windows_info *windows_info) +/* Sets a string property for the specified WIM image. */ +static int +set_image_property(WIMStruct *wim, int image, const xmlChar *name, + const tchar *value) { - int rc; - rc = xmlTextWriterStartElement(writer, "WINDOWS"); - if (rc < 0) - return rc; + struct wim_xml_info *info = wim->xml_info; - rc = xmlTextWriterWriteFormatElement(writer, "ARCH", "%"PRIu64, - windows_info->arch); - if (rc < 0) - return rc; + if (image < 1 || image > info->image_count) + return WIMLIB_ERR_INVALID_IMAGE; - if (windows_info->product_name) { - rc = xmlTextWriterWriteElement(writer, "PRODUCTNAME", - windows_info->product_name); - if (rc < 0) - return rc; - } + return xml_set_ttext_by_path(info->images[image - 1], name, value); +} - if (windows_info->edition_id) { - rc = xmlTextWriterWriteElement(writer, "EDITIONID", - windows_info->edition_id); - if (rc < 0) - return rc; - } +/* Gets a string property for the specified WIM image as a 'tchar' string. + * Returns a pointer to the property value if found; NULL if the image doesn't + * exist; or 'default_value' if the property doesn't exist in the image or if + * the property value could not be translated to a 'tchar' string. */ +static const tchar * +get_image_property(const WIMStruct *wim, int image, const xmlChar *name, + const tchar *default_value) +{ + struct wim_xml_info *info = wim->xml_info; + const tchar *value; - if (windows_info->installation_type) { - rc = xmlTextWriterWriteElement(writer, "INSTALLATIONTYPE", - windows_info->installation_type); - if (rc < 0) - return rc; - } + if (image < 1 || image > info->image_count) + return NULL; - if (windows_info->hal) { - rc = xmlTextWriterWriteElement(writer, "HAL", - windows_info->hal); - if (rc < 0) - return rc; - } + value = xml_get_ttext_by_path(info, info->images[image - 1], name); + return value ? value : default_value; +} - if (windows_info->product_type) { - rc = xmlTextWriterWriteElement(writer, "PRODUCTTYPE", - windows_info->product_type); - if (rc < 0) - return rc; - } +/* Unlink and return the node which represents the INDEX attribute of the + * specified IMAGE element. */ +static xmlAttr * +unlink_index_attribute(xmlNode *image_node) +{ + xmlAttr *attr = xmlHasProp(image_node, "INDEX"); + xmlUnlinkNode((xmlNode *)attr); + return attr; +} - if (windows_info->product_suite) { - rc = xmlTextWriterWriteElement(writer, "PRODUCTSUITE", - windows_info->product_suite); - if (rc < 0) - return rc; - } +/* Compute the total uncompressed size of the streams of the specified inode. */ +static u64 +inode_sum_stream_sizes(const struct wim_inode *inode, + const struct blob_table *blob_table) +{ + u64 total_size = 0; - if (windows_info->num_languages) { - rc = xmlTextWriterStartElement(writer, "LANGUAGES"); - if (rc < 0) - return rc; + for (unsigned i = 0; i < inode->i_num_streams; i++) { + const struct blob_descriptor *blob; - for (size_t i = 0; i < windows_info->num_languages; i++) { - rc = xmlTextWriterWriteElement(writer, "LANGUAGE", - windows_info->languages[i]); - if (rc < 0) - return rc; - } - rc = xmlTextWriterWriteElement(writer, "DEFAULT", - windows_info->default_language); - if (rc < 0) - return rc; - - rc = xmlTextWriterEndElement(writer); /* */ - if (rc < 0) - return rc; + blob = stream_blob(&inode->i_streams[i], blob_table); + if (blob) + total_size += blob->size; } + return total_size; +} - if (windows_info->windows_version_exists) { - rc = xml_write_windows_version(writer, &windows_info->windows_version); - if (rc < 0) - return rc; - } +static int +append_image_node(struct wim_xml_info *info, xmlNode *image_node) +{ + char buf[32]; + xmlNode **images; - if (windows_info->system_root) { - rc = xmlTextWriterWriteElement(writer, "SYSTEMROOT", - windows_info->system_root); - if (rc < 0) - return rc; - } + /* Limit exceeded? */ + if (unlikely(info->image_count >= MAX_IMAGES)) + return WIMLIB_ERR_IMAGE_COUNT; + + /* Add the INDEX attribute. */ + sprintf(buf, "%d", info->image_count + 1); + if (!xmlNewProp(image_node, "INDEX", buf)) + return WIMLIB_ERR_NOMEM; + + /* Append the IMAGE element to the 'images' array. */ + images = REALLOC(info->images, + (info->image_count + 1) * sizeof(info->images[0])); + if (unlikely(!images)) + return WIMLIB_ERR_NOMEM; + info->images = images; + images[info->image_count++] = image_node; - return xmlTextWriterEndElement(writer); /* */ + /* Add the IMAGE element to the document. */ + xmlAddChild(info->root, image_node); + return 0; } -/* Writes a time element to the XML document being constructed in memory. */ -static int xml_write_time(xmlTextWriter *writer, const char *element_name, - u64 time) +/*----------------------------------------------------------------------------* + * Functions for internal library use * + *----------------------------------------------------------------------------*/ + +/* Allocate an empty 'struct wim_xml_info', containing no images. */ +struct wim_xml_info * +xml_new_info_struct(void) { - int rc; - rc = xmlTextWriterStartElement(writer, element_name); - if (rc < 0) - return rc; + struct wim_xml_info *info; - rc = xmlTextWriterWriteFormatElement(writer, "HIGHPART", - "0x%08"PRIX32, (u32)(time >> 32)); - if (rc < 0) - return rc; + info = alloc_wim_xml_info(); + if (!info) + goto err; - rc = xmlTextWriterWriteFormatElement(writer, "LOWPART", - "0x%08"PRIX32, (u32)time); - if (rc < 0) - return rc; + info->doc = xmlNewDoc("1.0"); + if (!info->doc) + goto err_free_info; - rc = xmlTextWriterEndElement(writer); /* */ - if (rc < 0) - return rc; - return 0; -} + info->root = xmlNewNode(NULL, "WIM"); + if (!info->root) + goto err_free_doc; + xmlDocSetRootElement(info->doc, info->root); -/* Writes an element to the XML document. */ -static int xml_write_image_info(xmlTextWriter *writer, - const struct image_info *image_info) -{ - int rc; - rc = xmlTextWriterStartElement(writer, "IMAGE"); - if (rc < 0) - return rc; - - rc = xmlTextWriterWriteFormatAttribute(writer, "INDEX", "%d", - image_info->index); - if (rc < 0) - return rc; - - rc = xmlTextWriterWriteFormatElement(writer, "DIRCOUNT", "%"PRIu64, - image_info->dir_count); - if (rc < 0) - return rc; - - rc = xmlTextWriterWriteFormatElement(writer, "FILECOUNT", "%"PRIu64, - image_info->file_count); - if (rc < 0) - return rc; - - rc = xmlTextWriterWriteFormatElement(writer, "TOTALBYTES", "%"PRIu64, - image_info->total_bytes); - if (rc < 0) - return rc; - - rc = xmlTextWriterWriteFormatElement(writer, "HARDLINKBYTES", "%"PRIu64, - image_info->hard_link_bytes); - if (rc < 0) - return rc; - - rc = xml_write_time(writer, "CREATIONTIME", image_info->creation_time); - if (rc < 0) - return rc; - - rc = xml_write_time(writer, "LASTMODIFICATIONTIME", - image_info->last_modification_time); - if (rc < 0) - return rc; - - if (image_info->windows_info_exists) { - rc = xml_write_windows_info(writer, &image_info->windows_info); - if (rc < 0) - return rc; - } + info->images = NULL; + info->image_count = 0; + return info; - if (image_info->name) { - rc = xmlTextWriterWriteElement(writer, "NAME", - image_info->name); - if (rc < 0) - return rc; - } +err_free_doc: + xmlFreeDoc(info->doc); +err_free_info: + FREE(info); +err: + return NULL; +} - if (image_info->description) { - rc = xmlTextWriterWriteElement(writer, "DESCRIPTION", - image_info->description); - if (rc < 0) - return rc; - } - if (image_info->display_name) { - rc = xmlTextWriterWriteElement(writer, "DISPLAYNAME", - image_info->display_name); - if (rc < 0) - return rc; - } - if (image_info->display_description) { - rc = xmlTextWriterWriteElement(writer, "DISPLAYDESCRIPTION", - image_info->display_description); - if (rc < 0) - return rc; +/* Free a 'struct wim_xml_info'. */ +void +xml_free_info_struct(struct wim_xml_info *info) +{ + if (info) { + xmlFreeDoc(info->doc); + FREE(info->images); + for (size_t i = 0; i < info->num_strings; i++) + FREE(info->strings[i]); + FREE(info); } +} - if (image_info->flags) { - rc = xmlTextWriterWriteElement(writer, "FLAGS", - image_info->flags); - if (rc < 0) - return rc; - } +/* Retrieve the number of images for which there exist IMAGE elements in the XML + * document. */ +int +xml_get_image_count(const struct wim_xml_info *info) +{ + return info->image_count; +} - return xmlTextWriterEndElement(writer); /* */ +/* Retrieve the TOTALBYTES value for the WIM file, or 0 if this value is + * unavailable. */ +u64 +xml_get_total_bytes(const struct wim_xml_info *info) +{ + return xml_get_number_by_path(info->root, "TOTALBYTES"); } +/* Retrieve the TOTALBYTES value for the specified image, or 0 if this value is + * unavailable. */ +u64 +xml_get_image_total_bytes(const struct wim_xml_info *info, int image) +{ + return xml_get_number_by_path(info->images[image - 1], "TOTALBYTES"); +} +/* Retrieve the HARDLINKBYTES value for the specified image, or 0 if this value + * is unavailable. */ +u64 +xml_get_image_hard_link_bytes(const struct wim_xml_info *info, int image) +{ + return xml_get_number_by_path(info->images[image - 1], "HARDLINKBYTES"); +} -/* Makes space for another image in the XML information and return a pointer to - * it.*/ -static struct image_info *add_image_info_struct(struct wim_info *wim_info) +/* Retrieve the WIMBOOT value for the specified image, or false if this value is + * unavailable. */ +bool +xml_get_wimboot(const struct wim_xml_info *info, int image) { - struct image_info *images; + return xml_get_number_by_path(info->images[image - 1], "WIMBOOT"); +} - images = CALLOC(wim_info->num_images + 1, sizeof(struct image_info)); - if (!images) - return NULL; - memcpy(images, wim_info->images, - wim_info->num_images * sizeof(struct image_info)); - FREE(wim_info->images); - wim_info->images = images; - wim_info->num_images++; - return &images[wim_info->num_images - 1]; +/* Retrieve the Windows build number for the specified image, or 0 if this + * information is not available. */ +u64 +xml_get_windows_build_number(const struct wim_xml_info *info, int image) +{ + return xml_get_number_by_path(info->images[image - 1], + "WINDOWS/VERSION/BUILD"); } -static int clone_windows_info(const struct windows_info *old, - struct windows_info *new) +/* Set the WIMBOOT value for the specified image. */ +int +xml_set_wimboot(struct wim_xml_info *info, int image) { - if (old->product_name && !(new->product_name = STRDUP(old->product_name))) - return WIMLIB_ERR_NOMEM; - if (old->edition_id && !(new->edition_id = STRDUP(old->edition_id))) - return WIMLIB_ERR_NOMEM; - if (old->installation_type && !(new->installation_type = - STRDUP(old->installation_type))) - return WIMLIB_ERR_NOMEM; - if (old->hal && !(new->hal = STRDUP(old->hal))) - return WIMLIB_ERR_NOMEM; - if (old->product_type && !(new->product_type = STRDUP(old->product_type))) - return WIMLIB_ERR_NOMEM; - if (old->product_suite && !(new->product_suite = STRDUP(old->product_suite))) - return WIMLIB_ERR_NOMEM; + return xml_set_ttext_by_path(info->images[image - 1], "WIMBOOT", T("1")); +} - if (old->languages) { - new->languages = CALLOC(old->num_languages, sizeof(char*)); - if (!new->languages) - return WIMLIB_ERR_NOMEM; - new->num_languages = old->num_languages; - for (size_t i = 0; i < new->num_languages; i++) { - if (!old->languages[i]) - continue; - new->languages[i] = STRDUP(old->languages[i]); - if (!new->languages[i]) - return WIMLIB_ERR_NOMEM; - } +/* + * Update the DIRCOUNT, FILECOUNT, TOTALBYTES, HARDLINKBYTES, and + * LASTMODIFICATIONTIME elements for the specified WIM image. + * + * Note: since these stats are likely to be used for display purposes only, we + * no longer attempt to duplicate WIMGAPI's weird bugs when calculating them. + */ +int +xml_update_image_info(WIMStruct *wim, int image) +{ + const struct wim_image_metadata *imd = wim->image_metadata[image - 1]; + xmlNode *image_node = wim->xml_info->images[image - 1]; + const struct wim_inode *inode; + u64 dir_count = 0; + u64 file_count = 0; + u64 total_bytes = 0; + u64 hard_link_bytes = 0; + u64 size; + xmlNode *dircount_node; + xmlNode *filecount_node; + xmlNode *totalbytes_node; + xmlNode *hardlinkbytes_node; + xmlNode *lastmodificationtime_node; + + image_for_each_inode(inode, imd) { + if (inode_is_directory(inode)) + dir_count += inode->i_nlink; + else + file_count += inode->i_nlink; + size = inode_sum_stream_sizes(inode, wim->blob_table); + total_bytes += size * inode->i_nlink; + hard_link_bytes += size * (inode->i_nlink - 1); } - if (old->default_language && - !(new->default_language = STRDUP(old->default_language))) - return WIMLIB_ERR_NOMEM; - if (old->system_root && !(new->system_root = STRDUP(old->system_root))) + + dircount_node = new_element_with_u64(NULL, "DIRCOUNT", dir_count); + filecount_node = new_element_with_u64(NULL, "FILECOUNT", file_count); + totalbytes_node = new_element_with_u64(NULL, "TOTALBYTES", total_bytes); + hardlinkbytes_node = new_element_with_u64(NULL, "HARDLINKBYTES", + hard_link_bytes); + lastmodificationtime_node = + new_element_with_timestamp(NULL, "LASTMODIFICATIONTIME", + now_as_wim_timestamp()); + + if (unlikely(!dircount_node || !filecount_node || !totalbytes_node || + !hardlinkbytes_node || !lastmodificationtime_node)) { + xmlFreeNode(dircount_node); + xmlFreeNode(filecount_node); + xmlFreeNode(totalbytes_node); + xmlFreeNode(hardlinkbytes_node); + xmlFreeNode(lastmodificationtime_node); return WIMLIB_ERR_NOMEM; - if (old->windows_version_exists) { - new->windows_version_exists = true; - memcpy(&new->windows_version, &old->windows_version, - sizeof(old->windows_version)); } + + node_replace_child_element(image_node, dircount_node); + node_replace_child_element(image_node, filecount_node); + node_replace_child_element(image_node, totalbytes_node); + node_replace_child_element(image_node, hardlinkbytes_node); + node_replace_child_element(image_node, lastmodificationtime_node); return 0; } -static int clone_image_info(const struct image_info *old, struct image_info *new) +/* Add an image to the XML information. */ +int +xml_add_image(struct wim_xml_info *info, const tchar *name) { - new->dir_count = old->dir_count; - new->file_count = old->file_count; - new->total_bytes = old->total_bytes; - new->hard_link_bytes = old->hard_link_bytes; - new->creation_time = old->creation_time; - new->last_modification_time = old->last_modification_time; - - if (!(new->name = STRDUP(old->name))) - return WIMLIB_ERR_NOMEM; - - if (old->description) - if (!(new->description = STRDUP(old->description))) - return WIMLIB_ERR_NOMEM; - - if (old->display_name) - if (!(new->display_name = STRDUP(old->display_name))) - return WIMLIB_ERR_NOMEM; - - if (old->display_description) - if (!(new->display_description = STRDUP(old->display_description))) - return WIMLIB_ERR_NOMEM; + const u64 now = now_as_wim_timestamp(); + xmlNode *image_node; + int ret; - if (old->flags) - if (!(new->flags = STRDUP(old->flags))) - return WIMLIB_ERR_NOMEM; + ret = WIMLIB_ERR_NOMEM; + image_node = xmlNewNode(NULL, "IMAGE"); + if (!image_node) + goto err; - if (old->windows_info_exists) { - new->windows_info_exists = true; - return clone_windows_info(&old->windows_info, - &new->windows_info); + if (name && *name) { + ret = new_element_with_ttext(image_node, "NAME", name, NULL); + if (ret) + goto err; } + ret = WIMLIB_ERR_NOMEM; + if (!new_element_with_u64(image_node, "DIRCOUNT", 0)) + goto err; + if (!new_element_with_u64(image_node, "FILECOUNT", 0)) + goto err; + if (!new_element_with_u64(image_node, "TOTALBYTES", 0)) + goto err; + if (!new_element_with_u64(image_node, "HARDLINKBYTES", 0)) + goto err; + if (!new_element_with_timestamp(image_node, "CREATIONTIME", now)) + goto err; + if (!new_element_with_timestamp(image_node, "LASTMODIFICATIONTIME", now)) + goto err; + ret = append_image_node(info, image_node); + if (ret) + goto err; return 0; + +err: + xmlFreeNode(image_node); + return ret; } -/* Copies the XML information for an image between WIM files. - * - * @dest_image_name and @dest_image_description are ignored if they are NULL; - * otherwise, they are used to override the image name and/or image description - * from the XML data in the source WIM file. +/* + * Make a copy of the XML information for the image with index @src_image in the + * @src_info XML document and append it to the @dest_info XML document. * - * On failure, WIMLIB_ERR_NOMEM is returned and no changes are made. Otherwise, - * 0 is returned and the WIM information at *new_wim_info_p is modified. + * In the process, change the image's name and description to the values + * specified by @dest_image_name and @dest_image_description. Either or both + * may be NULL, which indicates that the corresponding element will not be + * included in the destination image. */ -int xml_export_image(const struct wim_info *old_wim_info, - int image, - struct wim_info **new_wim_info_p, - const char *dest_image_name, - const char *dest_image_description) -{ - struct wim_info *new_wim_info; - struct image_info *image_info; +int +xml_export_image(const struct wim_xml_info *src_info, int src_image, + struct wim_xml_info *dest_info, const tchar *dest_image_name, + const tchar *dest_image_description, bool wimboot) +{ + xmlNode *dest_node; int ret; - DEBUG("Copying XML data between WIM files for source image %d.", image); - - wimlib_assert(old_wim_info != NULL); - wimlib_assert(image >= 1 && image <= old_wim_info->num_images); + ret = WIMLIB_ERR_NOMEM; + dest_node = xmlDocCopyNode(src_info->images[src_image - 1], + dest_info->doc, 1); + if (!dest_node) + goto err; - if (*new_wim_info_p) { - new_wim_info = *new_wim_info_p; - } else { - new_wim_info = CALLOC(1, sizeof(struct wim_info)); - if (!new_wim_info) - goto err; - } + ret = xml_set_ttext_by_path(dest_node, "NAME", dest_image_name); + if (ret) + goto err; - image_info = add_image_info_struct(new_wim_info); - if (!image_info) + ret = xml_set_ttext_by_path(dest_node, "DESCRIPTION", + dest_image_description); + if (ret) goto err; - ret = clone_image_info(&old_wim_info->images[image - 1], image_info); - if (ret != 0) - goto err_destroy_image_info; + if (wimboot) { + ret = xml_set_ttext_by_path(dest_node, "WIMBOOT", T("1")); + if (ret) + goto err; + } - image_info->index = new_wim_info->num_images; + xmlFreeProp(unlink_index_attribute(dest_node)); - if (dest_image_name) { - FREE(image_info->name); - image_info->name = STRDUP(dest_image_name); - if (!image_info->name) - goto err_destroy_image_info; - } - if (dest_image_description) { - FREE(image_info->description); - image_info->description = STRDUP(dest_image_description); - if (!image_info->description) - goto err_destroy_image_info; - } - *new_wim_info_p = new_wim_info; + ret = append_image_node(dest_info, dest_node); + if (ret) + goto err; return 0; -err_destroy_image_info: - destroy_image_info(image_info); + err: - if (new_wim_info != *new_wim_info_p) - free_wim_info(new_wim_info); - return WIMLIB_ERR_NOMEM; + xmlFreeNode(dest_node); + return ret; } -/* Removes an image from the XML information. */ -void xml_delete_image(struct wim_info **wim_info_p, int image) +/* Remove the specified image from the XML document. */ +void +xml_delete_image(struct wim_xml_info *info, int image) { - struct wim_info *wim_info; - - DEBUG("Deleting image %d from the XML data.", image); + xmlNode *next_image; + xmlAttr *index_attr, *next_index_attr; + + /* Free the IMAGE element for the deleted image. Then, shift all + * higher-indexed IMAGE elements down by 1, in the process re-assigning + * their INDEX attributes. */ + + next_image = info->images[image - 1]; + next_index_attr = unlink_index_attribute(next_image); + unlink_and_free_tree(next_image); + + while (image < info->image_count) { + index_attr = next_index_attr; + next_image = info->images[image]; + next_index_attr = unlink_index_attribute(next_image); + xmlAddChild(next_image, (xmlNode *)index_attr); + info->images[image - 1] = next_image; + image++; + } - wim_info = *wim_info_p; + xmlFreeProp(next_index_attr); + info->image_count--; +} - destroy_image_info(&wim_info->images[image - 1]); +/* Architecture constants are from w64 mingw winnt.h */ +#define PROCESSOR_ARCHITECTURE_INTEL 0 +#define PROCESSOR_ARCHITECTURE_MIPS 1 +#define PROCESSOR_ARCHITECTURE_ALPHA 2 +#define PROCESSOR_ARCHITECTURE_PPC 3 +#define PROCESSOR_ARCHITECTURE_SHX 4 +#define PROCESSOR_ARCHITECTURE_ARM 5 +#define PROCESSOR_ARCHITECTURE_IA64 6 +#define PROCESSOR_ARCHITECTURE_ALPHA64 7 +#define PROCESSOR_ARCHITECTURE_MSIL 8 +#define PROCESSOR_ARCHITECTURE_AMD64 9 +#define PROCESSOR_ARCHITECTURE_IA32_ON_WIN64 10 + +static const tchar * +describe_arch(u64 arch) +{ + static const tchar * const descriptions[] = { + [PROCESSOR_ARCHITECTURE_INTEL] = T("x86"), + [PROCESSOR_ARCHITECTURE_MIPS] = T("MIPS"), + [PROCESSOR_ARCHITECTURE_ARM] = T("ARM"), + [PROCESSOR_ARCHITECTURE_IA64] = T("ia64"), + [PROCESSOR_ARCHITECTURE_AMD64] = T("x86_64"), + }; - memmove(&wim_info->images[image - 1], - &wim_info->images[image], - (wim_info->num_images - image) * sizeof(struct image_info)); + if (arch < ARRAY_LEN(descriptions) && descriptions[arch] != NULL) + return descriptions[arch]; - if (--wim_info->num_images == 0) { - free_wim_info(wim_info); - *wim_info_p = NULL; - } else { - for (int i = image - 1; i < wim_info->num_images; i++) - wim_info->images[i].index--; - } + return T("unknown"); } -size_t xml_get_max_image_name_len(const WIMStruct *w) +/* Print information from the WINDOWS element, if present. */ +static void +print_windows_info(struct wim_xml_info *info, xmlNode *image_node) { - size_t max_len = 0; - if (w->wim_info) { - size_t len; - for (int i = 0; i < w->wim_info->num_images; i++) { - len = strlen(w->wim_info->images[i].name); - if (len > max_len) - max_len = len; - } - } - return max_len; -} + xmlNode *windows_node; + xmlNode *langs_node; + xmlNode *version_node; + const tchar *text; -#ifdef ENABLE_CUSTOM_MEMORY_ALLOCATOR -void xml_set_memory_allocator(void *(*malloc_func)(size_t), - void (*free_func)(void *), - void *(*realloc_func)(void *, size_t)) -{ - xmlMemSetup(free_func, malloc_func, realloc_func, STRDUP); -} -#endif + windows_node = xml_get_node_by_path(image_node, "WINDOWS"); + if (!windows_node) + return; -static int calculate_dentry_statistics(struct wim_dentry *dentry, void *arg) -{ - struct image_info *info = arg; - struct wim_lookup_table *lookup_table = info->lookup_table; - const struct wim_inode *inode = dentry->d_inode; - struct wim_lookup_table_entry *lte; - - /* Update directory count and file count. - * - * Each dentry counts as either a file or a directory, but not both. - * The root directory is an exception: it is not counted at all. - * - * Symbolic links and junction points (and presumably other reparse - * points) count as regular files. This is despite the fact that - * junction points have FILE_ATTRIBUTE_DIRECTORY set. - */ - if (dentry_is_root(dentry)) - return 0; + tprintf(T("Architecture: %"TS"\n"), + describe_arch(xml_get_number_by_path(windows_node, "ARCH"))); - if (inode_is_directory(inode)) - info->dir_count++; - else - info->file_count++; - - /* - * Update total bytes and hard link bytes. - * - * Unfortunately there are some inconsistencies/bugs in the way this is - * done. - * - * If there are no alternate data streams in the image, the "total - * bytes" is the sum of the size of the un-named data stream of each - * inode times the link count of that inode. In other words, it would - * be the total number of bytes of regular files you would have if you - * extracted the full image without any hard-links. The "hard link - * bytes" is equal to the "total bytes" minus the size of the un-named - * data stream of each inode. In other words, the "hard link bytes" - * counts the size of the un-named data stream for all the links to each - * inode except the first one. - * - * Reparse points and directories don't seem to be counted in either the - * total bytes or the hard link bytes. - * - * And now we get to the most confusing part, the alternate data - * streams. They are not counted in the "total bytes". However, if the - * link count of an inode with alternate data streams is 2 or greater, - * the size of all the alternate data streams is included in the "hard - * link bytes", and this size is multiplied by the link count (NOT one - * less than the link count). - */ - lte = inode_unnamed_lte(inode, info->lookup_table); - if (lte) { - info->total_bytes += wim_resource_size(lte); - if (!dentry_is_first_in_inode(dentry)) - info->hard_link_bytes += wim_resource_size(lte); - } + text = xml_get_ttext_by_path(info, windows_node, "PRODUCTNAME"); + if (text) + tprintf(T("Product Name: %"TS"\n"), text); - if (inode->i_nlink >= 2 && dentry_is_first_in_inode(dentry)) { - for (unsigned i = 0; i < inode->i_num_ads; i++) { - if (inode->i_ads_entries[i].stream_name_len) { - lte = inode_stream_lte(inode, i + 1, lookup_table); - if (lte) { - info->hard_link_bytes += inode->i_nlink * - wim_resource_size(lte); - } - } - } - } - return 0; -} + text = xml_get_ttext_by_path(info, windows_node, "EDITIONID"); + if (text) + tprintf(T("Edition ID: %"TS"\n"), text); -/* - * Calculate what to put in the , , , and - * elements of each . - * - * Please note there is no official documentation for exactly how this is done. - * But, see calculate_dentry_statistics(). - */ -void xml_update_image_info(WIMStruct *w, int image) -{ - struct image_info *image_info; - char *flags_save; + text = xml_get_ttext_by_path(info, windows_node, "INSTALLATIONTYPE"); + if (text) + tprintf(T("Installation Type: %"TS"\n"), text); - DEBUG("Updating the image info for image %d", image); + text = xml_get_ttext_by_path(info, windows_node, "HAL"); + if (text) + tprintf(T("HAL: %"TS"\n"), text); - image_info = &w->wim_info->images[image - 1]; + text = xml_get_ttext_by_path(info, windows_node, "PRODUCTTYPE"); + if (text) + tprintf(T("Product Type: %"TS"\n"), text); - image_info->file_count = 0; - image_info->dir_count = 0; - image_info->total_bytes = 0; - image_info->hard_link_bytes = 0; + text = xml_get_ttext_by_path(info, windows_node, "PRODUCTSUITE"); + if (text) + tprintf(T("Product Suite: %"TS"\n"), text); - flags_save = image_info->flags; - image_info->lookup_table = w->lookup_table; - for_dentry_in_tree(w->image_metadata[image - 1].root_dentry, - calculate_dentry_statistics, - image_info); - image_info->flags = flags_save; - image_info->last_modification_time = get_wim_timestamp(); -} + langs_node = xml_get_node_by_path(windows_node, "LANGUAGES"); + if (langs_node) { + xmlNode *lang_node; -/* Adds an image to the XML information. */ -int xml_add_image(WIMStruct *w, const char *name) -{ - struct wim_info *wim_info; - struct image_info *image_info; + tprintf(T("Languages: ")); + node_for_each_child(langs_node, lang_node) { + if (!node_is_element(lang_node, "LANGUAGE")) + continue; + text = node_get_ttext(info, lang_node); + if (!text) + continue; + tprintf(T("%"TS" "), text); + } + tputchar(T('\n')); - wimlib_assert(name != NULL); + text = xml_get_ttext_by_path(info, langs_node, "DEFAULT"); + if (text) + tprintf(T("Default Language: %"TS"\n"), text); + } - /* If this is the first image, allocate the struct wim_info. Otherwise - * use the existing struct wim_info. */ - if (w->wim_info) { - wim_info = w->wim_info; - } else { - wim_info = CALLOC(1, sizeof(struct wim_info)); - if (!wim_info) - return WIMLIB_ERR_NOMEM; + text = xml_get_ttext_by_path(info, windows_node, "SYSTEMROOT"); + if (text) + tprintf(T("System Root: %"TS"\n"), text); + + version_node = xml_get_node_by_path(windows_node, "VERSION"); + if (version_node) { + tprintf(T("Major Version: %"PRIu64"\n"), + xml_get_number_by_path(version_node, "MAJOR")); + tprintf(T("Minor Version: %"PRIu64"\n"), + xml_get_number_by_path(version_node, "MINOR")); + tprintf(T("Build: %"PRIu64"\n"), + xml_get_number_by_path(version_node, "BUILD")); + tprintf(T("Service Pack Build: %"PRIu64"\n"), + xml_get_number_by_path(version_node, "SPBUILD")); + tprintf(T("Service Pack Level: %"PRIu64"\n"), + xml_get_number_by_path(version_node, "SPLEVEL")); } +} - image_info = add_image_info_struct(wim_info); - if (!image_info) - goto out_free_wim_info; +/* Prints information about the specified image. */ +void +xml_print_image_info(struct wim_xml_info *info, int image) +{ + xmlNode * const image_node = info->images[image - 1]; + const tchar *text; + tchar timebuf[64]; - if (!(image_info->name = STRDUP(name))) - goto out_destroy_image_info; + tprintf(T("Index: %d\n"), image); - w->wim_info = wim_info; - image_info->index = wim_info->num_images; - image_info->creation_time = get_wim_timestamp(); - xml_update_image_info(w, image_info->index); - return 0; + /* Always print the Name and Description, even if the corresponding XML + * elements are not present. */ + text = xml_get_ttext_by_path(info, image_node, "NAME"); + tprintf(T("Name: %"TS"\n"), text ? text : T("")); + text = xml_get_ttext_by_path(info, image_node, "DESCRIPTION"); + tprintf(T("Description: %"TS"\n"), text ? text : T("")); -out_destroy_image_info: - destroy_image_info(image_info); - wim_info->num_images--; -out_free_wim_info: - if (wim_info != w->wim_info) - FREE(wim_info); - return WIMLIB_ERR_NOMEM; -} + text = xml_get_ttext_by_path(info, image_node, "DISPLAYNAME"); + if (text) + tprintf(T("Display Name: %"TS"\n"), text); -/* Prints information about the specified image from struct wim_info structure. - * */ -void print_image_info(const struct wim_info *wim_info, int image) -{ - const struct image_info *image_info; - const char *desc; - char buf[50]; + text = xml_get_ttext_by_path(info, image_node, "DISPLAYDESCRIPTION"); + if (text) + tprintf(T("Display Description: %"TS"\n"), text); - wimlib_assert(image >= 1 && image <= wim_info->num_images); + tprintf(T("Directory Count: %"PRIu64"\n"), + xml_get_number_by_path(image_node, "DIRCOUNT")); - image_info = &wim_info->images[image - 1]; + tprintf(T("File Count: %"PRIu64"\n"), + xml_get_number_by_path(image_node, "FILECOUNT")); - printf("Index: %d\n", image_info->index); - printf("Name: %s\n", image_info->name); + tprintf(T("Total Bytes: %"PRIu64"\n"), + xml_get_number_by_path(image_node, "TOTALBYTES")); - /* Always print the Description: part even if there is no - * description. */ - if (image_info->description) - desc = image_info->description; - else - desc = ""; - printf("Description: %s\n", desc); + tprintf(T("Hard Link Bytes: %"PRIu64"\n"), + xml_get_number_by_path(image_node, "HARDLINKBYTES")); - if (image_info->display_name) - printf("Display Name: %s\n", - image_info->display_name); + wim_timestamp_to_str(xml_get_timestamp_by_path(image_node, + "CREATIONTIME"), + timebuf, ARRAY_LEN(timebuf)); + tprintf(T("Creation Time: %"TS"\n"), timebuf); - if (image_info->display_description) - printf("Display Description: %s\n", - image_info->display_description); + wim_timestamp_to_str(xml_get_timestamp_by_path(image_node, + "LASTMODIFICATIONTIME"), + timebuf, ARRAY_LEN(timebuf)); + tprintf(T("Last Modification Time: %"TS"\n"), timebuf); - printf("Directory Count: %"PRIu64"\n", image_info->dir_count); - printf("File Count: %"PRIu64"\n", image_info->file_count); - printf("Total Bytes: %"PRIu64"\n", image_info->total_bytes); - printf("Hard Link Bytes: %"PRIu64"\n", image_info->hard_link_bytes); + print_windows_info(info, image_node); - wim_timestamp_to_str(image_info->creation_time, buf, sizeof(buf)); - printf("Creation Time: %s\n", buf); + text = xml_get_ttext_by_path(info, image_node, "FLAGS"); + if (text) + tprintf(T("Flags: %"TS"\n"), text); - wim_timestamp_to_str(image_info->creation_time, buf, sizeof(buf)); - printf("Last Modification Time: %s\n", buf); - if (image_info->windows_info_exists) - print_windows_info(&image_info->windows_info); - if (image_info->flags) - printf("Flags: %s\n", image_info->flags); - putchar('\n'); + tprintf(T("WIMBoot compatible: %"TS"\n"), + xml_get_number_by_path(image_node, "WIMBOOT") ? + T("yes") : T("no")); + + tputchar('\n'); } -void libxml_global_init() +/*----------------------------------------------------------------------------* + * Reading and writing the XML data * + *----------------------------------------------------------------------------*/ + +static int +image_node_get_index(const xmlNode *node) { - xmlInitParser(); + u64 v = node_get_number((const xmlNode *)xmlHasProp(node, "INDEX"), 10); + return min(v, INT_MAX); } -void libxml_global_cleanup() +/* Prepare the 'images' array from the XML document tree. */ +static int +setup_images(struct wim_xml_info *info, xmlNode *root) { - xmlCleanupParser(); + xmlNode *child; + int index; + int max_index = 0; + int ret; + + info->images = NULL; + info->image_count = 0; + + node_for_each_child(root, child) { + if (!node_is_element(child, "IMAGE")) + continue; + index = image_node_get_index(child); + if (unlikely(index < 1 || info->image_count >= MAX_IMAGES)) + goto err_indices; + max_index = max(max_index, index); + info->image_count++; + } + if (unlikely(max_index != info->image_count)) + goto err_indices; + ret = WIMLIB_ERR_NOMEM; + info->images = CALLOC(info->image_count, sizeof(info->images[0])); + if (unlikely(!info->images)) + goto err; + node_for_each_child(root, child) { + if (!node_is_element(child, "IMAGE")) + continue; + index = image_node_get_index(child); + if (unlikely(info->images[index - 1])) + goto err_indices; + info->images[index - 1] = child; + } + return 0; + +err_indices: + ERROR("The WIM file's XML document does not contain exactly one IMAGE " + "element per image!"); + ret = WIMLIB_ERR_XML; +err: + FREE(info->images); + return ret; } -/* - * Reads the XML data from a WIM file. - */ -int read_xml_data(FILE *fp, const struct resource_entry *res_entry, - u8 **xml_data_ret, struct wim_info **info_ret) +/* Reads the XML data from a WIM file. */ +int +read_wim_xml_data(WIMStruct *wim) { - u8 *xml_data; + struct wim_xml_info *info; + void *buf; + size_t bufsize; xmlDoc *doc; xmlNode *root; int ret; - DEBUG("XML data is %"PRIu64" bytes at offset %"PRIu64"", - (u64)res_entry->size, res_entry->offset); + /* Allocate the 'struct wim_xml_info'. */ + ret = WIMLIB_ERR_NOMEM; + info = alloc_wim_xml_info(); + if (!info) + goto err; + + /* Read the raw UTF-16LE bytes. */ + ret = wimlib_get_xml_data(wim, &buf, &bufsize); + if (ret) + goto err_free_info; - if (resource_is_compressed(res_entry)) { - ERROR("XML data is supposed to be uncompressed"); + /* Parse the document with libxml2, creating the document tree. */ + doc = xmlReadMemory(buf, bufsize, NULL, "UTF-16LE", XML_PARSE_NONET); + FREE(buf); + buf = NULL; + if (!doc) { + ERROR("Unable to parse the WIM file's XML document!"); ret = WIMLIB_ERR_XML; - goto out; + goto err_free_info; } - if (res_entry->size < 2) { - ERROR("XML data must be at least 2 bytes long"); + /* Verify the root element. */ + root = xmlDocGetRootElement(doc); + if (!node_is_element(root, "WIM")) { + ERROR("The WIM file's XML document has an unexpected format!"); ret = WIMLIB_ERR_XML; - goto out; + goto err_free_doc; } - xml_data = MALLOC(res_entry->size + 2); - if (!xml_data) { - ret = WIMLIB_ERR_NOMEM; - goto out; + /* Verify the WIM file is not encrypted. */ + if (xml_get_node_by_path(root, "ESD/ENCRYPTED")) { + ret = WIMLIB_ERR_WIM_IS_ENCRYPTED; + goto err_free_doc; } - ret = read_uncompressed_resource(fp, res_entry->offset, - res_entry->size, xml_data); - if (ret != 0) - goto out_free_xml_data; + /* Validate the image elements and set up the images[] array. */ + ret = setup_images(info, root); + if (ret) + goto err_free_doc; - /* Null-terminate just in case */ - xml_data[res_entry->size] = 0; - xml_data[res_entry->size + 1] = 0; + /* Save the document and return. */ + info->doc = doc; + info->root = root; + wim->xml_info = info; + return 0; - DEBUG("Parsing XML using libxml2 to create XML tree"); +err_free_doc: + xmlFreeDoc(doc); +err_free_info: + FREE(info); +err: + return ret; +} - doc = xmlReadMemory(xml_data, res_entry->size, - "noname.xml", "UTF-16", 0); +/* Swap the INDEX attributes of two IMAGE elements. */ +static void +swap_index_attributes(xmlNode *image_node_1, xmlNode *image_node_2) +{ + xmlAttr *attr_1, *attr_2; - if (!doc) { - ERROR("Failed to parse XML data"); - ret = WIMLIB_ERR_XML; - goto out_free_xml_data; + if (image_node_1 != image_node_2) { + attr_1 = unlink_index_attribute(image_node_1); + attr_2 = unlink_index_attribute(image_node_2); + xmlAddChild(image_node_1, (xmlNode *)attr_2); + xmlAddChild(image_node_2, (xmlNode *)attr_1); } +} - DEBUG("Constructing WIM information structure from XML tree."); +static int +prepare_document_for_write(struct wim_xml_info *info, int image, u64 total_bytes, + xmlNode **orig_totalbytes_node_ret) +{ + xmlNode *totalbytes_node = NULL; + + /* Allocate the new TOTALBYTES element if needed. */ + if (total_bytes != WIM_TOTALBYTES_USE_EXISTING && + total_bytes != WIM_TOTALBYTES_OMIT) { + totalbytes_node = new_element_with_u64(NULL, "TOTALBYTES", + total_bytes); + if (!totalbytes_node) + return WIMLIB_ERR_NOMEM; + } - root = xmlDocGetRootElement(doc); - if (!root) { - ERROR("WIM XML data is an empty XML document"); - ret = WIMLIB_ERR_XML; - goto out_free_doc; + /* Adjust the IMAGE elements if needed. */ + if (image != WIMLIB_ALL_IMAGES) { + /* We're writing a single image only. Temporarily unlink all + * other IMAGE elements from the document. */ + for (int i = 0; i < info->image_count; i++) + if (i + 1 != image) + xmlUnlinkNode(info->images[i]); + + /* Temporarily set the INDEX attribute of the needed IMAGE + * element to 1. */ + swap_index_attributes(info->images[0], info->images[image - 1]); } - if (!node_is_element(root) || !node_name_is(root, "WIM")) { - ERROR("Expected for the root XML element (found <%s>)", - root->name); - ret = WIMLIB_ERR_XML; - goto out_free_doc; + /* Adjust (add, change, or remove) the TOTALBYTES element if needed. */ + *orig_totalbytes_node_ret = NULL; + if (total_bytes != WIM_TOTALBYTES_USE_EXISTING) { + /* Unlink the previous TOTALBYTES element, if any. */ + *orig_totalbytes_node_ret = xml_get_node_by_path(info->root, + "TOTALBYTES"); + if (*orig_totalbytes_node_ret) + xmlUnlinkNode(*orig_totalbytes_node_ret); + + /* Link in the new TOTALBYTES element, if any. */ + if (totalbytes_node) + xmlAddChild(info->root, totalbytes_node); } + return 0; +} - ret = xml_read_wim_info(root, info_ret); - if (ret != 0) - goto out_free_doc; +static void +restore_document_after_write(struct wim_xml_info *info, int image, + xmlNode *orig_totalbytes_node) +{ + /* Restore the IMAGE elements if needed. */ + if (image != WIMLIB_ALL_IMAGES) { + /* We wrote a single image only. Re-link all other IMAGE + * elements to the document. */ + for (int i = 0; i < info->image_count; i++) + if (i + 1 != image) + xmlAddChild(info->root, info->images[i]); + + /* Restore the original INDEX attributes. */ + swap_index_attributes(info->images[0], info->images[image - 1]); + } - *xml_data_ret = xml_data; - xml_data = NULL; -out_free_doc: - DEBUG("Freeing XML tree."); - xmlFreeDoc(doc); -out_free_xml_data: - FREE(xml_data); -out: - return ret; + /* Restore the original TOTALBYTES element if needed. */ + if (orig_totalbytes_node) + node_replace_child_element(info->root, orig_totalbytes_node); } -#define CHECK_RET ({ if (ret < 0) { \ - ERROR("Error writing XML data"); \ - ret = WIMLIB_ERR_WRITE; \ - goto out_free_text_writer; \ - } }) - /* - * Writes XML data to a WIM file. + * Writes the XML data to a WIM file. + * + * 'image' specifies the image(s) to include in the XML data. Normally it is + * WIMLIB_ALL_IMAGES, but it can also be a 1-based image index. * - * If @total_bytes is non-zero, it specifies what to write to the TOTALBYTES - * element in the XML data. If zero, TOTALBYTES is given the default value of - * the offset of the XML data. + * 'total_bytes' is the number to use in the top-level TOTALBYTES element, or + * WIM_TOTALBYTES_USE_EXISTING to use the existing value from the XML document + * (if any), or WIM_TOTALBYTES_OMIT to omit the TOTALBYTES element entirely. */ -int write_xml_data(const struct wim_info *wim_info, int image, FILE *out, - u64 total_bytes, struct resource_entry *out_res_entry) +int +write_wim_xml_data(WIMStruct *wim, int image, u64 total_bytes, + struct wim_reshdr *out_reshdr, int write_resource_flags) { - xmlCharEncodingHandler *encoding_handler; - xmlOutputBuffer *out_buffer; - xmlTextWriter *writer; - int ret; - off_t start_offset; - off_t end_offset; - - wimlib_assert(image == WIMLIB_ALL_IMAGES || - (wim_info != NULL && image >= 1 && - image <= wim_info->num_images)); + struct wim_xml_info *info = wim->xml_info; + long ret; + long ret2; + xmlBuffer *buffer; + xmlNode *orig_totalbytes_node; + xmlSaveCtxt *save_ctx; + + /* Make any needed temporary changes to the document. */ + ret = prepare_document_for_write(info, image, total_bytes, + &orig_totalbytes_node); + if (ret) + goto out; - start_offset = ftello(out); - if (start_offset == -1) - return WIMLIB_ERR_WRITE; + /* Create an in-memory buffer to hold the encoded document. */ + ret = WIMLIB_ERR_NOMEM; + buffer = xmlBufferCreate(); + if (!buffer) + goto out_restore_document; + + /* Encode the document in UTF-16LE, with a byte order mark, and with no + * XML declaration. Some other WIM software requires all of these + * characteristics. */ + ret = WIMLIB_ERR_NOMEM; + if (xmlBufferCat(buffer, "\xff\xfe")) + goto out_free_buffer; + save_ctx = xmlSaveToBuffer(buffer, "UTF-16LE", XML_SAVE_NO_DECL); + if (!save_ctx) + goto out_free_buffer; + ret = xmlSaveDoc(save_ctx, info->doc); + ret2 = xmlSaveClose(save_ctx); + if (ret < 0 || ret2 < 0) { + ERROR("Unable to serialize the WIM file's XML document!"); + ret = WIMLIB_ERR_NOMEM; + goto out_free_buffer; + } - DEBUG("Writing XML data for image %d at offset %"PRIu64, - image, start_offset); + /* Write the XML data uncompressed. Although wimlib can handle + * compressed XML data, some other WIM software cannot. */ + ret = write_wim_resource_from_buffer(xmlBufferContent(buffer), + xmlBufferLength(buffer), + true, + &wim->out_fd, + WIMLIB_COMPRESSION_TYPE_NONE, + 0, + out_reshdr, + NULL, + write_resource_flags); +out_free_buffer: + xmlBufferFree(buffer); +out_restore_document: + /* Revert any temporary changes we made to the document. */ + restore_document_after_write(info, image, orig_totalbytes_node); +out: + return ret; +} - /* 2 bytes endianness marker for UTF-16LE. This is _required_ for WIM - * XML data. */ - if ((putc(0xff, out)) == EOF || (putc(0xfe, out) == EOF)) { - ERROR_WITH_ERRNO("Error writing XML data"); - return WIMLIB_ERR_WRITE; - } +/*----------------------------------------------------------------------------* + * Global setup functions * + *----------------------------------------------------------------------------*/ - /* The contents of the element in the XML data, under the - * element (not the element), is for non-split WIMs the - * size of the WIM file excluding the XML data and integrity table. - * This should be equal to the current position in the output stream, - * since the XML data and integrity table are the last elements of the - * WIM. - * - * For split WIMs, takes into account the entire WIM, not - * just the current part. In that case, @total_bytes should be passed - * in to this function. */ - if (total_bytes == 0) - total_bytes = start_offset; - - xmlInitCharEncodingHandlers(); - - /* The encoding of the XML data must be UTF-16LE. */ - encoding_handler = xmlGetCharEncodingHandler(XML_CHAR_ENCODING_UTF16LE); - if (!encoding_handler) { - ERROR("Failed to get XML character encoding handler for UTF-16LE"); - ret = WIMLIB_ERR_LIBXML_UTF16_HANDLER_NOT_AVAILABLE; - goto out_cleanup_char_encoding_handlers; - } +void +xml_global_init(void) +{ + xmlInitParser(); +} - out_buffer = xmlOutputBufferCreateFile(out, encoding_handler); - if (!out_buffer) { - ERROR("Failed to allocate xmlOutputBuffer"); - ret = WIMLIB_ERR_NOMEM; - goto out_cleanup_char_encoding_handlers; - } +void +xml_global_cleanup(void) +{ + xmlCleanupParser(); +} - writer = xmlNewTextWriter(out_buffer); - if (!writer) { - ERROR("Failed to allocate xmlTextWriter"); - ret = WIMLIB_ERR_NOMEM; - goto out_output_buffer_close; - } +void +xml_set_memory_allocator(void *(*malloc_func)(size_t), + void (*free_func)(void *), + void *(*realloc_func)(void *, size_t)) +{ + xmlMemSetup(free_func, malloc_func, realloc_func, wimlib_strdup); +} - DEBUG("Writing element"); +/*----------------------------------------------------------------------------* + * Library API functions * + *----------------------------------------------------------------------------*/ - ret = xmlTextWriterStartElement(writer, "WIM"); - CHECK_RET; +WIMLIBAPI int +wimlib_get_xml_data(WIMStruct *wim, void **buf_ret, size_t *bufsize_ret) +{ + const struct wim_reshdr *xml_reshdr; - ret = xmlTextWriterWriteFormatElement(writer, "TOTALBYTES", "%"PRIu64, - total_bytes); - CHECK_RET; + if (wim->filename == NULL && filedes_is_seekable(&wim->in_fd)) + return WIMLIB_ERR_NO_FILENAME; - if (wim_info != NULL) { - int first, last; - if (image == WIMLIB_ALL_IMAGES) { - first = 1; - last = wim_info->num_images; - } else { - first = image; - last = image; - } - DEBUG("Writing %d elements", last - first + 1); - for (int i = first; i <= last; i++) { - ret = xml_write_image_info(writer, &wim_info->images[i - 1]); - CHECK_RET; - } - } + if (buf_ret == NULL || bufsize_ret == NULL) + return WIMLIB_ERR_INVALID_PARAM; - ret = xmlTextWriterEndElement(writer); - CHECK_RET; + xml_reshdr = &wim->hdr.xml_data_reshdr; - ret = xmlTextWriterEndDocument(writer); - CHECK_RET; + *bufsize_ret = xml_reshdr->uncompressed_size; + return wim_reshdr_to_data(xml_reshdr, wim, buf_ret); +} - DEBUG("Ended XML document"); +WIMLIBAPI int +wimlib_extract_xml_data(WIMStruct *wim, FILE *fp) +{ + int ret; + void *buf; + size_t bufsize; - /* Call xmlFreeTextWriter() before ftello() because the former will - * flush the file stream. */ - xmlFreeTextWriter(writer); - writer = NULL; + ret = wimlib_get_xml_data(wim, &buf, &bufsize); + if (ret) + return ret; - end_offset = ftello(out); - if (end_offset == -1) { + if (fwrite(buf, 1, bufsize, fp) != bufsize) { + ERROR_WITH_ERRNO("Failed to extract XML data"); ret = WIMLIB_ERR_WRITE; - } else { - ret = 0; - out_res_entry->offset = start_offset; - out_res_entry->size = end_offset - start_offset; - out_res_entry->original_size = end_offset - start_offset; - out_res_entry->flags = WIM_RESHDR_FLAG_METADATA; } -out_free_text_writer: - /* xmlFreeTextWriter will free the attached xmlOutputBuffer. */ - xmlFreeTextWriter(writer); - out_buffer = NULL; -out_output_buffer_close: - if (out_buffer != NULL) - xmlOutputBufferClose(out_buffer); -out_cleanup_char_encoding_handlers: - xmlCleanupCharEncodingHandlers(); -out: - if (ret == 0) - DEBUG("Successfully wrote XML data"); + FREE(buf); return ret; } -/* Returns the name of the specified image. */ -WIMLIBAPI const char *wimlib_get_image_name(const WIMStruct *w, int image) +static bool +image_name_in_use(const WIMStruct *wim, const tchar *name, int excluded_image) { - if (image < 1 || image > w->hdr.image_count) - return NULL; - return w->wim_info->images[image - 1].name; + const struct wim_xml_info *info = wim->xml_info; + const xmlChar *name_utf8; + bool found = false; + + /* Any number of images can have "no name". */ + if (!name || !*name) + return false; + + /* Check for images that have the specified name. */ + if (tstr_get_utf8(name, &name_utf8)) + return false; + for (int i = 0; i < info->image_count && !found; i++) { + if (i + 1 == excluded_image) + continue; + found = xmlStrEqual(name_utf8, xml_get_text_by_path( + info->images[i], "NAME")); + } + tstr_put_utf8(name_utf8); + return found; } -/* Returns the description of the specified image. */ -WIMLIBAPI const char *wimlib_get_image_description(const WIMStruct *w, - int image) +WIMLIBAPI bool +wimlib_image_name_in_use(const WIMStruct *wim, const tchar *name) { - if (image < 1 || image > w->hdr.image_count) - return NULL; - return w->wim_info->images[image - 1].description; + return image_name_in_use(wim, name, WIMLIB_NO_IMAGE); } -/* Determines if an image name is already used by some image in the WIM. */ -WIMLIBAPI bool wimlib_image_name_in_use(const WIMStruct *w, const char *name) +WIMLIBAPI const tchar * +wimlib_get_image_name(const WIMStruct *wim, int image) { - if (!name || !*name) - return false; - for (int i = 1; i <= w->hdr.image_count; i++) - if (strcmp(w->wim_info->images[i - 1].name, name) == 0) - return true; - return false; + return get_image_property(wim, image, "NAME", T("")); } -/* Extracts the raw XML data to a file stream. */ -WIMLIBAPI int wimlib_extract_xml_data(WIMStruct *w, FILE *fp) +WIMLIBAPI const tchar * +wimlib_get_image_description(const WIMStruct *wim, int image) { - if (!w->xml_data) - return WIMLIB_ERR_INVALID_PARAM; - if (fwrite(w->xml_data, 1, w->hdr.xml_res_entry.size, fp) != - w->hdr.xml_res_entry.size) { - ERROR_WITH_ERRNO("Failed to extract XML data"); - return WIMLIB_ERR_WRITE; - } - return 0; + return get_image_property(wim, image, "DESCRIPTION", NULL); } -/* Sets the name of an image in the WIM. */ -WIMLIBAPI int wimlib_set_image_name(WIMStruct *w, int image, const char *name) +WIMLIBAPI const tchar * +wimlib_get_image_property(const WIMStruct *wim, int image, + const tchar *property_name) { - char *p; - int i; - - DEBUG("Setting the name of image %d to %s", image, name); - - if (!name || !*name) { - ERROR("Must specify a non-empty string for the image name"); - return WIMLIB_ERR_INVALID_PARAM; - } - - if (image < 1 || image > w->hdr.image_count) { - ERROR("%d is not a valid image", image); - return WIMLIB_ERR_INVALID_IMAGE; - } + const xmlChar *name; + const tchar *value; - for (i = 1; i <= w->hdr.image_count; i++) { - if (i == image) - continue; - if (strcmp(w->wim_info->images[i - 1].name, name) == 0) { - ERROR("The name `%s' is already used for image %d", - name, i); - return WIMLIB_ERR_IMAGE_NAME_COLLISION; - } - } + if (!property_name || !*property_name) + return NULL; + if (tstr_get_utf8(property_name, &name)) + return NULL; + value = get_image_property(wim, image, name, NULL); + tstr_put_utf8(name); + return value; +} - p = STRDUP(name); - if (!p) - return WIMLIB_ERR_NOMEM; +WIMLIBAPI int +wimlib_set_image_name(WIMStruct *wim, int image, const tchar *name) +{ + if (image_name_in_use(wim, name, image)) + return WIMLIB_ERR_IMAGE_NAME_COLLISION; - FREE(w->wim_info->images[image - 1].name); - w->wim_info->images[image - 1].name = p; - return 0; + return set_image_property(wim, image, "NAME", name); } -/* Sets the description of an image in the WIM. */ -WIMLIBAPI int wimlib_set_image_descripton(WIMStruct *w, int image, - const char *description) +WIMLIBAPI int +wimlib_set_image_descripton(WIMStruct *wim, int image, const tchar *description) { - char *p; + return set_image_property(wim, image, "DESCRIPTION", description); +} - if (image < 1 || image > w->hdr.image_count) { - ERROR("%d is not a valid image", image); - return WIMLIB_ERR_INVALID_IMAGE; - } - if (description) { - p = STRDUP(description); - if (!p) - return WIMLIB_ERR_NOMEM; - } else { - p = NULL; - } - FREE(w->wim_info->images[image - 1].description); - w->wim_info->images[image - 1].description = p; - return 0; +WIMLIBAPI int +wimlib_set_image_flags(WIMStruct *wim, int image, const tchar *flags) +{ + return set_image_property(wim, image, "FLAGS", flags); } -/* Set the element of a WIM image */ -WIMLIBAPI int wimlib_set_image_flags(WIMStruct *w, int image, - const char *flags) +WIMLIBAPI int +wimlib_set_image_property(WIMStruct *wim, int image, const tchar *property_name, + const tchar *property_value) { - char *p; + const xmlChar *name; + int ret; - if (image < 1 || image > w->hdr.image_count) { - ERROR("%d is not a valid image", image); - return WIMLIB_ERR_INVALID_IMAGE; - } - if (flags) { - p = STRDUP(flags); - if (!p) - return WIMLIB_ERR_NOMEM; - } else { - p = NULL; - } - FREE(w->wim_info->images[image - 1].flags); - w->wim_info->images[image - 1].flags = p; - return 0; + if (!property_name || !*property_name) + return WIMLIB_ERR_INVALID_PARAM; + + ret = tstr_get_utf8(property_name, &name); + if (ret) + return ret; + ret = set_image_property(wim, image, name, property_value); + tstr_put_utf8(name); + return ret; }