/*
- * xml.c
- *
- * Deals with the XML information in WIM files. Uses the C library libxml2.
+ * xml.c - deals with the XML information in WIM files
*/
/*
- * Copyright (C) 2012, 2013 Eric Biggers
+ * Copyright 2012-2023 Eric Biggers
*
* 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
* details.
*
* 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/.
+ * along with this file; if not, see https://www.gnu.org/licenses/.
*/
#ifdef HAVE_CONFIG_H
# include "config.h"
#endif
-#include <libxml/encoding.h>
-#include <libxml/parser.h>
-#include <libxml/tree.h>
-#include <libxml/xmlwriter.h>
+#include <stdlib.h>
#include <string.h>
-#include "wimlib/assert.h"
#include "wimlib/blob_table.h"
#include "wimlib/dentry.h"
#include "wimlib/encoding.h"
#include "wimlib/resource.h"
#include "wimlib/timestamp.h"
#include "wimlib/xml.h"
+#include "wimlib/xmlproc.h"
#include "wimlib/write.h"
-/* 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;
-};
-
-struct windows_info {
- u64 arch;
- tchar *product_name;
- tchar *edition_id;
- tchar *installation_type;
- tchar *pkeyconfigversion;
- tchar *hal;
- tchar *product_type;
- tchar *product_suite;
- tchar **languages;
- tchar *default_language;
- size_t num_languages;
- tchar *system_root;
- bool windows_version_exists;
- struct windows_version windows_version;
-};
-
-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;
- tchar *name;
- tchar *description;
- tchar *display_name;
- tchar *display_description;
- tchar *flags;
- bool wimboot;
-
- /* Note: must update clone_image_info() if adding new fields here */
-
- struct blob_table *blob_table; /* temporary field */
-};
+/*
+ * 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 {
-/* A struct wim_info structure corresponds to the entire XML data for a WIM file. */
-struct wim_info {
- u64 total_bytes;
- int num_images;
- /* Array of `struct image_info's, one for each image in the WIM that is
- * mentioned in the XML data. */
- struct image_info *images;
-};
+ /* The XML document in tree form */
+ struct xml_node *root;
-struct xml_string_spec {
- const char *name;
- size_t offset;
-};
+ /* 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. */
+ struct xml_node **images;
-#define ELEM(STRING_NAME, MEMBER_NAME) \
- {STRING_NAME, offsetof(struct image_info, MEMBER_NAME)}
-static const struct xml_string_spec
-image_info_xml_string_specs[] = {
- ELEM("NAME", name),
- ELEM("DESCRIPTION", description),
- ELEM("DISPLAYNAME", display_name),
- ELEM("DISPLAYDESCRIPTION", display_description),
- ELEM("FLAGS", flags),
+ /* The number of WIM images (the length of 'images') */
+ int image_count;
};
-#undef ELEM
-
-#define ELEM(STRING_NAME, MEMBER_NAME) \
- {STRING_NAME, offsetof(struct windows_info, MEMBER_NAME)}
-static const struct xml_string_spec
-windows_info_xml_string_specs[] = {
- ELEM("PRODUCTNAME", product_name),
- ELEM("EDITIONID", edition_id),
- ELEM("INSTALLATIONTYPE", installation_type),
- ELEM("HAL", hal),
- ELEM("PRODUCTTYPE", product_type),
- ELEM("PRODUCTSUITE", product_suite),
-};
-#undef ELEM
-u64
-wim_info_get_total_bytes(const struct wim_info *info)
+static u64
+parse_number(const tchar *str, int base)
{
- if (info)
- return info->total_bytes;
- else
- return 0;
-}
+ tchar *end;
+ unsigned long long v;
-u64
-wim_info_get_image_hard_link_bytes(const struct wim_info *info, int image)
-{
- if (info)
- return info->images[image - 1].hard_link_bytes;
- else
+ if (!str)
return 0;
-}
-
-u64
-wim_info_get_image_total_bytes(const struct wim_info *info, int image)
-{
- if (info)
- return info->images[image - 1].total_bytes;
- else
+ v = tstrtoull(str, &end, base);
+ if (end == str || *end || v >= UINT64_MAX)
return 0;
+ return v;
}
-unsigned
-wim_info_get_num_images(const struct wim_info *info)
+/*
+ * Retrieve an unsigned integer from the contents of the specified element,
+ * decoding it using the specified base. If the element has no contents or does
+ * not contain a valid number, returns 0.
+ */
+static u64
+xml_element_get_number(const struct xml_node *element, int base)
{
- if (info)
- return info->num_images;
- else
- return 0;
+ return parse_number(xml_element_get_text(element), base);
}
-void
-wim_info_set_wimboot(struct wim_info *info, int image, bool value)
+/*
+ * Retrieve the timestamp from a time element. This element should have child
+ * elements HIGHPART and LOWPART; these elements will be used to construct a
+ * Windows-style timestamp.
+ */
+static u64
+xml_element_get_timestamp(const struct xml_node *element)
{
- info->images[image - 1].wimboot = value;
+ u64 timestamp = 0;
+ const struct xml_node *child;
+
+ xml_node_for_each_child(element, child) {
+ if (xml_node_is_element(child, T("HIGHPART")))
+ timestamp |= xml_element_get_number(child, 16) << 32;
+ else if (xml_node_is_element(child, T("LOWPART")))
+ timestamp |= xml_element_get_number(child, 16);
+ }
+ return timestamp;
}
-bool
-wim_info_get_wimboot(const struct wim_info *info, int image)
+/* Create a new timestamp element and optionally link it into a tree. */
+static struct xml_node *
+xml_new_element_with_timestamp(struct xml_node *parent, const tchar *name,
+ u64 timestamp)
{
- return info->images[image - 1].wimboot;
-}
+ struct xml_node *element;
+ tchar buf[32];
-/* 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
-
-/* Returns a statically allocated string that is a string representation of the
- * architecture number. */
-static const tchar *
-get_arch(int arch)
-{
- switch (arch) {
- case PROCESSOR_ARCHITECTURE_INTEL:
- return T("x86");
- case PROCESSOR_ARCHITECTURE_MIPS:
- return T("MIPS");
- case PROCESSOR_ARCHITECTURE_ARM:
- return T("ARM");
- case PROCESSOR_ARCHITECTURE_IA64:
- return T("ia64");
- case PROCESSOR_ARCHITECTURE_AMD64:
- return T("x86_64");
- default:
- return T("unknown");
- }
-}
+ element = xml_new_element(NULL, name);
+ if (!element)
+ goto err;
+ tsprintf(buf, T("0x%08"PRIX32), (u32)(timestamp >> 32));
+ if (!xml_new_element_with_text(element, T("HIGHPART"), buf))
+ goto err;
-/* Iterate through the children of an xmlNode. */
-#define for_node_child(parent, child) \
- for (child = parent->children; child != NULL; child = child->next)
+ tsprintf(buf, T("0x%08"PRIX32), (u32)timestamp);
+ if (!xml_new_element_with_text(element, T("LOWPART"), buf))
+ goto err;
-/* Utility functions for xmlNodes */
-static inline bool
-node_is_element(xmlNode *node)
-{
- return node->type == XML_ELEMENT_NODE;
-}
+ if (parent)
+ xml_add_child(parent, element);
+ return element;
-static inline bool
-node_is_text(xmlNode *node)
-{
- return node->type == XML_TEXT_NODE;
+err:
+ xml_free_node(element);
+ return NULL;
}
-static inline bool
-node_name_is(xmlNode *node, const char *name)
+/* Create a new number element and optionally link it into a tree. */
+static struct xml_node *
+xml_new_element_with_u64(struct xml_node *parent, const tchar *name, u64 value)
{
- /* For now, both upper case and lower case element names are accepted. */
- return strcasecmp((const char *)node->name, name) == 0;
-}
+ tchar buf[32];
-static u64
-node_get_number(const xmlNode *u64_node, int base)
-{
- xmlNode *child;
- for_node_child(u64_node, child)
- if (node_is_text(child))
- return strtoull(child->content, NULL, base);
- return 0;
+ tsprintf(buf, T("%"PRIu64), value);
+ return xml_new_element_with_text(parent, name, buf);
}
-/* 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)
+static bool
+parse_index(tchar **pp, u32 *index_ret)
{
- return node_get_number(u64_node, 10);
-}
+ tchar *p = *pp;
+ u32 index = 0;
+
+ *p++ = '\0'; /* overwrite '[' */
+ while (*p >= '0' && *p <= '9') {
+ u32 n = (index * 10) + (*p++ - '0');
+ if (n < index)
+ return false;
+ index = n;
+ }
+ if (index == 0)
+ return false;
+ if (*p != ']')
+ return false;
+ p++;
+ if (*p != '/' && *p != '\0')
+ return false;
-/* Like node_get_u64(), but expects a number in base 16. */
-static u64
-node_get_hex_u64(const xmlNode *u64_node)
-{
- return node_get_number(u64_node, 16);
+ *pp = p;
+ *index_ret = index;
+ return true;
}
static int
-node_get_string(const xmlNode *string_node, tchar **tstr_ret)
+do_xml_path_walk(struct xml_node *element, const tchar *path, bool create,
+ struct xml_node **result_ret)
{
- xmlNode *child;
+ size_t n = tstrlen(path) + 1;
+ tchar buf[n];
+ tchar *p;
+ tchar c;
+
+ *result_ret = NULL;
- if (*tstr_ret)
+ if (!element)
return 0;
- for_node_child(string_node, child)
- if (node_is_text(child) && child->content)
- return utf8_to_tstr_simple(child->content, tstr_ret);
- return 0;
-}
+ /* Copy the path to a temporary buffer. */
+ tmemcpy(buf, path, n);
+ p = buf;
-/* Returns the timestamp from a time node. It has child elements <HIGHPART> and
- * <LOWPART> that are then used to construct a 64-bit timestamp. */
-static u64
-node_get_timestamp(const xmlNode *time_node)
-{
- u32 high_part = 0;
- u32 low_part = 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 (*p == '/')
+ goto bad_syntax;
+ c = *p;
+
+ while (c != '\0') {
+ const tchar *name;
+ struct xml_node *child;
+ u32 index = 1;
+
+ /* We have another path component. */
+
+ /* Parse the element name. */
+ name = p;
+ while (*p != '/' && *p != '\0' && *p != '[')
+ p++;
+ if (p == name) /* empty name? */
+ goto bad_syntax;
+
+ /* Handle a bracketed index, if one was specified. */
+ if (*p == '[' && !parse_index(&p, &index))
+ goto bad_syntax;
+
+ c = *p;
+ *p = '\0';
+
+ /* Look for a matching child. */
+ xml_node_for_each_child(element, child)
+ if (xml_node_is_element(child, name) && !--index)
+ 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;
+
+ /* We can't create an element at index 'n' if indices 1...n-1
+ * didn't already exist. */
+ if (index != 1)
+ return WIMLIB_ERR_INVALID_PARAM;
+
+ child = xml_new_element(element, name);
+ if (!child)
+ return WIMLIB_ERR_NOMEM;
+ next_step:
+ /* Continue to the next path component, if there is one. */
+ element = child;
+ p++;
}
- return (u64)low_part | ((u64)high_part << 32);
+
+ *result_ret = element;
+ return 0;
+
+bad_syntax:
+ ERROR("The XML path \"%"TS"\" has invalid syntax.", path);
+ return WIMLIB_ERR_INVALID_PARAM;
}
-/* Used to sort an array of struct image_infos by their image indices. */
-static int
-sort_by_index(const void *p1, const void *p2)
+/* 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 struct xml_node *
+xml_get_element_by_path(struct xml_node *root, const tchar *path)
{
- 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
- return 0;
-}
+ struct xml_node *element;
+ do_xml_path_walk(root, path, false, &element);
+ return element;
+}
-/* Frees memory allocated inside a struct windows_info structure. */
-static void
-destroy_windows_info(struct windows_info *windows_info)
+/*
+ * Similar to xml_get_element_by_path(), but creates the element and any
+ * requisite ancestor elements as needed. If successful, 0 is returned and
+ * *element_ret is set to a pointer to the resulting element. If unsuccessful,
+ * an error code is returned and *element_ret is set to NULL.
+ */
+static int
+xml_ensure_element_by_path(struct xml_node *root, const tchar *path,
+ struct xml_node **element_ret)
{
- 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);
- FREE(windows_info->pkeyconfigversion);
- 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);
+ return do_xml_path_walk(root, path, true, element_ret);
}
-/* Frees memory allocated inside a struct image_info structure. */
-static void
-destroy_image_info(struct image_info *image_info)
+static u64
+xml_get_number_by_path(struct xml_node *root, const tchar *path)
{
- 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));
+ return xml_element_get_number(xml_get_element_by_path(root, path), 10);
}
-void
-free_wim_info(struct wim_info *info)
+static u64
+xml_get_timestamp_by_path(struct xml_node *root, const tchar *path)
{
- 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);
- }
+ return xml_element_get_timestamp(xml_get_element_by_path(root, path));
}
-/* Reads the information from a <VERSION> element inside the <WINDOWS> element.
- * */
-static void
-xml_read_windows_version(const xmlNode *version_node,
- struct windows_version* windows_version)
+static const tchar *
+xml_get_text_by_path(struct xml_node *root, const tchar *path)
{
- 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);
- }
+ return xml_element_get_text(xml_get_element_by_path(root, path));
}
-/* Reads the information from a <LANGUAGE> element inside a <WINDOWS> element.
- * */
+/*
+ * Create/replace (if text is not NULL and not empty) or remove (if text is NULL
+ * or empty) an element containing text.
+ */
static int
-xml_read_languages(const xmlNode *languages_node,
- tchar ***languages_ret,
- size_t *num_languages_ret,
- tchar **default_language_ret)
+xml_set_text_by_path(struct xml_node *root, const tchar *path,
+ const tchar *text)
{
- xmlNode *child;
- size_t num_languages = 0;
- tchar **languages;
int ret;
+ struct xml_node *element;
- 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;
-
- *languages_ret = languages;
- *num_languages_ret = num_languages;
-
- 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;
+ if (text && *text) {
+ /* Create or replace */
+ ret = xml_ensure_element_by_path(root, path, &element);
+ if (ret)
+ return ret;
+ return xml_element_set_text(element, text);
+ } else {
+ /* Remove */
+ xml_free_node(xml_get_element_by_path(root, path));
+ return 0;
}
- return ret;
}
-/* Reads the information from a <WINDOWS> element inside an <IMAGE> element. */
-static int
-xml_read_windows_info(const xmlNode *windows_node,
- struct windows_info *windows_info)
+/* Unlink and return the node which represents the INDEX attribute of the
+ * specified IMAGE element. */
+static struct xml_node *
+unlink_index_attribute(struct xml_node *image_node)
{
- xmlNode *child;
- int ret = 0;
+ struct xml_node *attr = xml_get_attrib(image_node, T("INDEX"));
- 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);
- } else if (node_name_is(child, "SERVICINGDATA")) {
- xmlNode *grandchild;
-
- for_node_child(child, grandchild) {
- if (node_is_element(grandchild) &&
- node_name_is(grandchild, "PKEYCONFIGVERSION"))
- {
- ret = node_get_string(grandchild,
- &windows_info->pkeyconfigversion);
- }
- }
- }
-
- if (ret != 0)
- return ret;
- }
- return ret;
+ xml_unlink_node(attr);
+ return attr;
}
-/* Reads the information from an <IMAGE> element. */
-static int
-xml_read_image_info(xmlNode *image_node, struct image_info *image_info)
+/* 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)
{
- xmlNode *child;
- xmlChar *index_prop;
- int ret;
+ u64 total_size = 0;
- index_prop = xmlGetProp(image_node, "INDEX");
- if (index_prop) {
- image_info->index = atoi(index_prop);
- xmlFree(index_prop);
- } else {
- image_info->index = 1;
- }
+ for (unsigned i = 0; i < inode->i_num_streams; i++) {
+ const struct blob_descriptor *blob;
- 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")) {
- 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);
- } else if (node_name_is(child, "WIMBOOT")) {
- if (node_get_u64(child) == 1) {
- image_info->wimboot = true;
- }
- }
- if (ret != 0)
- return ret;
- }
- if (!image_info->name) {
- tchar *empty_name;
- empty_name = MALLOC(sizeof(tchar));
- if (!empty_name)
- return WIMLIB_ERR_NOMEM;
- *empty_name = T('\0');
- image_info->name = empty_name;
+ blob = stream_blob(&inode->i_streams[i], blob_table);
+ if (blob)
+ total_size += blob->size;
}
- return ret;
+ return total_size;
}
-/* Reads the information from a <WIM> 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)
+append_image_node(struct wim_xml_info *info, struct xml_node *image_node)
{
- struct wim_info *wim_info;
- xmlNode *child;
+ tchar buf[32];
+ struct xml_node **images;
int ret;
- int num_images;
- int i;
-
- wim_info = CALLOC(1, sizeof(struct wim_info));
- if (!wim_info)
- return WIMLIB_ERR_NOMEM;
- /* 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 (unlikely(num_images == MAX_IMAGES)) {
- ret = WIMLIB_ERR_IMAGE_COUNT;
- goto err;
- }
- num_images++;
- }
- }
+ /* Limit exceeded? */
+ if (unlikely(info->image_count >= MAX_IMAGES))
+ return WIMLIB_ERR_IMAGE_COUNT;
- 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")) {
- 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);
- } else if (node_name_is(child, "ESD")) {
- xmlNode *esdchild;
- for_node_child(child, esdchild) {
- if (node_is_element(esdchild) &&
- node_name_is(esdchild, "ENCRYPTED"))
- {
- ret = WIMLIB_ERR_WIM_IS_ENCRYPTED;
- goto err;
- }
- }
- }
- }
+ /* Set the INDEX attribute. */
+ tsprintf(buf, T("%d"), info->image_count + 1);
+ ret = xml_set_attrib(image_node, T("INDEX"), buf);
+ if (ret)
+ return ret;
- /* 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);
- ret = WIMLIB_ERR_IMAGE_COUNT;
- goto err;
- }
- }
+ /* 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;
- }
- *wim_info_ret = wim_info;
+ /* Add the IMAGE element to the document. */
+ xml_add_child(info->root, image_node);
return 0;
-err:
- free_wim_info(wim_info);
- return ret;
}
-/* Prints the information contained in a `struct windows_info'. */
-static void
-print_windows_info(const struct windows_info *windows_info)
-{
- const struct windows_version *windows_version;
-
- tprintf(T("Architecture: %"TS"\n"),
- get_arch(windows_info->arch));
-
- if (windows_info->product_name) {
- tprintf(T("Product Name: %"TS"\n"),
- windows_info->product_name);
- }
-
- if (windows_info->edition_id) {
- tprintf(T("Edition ID: %"TS"\n"),
- windows_info->edition_id);
- }
-
- if (windows_info->installation_type) {
- tprintf(T("Installation Type: %"TS"\n"),
- windows_info->installation_type);
- }
-
- if (windows_info->hal) {
- tprintf(T("HAL: %"TS"\n"),
- windows_info->hal);
- }
-
- if (windows_info->product_type) {
- tprintf(T("Product Type: %"TS"\n"),
- windows_info->product_type);
- }
-
- if (windows_info->product_suite) {
- tprintf(T("Product Suite: %"TS"\n"),
- windows_info->product_suite);
- }
+/*----------------------------------------------------------------------------*
+ * Functions for internal library use *
+ *----------------------------------------------------------------------------*/
- tprintf(T("Languages: "));
- for (size_t i = 0; i < windows_info->num_languages; i++) {
+/* Allocate an empty 'struct wim_xml_info', containing no images. */
+struct wim_xml_info *
+xml_new_info_struct(void)
+{
+ struct wim_xml_info *info = CALLOC(1, sizeof(*info));
- tfputs(windows_info->languages[i], stdout);
- tputchar(T(' '));
- }
- tputchar(T('\n'));
- if (windows_info->default_language) {
- tprintf(T("Default Language: %"TS"\n"),
- windows_info->default_language);
- }
- if (windows_info->system_root) {
- tprintf(T("System Root: %"TS"\n"),
- windows_info->system_root);
- }
+ if (!info)
+ return NULL;
- if (windows_info->windows_version_exists) {
- windows_version = &windows_info->windows_version;
- tprintf(T("Major Version: %"PRIu64"\n"),
- windows_version->major);
- tprintf(T("Minor Version: %"PRIu64"\n"),
- windows_version->minor);
- tprintf(T("Build: %"PRIu64"\n"),
- windows_version->build);
- tprintf(T("Service Pack Build: %"PRIu64"\n"),
- windows_version->sp_build);
- tprintf(T("Service Pack Level: %"PRIu64"\n"),
- windows_version->sp_level);
+ info->root = xml_new_element(NULL, T("WIM"));
+ if (!info->root) {
+ FREE(info);
+ return NULL;
}
+ return info;
}
-static int
-xml_write_string(xmlTextWriter *writer, const char *name,
- const tchar *tstr)
+/* Free a 'struct wim_xml_info'. */
+void
+xml_free_info_struct(struct wim_xml_info *info)
{
- if (tstr) {
- char *utf8_str;
- int rc = tstr_to_utf8_simple(tstr, &utf8_str);
- if (rc)
- return rc;
- rc = xmlTextWriterWriteElement(writer, name, utf8_str);
- FREE(utf8_str);
- if (rc < 0)
- return rc;
+ if (info) {
+ xml_free_node(info->root);
+ FREE(info->images);
+ FREE(info);
}
- return 0;
}
-static int
-xml_write_strings_from_specs(xmlTextWriter *writer,
- const void *struct_with_strings,
- const struct xml_string_spec specs[],
- size_t num_specs)
+/* 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)
{
- for (size_t i = 0; i < num_specs; i++) {
- int rc = xml_write_string(writer, specs[i].name,
- *(const tchar * const *)
- (struct_with_strings + specs[i].offset));
- if (rc)
- return rc;
- }
- return 0;
+ return info->image_count;
}
-static int
-dup_strings_from_specs(const void *old_struct_with_strings,
- void *new_struct_with_strings,
- const struct xml_string_spec specs[],
- size_t num_specs)
+/* 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)
{
- for (size_t i = 0; i < num_specs; i++) {
- const tchar *old_str = *(const tchar * const *)
- ((const void*)old_struct_with_strings + specs[i].offset);
- tchar **new_str_p = (tchar **)((void*)new_struct_with_strings + specs[i].offset);
- if (old_str) {
- *new_str_p = TSTRDUP(old_str);
- if (!*new_str_p)
- return WIMLIB_ERR_NOMEM;
- }
- }
- return 0;
+ return xml_get_number_by_path(info->root, T("TOTALBYTES"));
}
-/* Writes the information contained in a `struct windows_version' to the XML
- * document being written. This is the <VERSION> element inside the <WINDOWS>
- * element. */
-static int
-xml_write_windows_version(xmlTextWriter *writer,
- const struct windows_version *version)
+/* 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)
{
- int rc;
- rc = xmlTextWriterStartElement(writer, "VERSION");
- if (rc < 0)
- return rc;
-
- rc = xmlTextWriterWriteFormatElement(writer, "MAJOR", "%"PRIu64,
- version->major);
- if (rc < 0)
- return rc;
-
- rc = xmlTextWriterWriteFormatElement(writer, "MINOR", "%"PRIu64,
- version->minor);
- if (rc < 0)
- return rc;
-
- rc = xmlTextWriterWriteFormatElement(writer, "BUILD", "%"PRIu64,
- version->build);
- if (rc < 0)
- return rc;
-
- rc = xmlTextWriterWriteFormatElement(writer, "SPBUILD", "%"PRIu64,
- version->sp_build);
- if (rc < 0)
- return rc;
-
- rc = xmlTextWriterWriteFormatElement(writer, "SPLEVEL", "%"PRIu64,
- version->sp_level);
- if (rc < 0)
- return rc;
-
- rc = xmlTextWriterEndElement(writer); /* </VERSION> */
- if (rc < 0)
- return rc;
-
- return 0;
+ return xml_get_number_by_path(info->images[image - 1], T("TOTALBYTES"));
}
-/* Writes the information contained in a `struct windows_info' to the XML
- * document being written. This is the <WINDOWS> element. */
-static int
-xml_write_windows_info(xmlTextWriter *writer,
- const struct windows_info *windows_info)
+/* 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)
{
- int rc;
- rc = xmlTextWriterStartElement(writer, "WINDOWS");
- if (rc < 0)
- return rc;
-
- rc = xmlTextWriterWriteFormatElement(writer, "ARCH", "%"PRIu64,
- windows_info->arch);
- if (rc < 0)
- return rc;
-
- rc = xml_write_strings_from_specs(writer,
- windows_info,
- windows_info_xml_string_specs,
- ARRAY_LEN(windows_info_xml_string_specs));
- if (rc)
- return rc;
-
- if (windows_info->num_languages) {
- rc = xmlTextWriterStartElement(writer, "LANGUAGES");
- if (rc < 0)
- return rc;
-
- for (size_t i = 0; i < windows_info->num_languages; i++) {
- rc = xml_write_string(writer, "LANGUAGE",
- windows_info->languages[i]);
- if (rc)
- return rc;
- }
-
- rc = xml_write_string(writer, "DEFAULT",
- windows_info->default_language);
- if (rc)
- return rc;
-
- rc = xmlTextWriterEndElement(writer); /* </LANGUAGES> */
- if (rc < 0)
- return rc;
- }
-
- if (windows_info->pkeyconfigversion) {
- rc = xmlTextWriterStartElement(writer, "SERVICINGDATA");
- if (rc < 0)
- return rc;
-
- rc = xml_write_string(writer, "PKEYCONFIGVERSION",
- windows_info->pkeyconfigversion);
- if (rc)
- return rc;
-
- rc = xmlTextWriterEndElement(writer);
- if (rc < 0)
- return rc;
- }
-
- if (windows_info->windows_version_exists) {
- rc = xml_write_windows_version(writer, &windows_info->windows_version);
- if (rc)
- return rc;
- }
-
- rc = xml_write_string(writer, "SYSTEMROOT", windows_info->system_root);
- if (rc)
- return rc;
-
- rc = xmlTextWriterEndElement(writer); /* </WINDOWS> */
- if (rc < 0)
- return rc;
-
- return 0;
+ return xml_get_number_by_path(info->images[image - 1],
+ T("HARDLINKBYTES"));
}
-/* 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)
+/* 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)
{
- int rc;
- rc = xmlTextWriterStartElement(writer, element_name);
- if (rc < 0)
- return rc;
-
- rc = xmlTextWriterWriteFormatElement(writer, "HIGHPART",
- "0x%08"PRIX32, (u32)(time >> 32));
- if (rc < 0)
- return rc;
-
- rc = xmlTextWriterWriteFormatElement(writer, "LOWPART",
- "0x%08"PRIX32, (u32)time);
- if (rc < 0)
- return rc;
-
- rc = xmlTextWriterEndElement(writer); /* </@element_name> */
- if (rc < 0)
- return rc;
- return 0;
+ return xml_get_number_by_path(info->images[image - 1], T("WIMBOOT"));
}
-/* Writes an <IMAGE> element to the XML document. */
-static int
-xml_write_image_info(xmlTextWriter *writer, const struct image_info *image_info)
+/* 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)
{
- 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)
- return rc;
- }
-
- rc = xml_write_strings_from_specs(writer, image_info,
- image_info_xml_string_specs,
- ARRAY_LEN(image_info_xml_string_specs));
- if (rc)
- return rc;
-
- if (image_info->wimboot) {
- rc = xmlTextWriterWriteFormatElement(writer, "WIMBOOT", "%d", 1);
- if (rc < 0)
- return rc;
- }
-
- rc = xmlTextWriterEndElement(writer); /* </IMAGE> */
- if (rc < 0)
- return rc;
-
- return 0;
+ return xml_get_number_by_path(info->images[image - 1],
+ T("WINDOWS/VERSION/BUILD"));
}
-
-
-/* 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)
+/* Set the WIMBOOT value for the specified image. */
+int
+xml_set_wimboot(struct wim_xml_info *info, int image)
{
- struct image_info *images;
-
- 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];
+ return xml_set_text_by_path(info->images[image - 1],
+ T("WIMBOOT"), T("1"));
}
-static int
-clone_windows_info(const struct windows_info *old, struct windows_info *new)
+/*
+ * 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)
{
- int ret;
-
- new->arch = old->arch;
-
- ret = dup_strings_from_specs(old, new, windows_info_xml_string_specs,
- ARRAY_LEN(windows_info_xml_string_specs));
- if (ret)
- return ret;
-
- if (old->pkeyconfigversion) {
- new->pkeyconfigversion = TSTRDUP(old->pkeyconfigversion);
- if (new->pkeyconfigversion == NULL)
- return WIMLIB_ERR_NOMEM;
+ const struct wim_image_metadata *imd = wim->image_metadata[image - 1];
+ struct xml_node *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;
+ struct xml_node *dircount_node;
+ struct xml_node *filecount_node;
+ struct xml_node *totalbytes_node;
+ struct xml_node *hardlinkbytes_node;
+ struct xml_node *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->languages) {
- new->languages = CALLOC(old->num_languages, sizeof(new->languages[0]));
- 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] = TSTRDUP(old->languages[i]);
- if (!new->languages[i])
- return WIMLIB_ERR_NOMEM;
- }
- }
- if (old->default_language &&
- !(new->default_language = TSTRDUP(old->default_language)))
- return WIMLIB_ERR_NOMEM;
- if (old->system_root && !(new->system_root = TSTRDUP(old->system_root)))
+ dircount_node = xml_new_element_with_u64(NULL, T("DIRCOUNT"),
+ dir_count);
+ filecount_node = xml_new_element_with_u64(NULL, T("FILECOUNT"),
+ file_count);
+ totalbytes_node = xml_new_element_with_u64(NULL, T("TOTALBYTES"),
+ total_bytes);
+ hardlinkbytes_node = xml_new_element_with_u64(NULL, T("HARDLINKBYTES"),
+ hard_link_bytes);
+ lastmodificationtime_node = xml_new_element_with_timestamp(NULL,
+ T("LASTMODIFICATIONTIME"), now_as_wim_timestamp());
+
+ if (unlikely(!dircount_node || !filecount_node || !totalbytes_node ||
+ !hardlinkbytes_node || !lastmodificationtime_node)) {
+ xml_free_node(dircount_node);
+ xml_free_node(filecount_node);
+ xml_free_node(totalbytes_node);
+ xml_free_node(hardlinkbytes_node);
+ xml_free_node(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));
}
+
+ xml_replace_child(image_node, dircount_node);
+ xml_replace_child(image_node, filecount_node);
+ xml_replace_child(image_node, totalbytes_node);
+ xml_replace_child(image_node, hardlinkbytes_node);
+ xml_replace_child(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)
{
+ const u64 now = now_as_wim_timestamp();
+ struct xml_node *image_node;
int ret;
- 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 (name && !xml_legal_value(name)) {
+ ERROR("Name of new image contains illegal characters");
+ return WIMLIB_ERR_INVALID_PARAM;
+ }
- ret = dup_strings_from_specs(old, new,
- image_info_xml_string_specs,
- ARRAY_LEN(image_info_xml_string_specs));
+ ret = WIMLIB_ERR_NOMEM;
+ image_node = xml_new_element(NULL, T("IMAGE"));
+ if (!image_node)
+ goto err;
+ if (name && *name &&
+ !xml_new_element_with_text(image_node, T("NAME"), name))
+ goto err;
+ if (!xml_new_element_with_u64(image_node, T("DIRCOUNT"), 0))
+ goto err;
+ if (!xml_new_element_with_u64(image_node, T("FILECOUNT"), 0))
+ goto err;
+ if (!xml_new_element_with_u64(image_node, T("TOTALBYTES"), 0))
+ goto err;
+ if (!xml_new_element_with_u64(image_node, T("HARDLINKBYTES"), 0))
+ goto err;
+ if (!xml_new_element_with_timestamp(image_node, T("CREATIONTIME"), now))
+ goto err;
+ if (!xml_new_element_with_timestamp(image_node,
+ T("LASTMODIFICATIONTIME"), now))
+ goto err;
+ ret = append_image_node(info, image_node);
if (ret)
- return ret;
-
- if (old->windows_info_exists) {
- new->windows_info_exists = true;
- ret = clone_windows_info(&old->windows_info,
- &new->windows_info);
- if (ret)
- return ret;
- }
- new->wimboot = old->wimboot;
+ goto err;
return 0;
+
+err:
+ xml_free_node(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 tchar *dest_image_name,
- const tchar *dest_image_description)
+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)
{
- struct wim_info *new_wim_info;
- struct image_info *image_info;
+ struct xml_node *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);
-
- 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;
+ if (dest_image_name && !xml_legal_value(dest_image_name)) {
+ ERROR("Destination image name contains illegal characters");
+ return WIMLIB_ERR_INVALID_PARAM;
+ }
+ if (dest_image_description &&
+ !xml_legal_value(dest_image_description)) {
+ ERROR("Destination image description contains illegal characters");
+ return WIMLIB_ERR_INVALID_PARAM;
}
- image_info = add_image_info_struct(new_wim_info);
- if (!image_info)
+ ret = WIMLIB_ERR_NOMEM;
+ dest_node = xml_clone_tree(src_info->images[src_image - 1]);
+ if (!dest_node)
goto err;
- ret = clone_image_info(&old_wim_info->images[image - 1], image_info);
- if (ret != 0)
- goto err_destroy_image_info;
+ ret = xml_set_text_by_path(dest_node, T("NAME"), dest_image_name);
+ if (ret)
+ goto err;
- image_info->index = new_wim_info->num_images;
+ ret = xml_set_text_by_path(dest_node, T("DESCRIPTION"),
+ dest_image_description);
+ if (ret)
+ goto err;
- if (dest_image_name) {
- FREE(image_info->name);
- image_info->name = TSTRDUP(dest_image_name);
- if (!image_info->name)
- goto err_destroy_image_info;
- }
- if (dest_image_description) {
- FREE(image_info->description);
- image_info->description = TSTRDUP(dest_image_description);
- if (!image_info->description)
- goto err_destroy_image_info;
+ if (wimboot) {
+ ret = xml_set_text_by_path(dest_node, T("WIMBOOT"), T("1"));
+ if (ret)
+ goto err;
}
- *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;
+ xml_free_node(dest_node);
+ return ret;
}
-/* Removes an image from the XML information. */
+/* Remove the specified image from the XML document. */
void
-xml_delete_image(struct wim_info **wim_info_p, int image)
+xml_delete_image(struct wim_xml_info *info, int image)
{
- struct wim_info *wim_info;
-
- wim_info = *wim_info_p;
- wimlib_assert(image >= 1 && image <= wim_info->num_images);
- DEBUG("Deleting image %d from the XML data.", image);
-
- destroy_image_info(&wim_info->images[image - 1]);
-
- memmove(&wim_info->images[image - 1],
- &wim_info->images[image],
- (wim_info->num_images - image) * sizeof(struct image_info));
-
- 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--;
+ struct xml_node *next_image;
+ struct xml_node *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);
+ xml_free_node(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);
+ xml_add_child(next_image, index_attr);
+ info->images[image - 1] = next_image;
+ image++;
}
-}
-size_t
-xml_get_max_image_name_len(const WIMStruct *wim)
-{
- size_t max_len = 0;
- for (u32 i = 0; i < wim->hdr.image_count; i++)
- max_len = max(max_len, tstrlen(wim->wim_info->images[i].name));
- return max_len;
+ xml_free_node(next_index_attr);
+ info->image_count--;
}
-void
-xml_set_memory_allocator(void *(*malloc_func)(size_t),
- void (*free_func)(void *),
- void *(*realloc_func)(void *, size_t))
+/* 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
+#define PROCESSOR_ARCHITECTURE_ARM64 12
+
+static const tchar *
+describe_arch(u64 arch)
{
- xmlMemSetup(free_func, malloc_func, realloc_func, STRDUP);
+ 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"),
+ [PROCESSOR_ARCHITECTURE_ARM64] = T("ARM64"),
+ };
+
+ if (arch < ARRAY_LEN(descriptions) && descriptions[arch] != NULL)
+ return descriptions[arch];
+
+ return T("unknown");
}
-static int
-calculate_dentry_statistics(struct wim_dentry *dentry, void *arg)
+/* Print information from the WINDOWS element, if present. */
+static void
+print_windows_info(struct xml_node *image_node)
{
- struct image_info *info = arg;
- const struct wim_inode *inode = dentry->d_inode;
-
- /* 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)) {
- if (inode_is_directory(inode))
- info->dir_count++;
- else
- info->file_count++;
- }
+ struct xml_node *windows_node;
+ struct xml_node *langs_node;
+ struct xml_node *version_node;
+ const tchar *text;
- /*
- * Update total bytes and hard link bytes.
- *
- * We try to act the same as the MS implementation, even though there
- * are some inconsistencies/bugs in the way it operates.
- *
- * 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).
- */
- if (!(inode->i_attributes & (FILE_ATTRIBUTE_DIRECTORY |
- FILE_ATTRIBUTE_REPARSE_POINT)))
- {
- struct blob_descriptor *blob;
-
- blob = inode_get_blob_for_unnamed_data_stream(inode,
- info->blob_table);
- if (blob) {
- info->total_bytes += blob->size;
- if (!dentry_is_first_in_inode(dentry))
- info->hard_link_bytes += blob->size;
- }
+ windows_node = xml_get_element_by_path(image_node, T("WINDOWS"));
+ if (!windows_node)
+ return;
- if (inode->i_nlink >= 2 && dentry_is_first_in_inode(dentry)) {
- for (unsigned i = 0; i < inode->i_num_streams; i++) {
- if (stream_is_named_data_stream(&inode->i_streams[i])) {
- blob = stream_blob(&inode->i_streams[i],
- info->blob_table);
- if (blob) {
- info->hard_link_bytes += inode->i_nlink *
- blob->size;
- }
- }
- }
- }
- }
- return 0;
-}
+ tprintf(T("Architecture: %"TS"\n"),
+ describe_arch(xml_get_number_by_path(windows_node, T("ARCH"))));
-/*
- * Calculate what to put in the <FILECOUNT>, <DIRCOUNT>, <TOTALBYTES>, and
- * <HARDLINKBYTES> elements of the specified WIM image.
- */
-void
-xml_update_image_info(WIMStruct *wim, int image)
-{
- struct image_info *image_info;
+ text = xml_get_text_by_path(windows_node, T("PRODUCTNAME"));
+ if (text)
+ tprintf(T("Product Name: %"TS"\n"), text);
- DEBUG("Updating the image info for image %d", image);
+ text = xml_get_text_by_path(windows_node, T("EDITIONID"));
+ if (text)
+ tprintf(T("Edition ID: %"TS"\n"), text);
- image_info = &wim->wim_info->images[image - 1];
+ text = xml_get_text_by_path(windows_node, T("INSTALLATIONTYPE"));
+ if (text)
+ tprintf(T("Installation 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;
- image_info->blob_table = wim->blob_table;
+ text = xml_get_text_by_path(windows_node, T("HAL"));
+ if (text)
+ tprintf(T("HAL: %"TS"\n"), text);
- for_dentry_in_tree(wim->image_metadata[image - 1]->root_dentry,
- calculate_dentry_statistics,
- image_info);
- image_info->last_modification_time = now_as_wim_timestamp();
-}
+ text = xml_get_text_by_path(windows_node, T("PRODUCTTYPE"));
+ if (text)
+ tprintf(T("Product Type: %"TS"\n"), text);
-/* Adds an image to the XML information. */
-int
-xml_add_image(WIMStruct *wim, const tchar *name)
-{
- struct wim_info *wim_info;
- struct image_info *image_info;
+ text = xml_get_text_by_path(windows_node, T("PRODUCTSUITE"));
+ if (text)
+ tprintf(T("Product Suite: %"TS"\n"), text);
- wimlib_assert(name != NULL);
+ langs_node = xml_get_element_by_path(windows_node, T("LANGUAGES"));
+ if (langs_node) {
+ struct xml_node *lang_node;
- /* If this is the first image, allocate the struct wim_info. Otherwise
- * use the existing struct wim_info. */
- if (wim->wim_info) {
- wim_info = wim->wim_info;
- } else {
- wim_info = CALLOC(1, sizeof(struct wim_info));
- if (!wim_info)
- return WIMLIB_ERR_NOMEM;
- }
-
- image_info = add_image_info_struct(wim_info);
- if (!image_info)
- goto out_free_wim_info;
+ tprintf(T("Languages: "));
+ xml_node_for_each_child(langs_node, lang_node) {
+ if (!xml_node_is_element(lang_node, T("LANGUAGE")))
+ continue;
+ text = xml_element_get_text(lang_node);
+ if (!text)
+ continue;
+ tprintf(T("%"TS" "), text);
+ }
+ tputchar(T('\n'));
- if (!(image_info->name = TSTRDUP(name)))
- goto out_destroy_image_info;
+ text = xml_get_text_by_path(langs_node, T("DEFAULT"));
+ if (text)
+ tprintf(T("Default Language: %"TS"\n"), text);
+ }
- wim->wim_info = wim_info;
- image_info->index = wim_info->num_images;
- image_info->creation_time = now_as_wim_timestamp();
- xml_update_image_info(wim, image_info->index);
- return 0;
+ text = xml_get_text_by_path(windows_node, T("SYSTEMROOT"));
+ if (text)
+ tprintf(T("System Root: %"TS"\n"), text);
-out_destroy_image_info:
- destroy_image_info(image_info);
- wim_info->num_images--;
-out_free_wim_info:
- if (wim_info != wim->wim_info)
- FREE(wim_info);
- return WIMLIB_ERR_NOMEM;
+ version_node = xml_get_element_by_path(windows_node, T("VERSION"));
+ if (version_node) {
+ tprintf(T("Major Version: %"PRIu64"\n"),
+ xml_get_number_by_path(version_node, T("MAJOR")));
+ tprintf(T("Minor Version: %"PRIu64"\n"),
+ xml_get_number_by_path(version_node, T("MINOR")));
+ tprintf(T("Build: %"PRIu64"\n"),
+ xml_get_number_by_path(version_node, T("BUILD")));
+ tprintf(T("Service Pack Build: %"PRIu64"\n"),
+ xml_get_number_by_path(version_node, T("SPBUILD")));
+ tprintf(T("Service Pack Level: %"PRIu64"\n"),
+ xml_get_number_by_path(version_node, T("SPLEVEL")));
+ }
}
-/* Prints information about the specified image from struct wim_info structure.
- * */
+/* Prints information about the specified image. */
void
-print_image_info(const struct wim_info *wim_info, int image)
+xml_print_image_info(struct wim_xml_info *info, int image)
{
- const struct image_info *image_info;
- const tchar *desc;
- tchar buf[50];
+ struct xml_node * const image_node = info->images[image - 1];
+ const tchar *text;
+ tchar timebuf[64];
- wimlib_assert(image >= 1 && image <= wim_info->num_images);
+ tprintf(T("Index: %d\n"), image);
- image_info = &wim_info->images[image - 1];
+ /* Always print the Name and Description, even if the corresponding XML
+ * elements are not present. */
+ text = xml_get_text_by_path(image_node, T("NAME"));
+ tprintf(T("Name: %"TS"\n"), text ? text : T(""));
+ text = xml_get_text_by_path(image_node, T("DESCRIPTION"));
+ tprintf(T("Description: %"TS"\n"), text ? text : T(""));
- tprintf(T("Index: %d\n"), image_info->index);
- tprintf(T("Name: %"TS"\n"), image_info->name);
+ text = xml_get_text_by_path(image_node, T("DISPLAYNAME"));
+ if (text)
+ tprintf(T("Display Name: %"TS"\n"), text);
- /* Always print the Description: part even if there is no
- * description. */
- if (image_info->description)
- desc = image_info->description;
- else
- desc = T("");
- tprintf(T("Description: %"TS"\n"), desc);
+ text = xml_get_text_by_path(image_node, T("DISPLAYDESCRIPTION"));
+ if (text)
+ tprintf(T("Display Description: %"TS"\n"), text);
- if (image_info->display_name) {
- tprintf(T("Display Name: %"TS"\n"),
- image_info->display_name);
- }
+ tprintf(T("Directory Count: %"PRIu64"\n"),
+ xml_get_number_by_path(image_node, T("DIRCOUNT")));
- if (image_info->display_description) {
- tprintf(T("Display Description: %"TS"\n"),
- image_info->display_description);
- }
+ tprintf(T("File Count: %"PRIu64"\n"),
+ xml_get_number_by_path(image_node, T("FILECOUNT")));
+
+ tprintf(T("Total Bytes: %"PRIu64"\n"),
+ xml_get_number_by_path(image_node, T("TOTALBYTES")));
- tprintf(T("Directory Count: %"PRIu64"\n"), image_info->dir_count);
- tprintf(T("File Count: %"PRIu64"\n"), image_info->file_count);
- tprintf(T("Total Bytes: %"PRIu64"\n"), image_info->total_bytes);
- tprintf(T("Hard Link Bytes: %"PRIu64"\n"), image_info->hard_link_bytes);
+ tprintf(T("Hard Link Bytes: %"PRIu64"\n"),
+ xml_get_number_by_path(image_node, T("HARDLINKBYTES")));
- wim_timestamp_to_str(image_info->creation_time, buf, sizeof(buf));
- tprintf(T("Creation Time: %"TS"\n"), buf);
+ wim_timestamp_to_str(xml_get_timestamp_by_path(image_node,
+ T("CREATIONTIME")),
+ timebuf, ARRAY_LEN(timebuf));
+ tprintf(T("Creation Time: %"TS"\n"), timebuf);
+
+ wim_timestamp_to_str(xml_get_timestamp_by_path(image_node,
+ T("LASTMODIFICATIONTIME")),
+ timebuf, ARRAY_LEN(timebuf));
+ tprintf(T("Last Modification Time: %"TS"\n"), timebuf);
+
+ print_windows_info(image_node);
+
+ text = xml_get_text_by_path(image_node, T("FLAGS"));
+ if (text)
+ tprintf(T("Flags: %"TS"\n"), text);
- wim_timestamp_to_str(image_info->last_modification_time, buf, sizeof(buf));
- tprintf(T("Last Modification Time: %"TS"\n"), buf);
- if (image_info->windows_info_exists)
- print_windows_info(&image_info->windows_info);
- if (image_info->flags)
- tprintf(T("Flags: %"TS"\n"), image_info->flags);
tprintf(T("WIMBoot compatible: %"TS"\n"),
- image_info->wimboot ? T("yes") : T("no"));
+ xml_get_number_by_path(image_node, T("WIMBOOT")) ?
+ T("yes") : T("no"));
+
tputchar('\n');
}
-void
-libxml_global_init(void)
-{
- xmlInitParser();
- xmlInitCharEncodingHandlers();
-}
+/*----------------------------------------------------------------------------*
+ * Reading and writing the XML data *
+ *----------------------------------------------------------------------------*/
-void
-libxml_global_cleanup(void)
+static int
+image_element_get_index(struct xml_node *element)
{
- xmlCleanupParser();
- xmlCleanupCharEncodingHandlers();
+ struct xml_node *attrib = xml_get_attrib(element, T("INDEX"));
+
+ if (!attrib)
+ return 0;
+ return min(INT_MAX, parse_number(attrib->value, 10));
}
-/* Reads the XML data from a WIM file. */
-int
-read_wim_xml_data(WIMStruct *wim)
+/* Prepare the 'images' array from the XML document tree. */
+static int
+setup_images(struct wim_xml_info *info, struct xml_node *root)
{
- void *buf;
- size_t bufsize;
- u8 *xml_data;
- xmlDoc *doc;
- xmlNode *root;
+ struct xml_node *child;
+ int index;
+ int max_index = 0;
int ret;
- ret = wimlib_get_xml_data(wim, &buf, &bufsize);
- if (ret)
- goto out;
- xml_data = buf;
-
- doc = xmlReadMemory((const char *)xml_data, bufsize,
- NULL, "UTF-16LE", 0);
- if (!doc) {
- ERROR("Failed to parse XML data");
- ret = WIMLIB_ERR_XML;
- goto out_free_xml_data;
+ xml_node_for_each_child(root, child) {
+ if (!xml_node_is_element(child, T("IMAGE")))
+ continue;
+ index = image_element_get_index(child);
+ if (unlikely(index < 1 || info->image_count >= MAX_IMAGES))
+ goto err_indices;
+ max_index = max(max_index, index);
+ info->image_count++;
}
-
- root = xmlDocGetRootElement(doc);
- if (!root || !node_is_element(root) || !node_name_is(root, "WIM")) {
- ERROR("WIM XML data is invalid");
- ret = WIMLIB_ERR_XML;
- goto out_free_doc;
+ 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;
+ xml_node_for_each_child(root, child) {
+ if (!xml_node_is_element(child, T("IMAGE")))
+ continue;
+ index = image_element_get_index(child);
+ if (unlikely(info->images[index - 1]))
+ goto err_indices;
+ info->images[index - 1] = child;
}
+ return 0;
- ret = xml_read_wim_info(root, &wim->wim_info);
-out_free_doc:
- xmlFreeDoc(doc);
-out_free_xml_data:
- FREE(xml_data);
-out:
+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;
}
-/* Prepares an in-memory buffer containing the UTF-16LE XML data for a WIM file.
- *
- * total_bytes is the number to write in <TOTALBYTES>, or
- * WIM_TOTALBYTES_USE_EXISTING to use the existing value in memory, or
- * WIM_TOTALBYTES_OMIT to omit <TOTALBYTES> entirely.
- */
static int
-prepare_wim_xml_data(WIMStruct *wim, int image, u64 total_bytes,
- u8 **xml_data_ret, size_t *xml_len_ret)
+parse_wim_xml_document(const utf16lechar *raw_doc, size_t raw_doc_size,
+ struct xml_node **root_ret)
{
- xmlCharEncodingHandler *encoding_handler;
- xmlBuffer *buf;
- xmlOutputBuffer *outbuf;
- xmlTextWriter *writer;
+ tchar *doc;
int ret;
- int first, last;
- const xmlChar *content;
- int len;
- u8 *xml_data;
- size_t xml_len;
-
- /* Open an xmlTextWriter that writes to an in-memory buffer using
- * UTF-16LE encoding. */
-
- 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;
- }
- buf = xmlBufferCreate();
- if (!buf) {
- ERROR("Failed to create xmlBuffer");
- ret = WIMLIB_ERR_NOMEM;
- goto out;
- }
+ ret = utf16le_to_tstr(raw_doc, raw_doc_size, &doc, NULL);
+ if (ret)
+ return ret;
+ ret = xml_parse_document(doc, root_ret);
+ FREE(doc);
+ return ret;
+}
- outbuf = xmlOutputBufferCreateBuffer(buf, encoding_handler);
- if (!outbuf) {
- ERROR("Failed to allocate xmlOutputBuffer");
- ret = WIMLIB_ERR_NOMEM;
- goto out_buffer_free;
- }
+/* Reads the XML data from a WIM file. */
+int
+read_wim_xml_data(WIMStruct *wim)
+{
+ struct wim_xml_info *info;
+ void *raw_doc;
+ size_t raw_doc_size;
+ struct xml_node *root;
+ int ret;
- writer = xmlNewTextWriter(outbuf);
- if (!writer) {
- ERROR("Failed to allocate xmlTextWriter");
- ret = WIMLIB_ERR_NOMEM;
- goto out_output_buffer_close;
- }
+ /* Allocate the 'struct wim_xml_info'. */
+ ret = WIMLIB_ERR_NOMEM;
+ info = CALLOC(1, sizeof(*info));
+ if (!info)
+ goto err;
- /* Write the XML document. */
-
- ret = xmlTextWriterStartElement(writer, "WIM");
- if (ret < 0)
- goto out_write_error;
-
- /* The contents of the <TOTALBYTES> element in the XML data, under the
- * <WIM> element (not the <IMAGE> element), is for non-split WIMs the
- * size of the WIM file excluding the XML data and integrity table.
- * For split WIMs, <TOTALBYTES> takes into account the entire WIM, not
- * just the current part. */
- if (total_bytes != WIM_TOTALBYTES_OMIT) {
- if (total_bytes == WIM_TOTALBYTES_USE_EXISTING) {
- if (wim->wim_info)
- total_bytes = wim->wim_info->total_bytes;
- else
- total_bytes = 0;
- }
- ret = xmlTextWriterWriteFormatElement(writer, "TOTALBYTES",
- "%"PRIu64, total_bytes);
- if (ret < 0)
- goto out_write_error;
+ /* Read the raw UTF-16LE XML document. */
+ ret = wimlib_get_xml_data(wim, &raw_doc, &raw_doc_size);
+ if (ret)
+ goto err;
+
+ /* Parse the document, creating the document tree. */
+ ret = parse_wim_xml_document(raw_doc, raw_doc_size, &info->root);
+ FREE(raw_doc);
+ raw_doc = NULL;
+ if (ret) {
+ if (ret != WIMLIB_ERR_NOMEM)
+ ret = WIMLIB_ERR_XML;
+ ERROR("Unable to parse the WIM file's XML document!");
+ goto err;
}
+ root = info->root;
- if (image == WIMLIB_ALL_IMAGES) {
- first = 1;
- last = wim->hdr.image_count;
- } else {
- first = image;
- last = image;
+ /* Verify the root element. */
+ if (!xml_node_is_element(root, T("WIM"))) {
+ ERROR("The WIM file's XML document has an unexpected format!");
+ ret = WIMLIB_ERR_XML;
+ goto err;
}
- for (int i = first; i <= last; i++) {
- ret = xml_write_image_info(writer, &wim->wim_info->images[i - 1]);
- if (ret) {
- if (ret < 0)
- goto out_write_error;
- goto out_free_text_writer;
- }
+ /* Verify the WIM file is not encrypted. */
+ if (xml_get_element_by_path(root, T("ESD/ENCRYPTED"))) {
+ ret = WIMLIB_ERR_WIM_IS_ENCRYPTED;
+ goto err;
}
- ret = xmlTextWriterEndElement(writer);
- if (ret < 0)
- goto out_write_error;
+ /* Validate the image elements and set up the images[] array. */
+ ret = setup_images(info, root);
+ if (ret)
+ goto err;
- ret = xmlTextWriterEndDocument(writer);
- if (ret < 0)
- goto out_write_error;
+ /* Success! */
+ wim->xml_info = info;
+ return 0;
- ret = xmlTextWriterFlush(writer);
- if (ret < 0)
- goto out_write_error;
+err:
+ xml_free_info_struct(info);
+ return ret;
+}
- /* Retrieve the buffer into which the document was written. */
+/* Swap the INDEX attributes of two IMAGE elements. */
+static void
+swap_index_attributes(struct xml_node *image_element_1,
+ struct xml_node *image_element_2)
+{
+ struct xml_node *attr_1, *attr_2;
- content = xmlBufferContent(buf);
- len = xmlBufferLength(buf);
+ if (image_element_1 != image_element_2) {
+ attr_1 = unlink_index_attribute(image_element_1);
+ attr_2 = unlink_index_attribute(image_element_2);
+ xml_add_child(image_element_1, attr_2);
+ xml_add_child(image_element_2, attr_1);
+ }
+}
- /* Copy the data into a new buffer, and prefix it with the UTF-16LE BOM
- * (byte order mark), which is required by MS's software to understand
- * the data. */
+static int
+prepare_document_for_write(struct wim_xml_info *info, int image, u64 total_bytes,
+ struct xml_node **orig_totalbytes_element_ret)
+{
+ struct xml_node *totalbytes_element = NULL;
+
+ /* Allocate the new TOTALBYTES element if needed. */
+ if (total_bytes != WIM_TOTALBYTES_USE_EXISTING &&
+ total_bytes != WIM_TOTALBYTES_OMIT) {
+ totalbytes_element = xml_new_element_with_u64(
+ NULL, T("TOTALBYTES"), total_bytes);
+ if (!totalbytes_element)
+ return WIMLIB_ERR_NOMEM;
+ }
- xml_len = len + 2;
- xml_data = MALLOC(xml_len);
- if (!xml_data) {
- ret = WIMLIB_ERR_NOMEM;
- goto out_free_text_writer;
+ /* 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)
+ xml_unlink_node(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]);
}
- xml_data[0] = 0xff;
- xml_data[1] = 0xfe;
- memcpy(&xml_data[2], content, len);
-
- /* Clean up libxml objects and return success. */
- *xml_data_ret = xml_data;
- *xml_len_ret = xml_len;
- ret = 0;
-out_free_text_writer:
- /* xmlFreeTextWriter will free the attached xmlOutputBuffer. */
- xmlFreeTextWriter(writer);
- goto out_buffer_free;
-out_output_buffer_close:
- xmlOutputBufferClose(outbuf);
-out_buffer_free:
- xmlBufferFree(buf);
-out:
- DEBUG("ret=%d", ret);
- return ret;
-out_write_error:
- ERROR("Error writing XML data");
- ret = WIMLIB_ERR_WRITE;
- goto out_free_text_writer;
+ /* Adjust (add, change, or remove) the TOTALBYTES element if needed. */
+ *orig_totalbytes_element_ret = NULL;
+ if (total_bytes != WIM_TOTALBYTES_USE_EXISTING) {
+ /* Unlink the previous TOTALBYTES element, if any. */
+ *orig_totalbytes_element_ret = xml_get_element_by_path(
+ info->root, T("TOTALBYTES"));
+ if (*orig_totalbytes_element_ret)
+ xml_unlink_node(*orig_totalbytes_element_ret);
+
+ /* Link in the new TOTALBYTES element, if any. */
+ if (totalbytes_element)
+ xml_add_child(info->root, totalbytes_element);
+ }
+ return 0;
}
-/* Writes the XML data to a WIM file. */
+static void
+restore_document_after_write(struct wim_xml_info *info, int image,
+ struct xml_node *orig_totalbytes_element)
+{
+ /* 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)
+ xml_add_child(info->root, info->images[i]);
+
+ /* Restore the original INDEX attributes. */
+ swap_index_attributes(info->images[0], info->images[image - 1]);
+ }
+
+ /* Restore the original TOTALBYTES element if needed. */
+ if (orig_totalbytes_element)
+ xml_replace_child(info->root, orig_totalbytes_element);
+}
+
+/*
+ * 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.
+ *
+ * '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_wim_xml_data(WIMStruct *wim, int image, u64 total_bytes,
- struct wim_reshdr *out_reshdr,
- int write_resource_flags)
+ struct wim_reshdr *out_reshdr, int write_resource_flags)
{
+ struct wim_xml_info *info = wim->xml_info;
int ret;
- u8 *xml_data;
- size_t xml_len;
+ struct xml_node *orig_totalbytes_element;
+ struct xml_out_buf buf = {};
+ const utf16lechar *raw_doc;
+ size_t raw_doc_size;
+
+ /* Make any needed temporary changes to the document. */
+ ret = prepare_document_for_write(info, image, total_bytes,
+ &orig_totalbytes_element);
+ if (ret)
+ goto out;
- DEBUG("Writing WIM XML data (image=%d, offset=%"PRIu64")",
- image, wim->out_fd.offset);
+ ret = xml_write_document(info->root, &buf);
+ if (ret)
+ goto out_restore_document;
- ret = prepare_wim_xml_data(wim, image, total_bytes,
- &xml_data, &xml_len);
+ ret = tstr_get_utf16le_and_len(buf.buf, &raw_doc, &raw_doc_size);
if (ret)
- return ret;
+ goto out_restore_document;
/* Write the XML data uncompressed. Although wimlib can handle
- * compressed XML data, MS software cannot. */
- ret = write_wim_resource_from_buffer(xml_data,
- xml_len,
- WIM_RESHDR_FLAG_METADATA,
+ * compressed XML data, some other WIM software cannot. */
+ ret = write_wim_resource_from_buffer(raw_doc, raw_doc_size,
+ true,
&wim->out_fd,
WIMLIB_COMPRESSION_TYPE_NONE,
0,
out_reshdr,
NULL,
write_resource_flags);
- FREE(xml_data);
- DEBUG("ret=%d", ret);
+ tstr_put_utf16le(raw_doc);
+out_restore_document:
+ /* Revert any temporary changes we made to the document. */
+ restore_document_after_write(info, image, orig_totalbytes_element);
+ FREE(buf.buf);
+out:
return ret;
}
-/* API function documented in wimlib.h */
-WIMLIBAPI const tchar *
-wimlib_get_image_name(const WIMStruct *wim, int image)
-{
- if (image < 1 || image > wim->hdr.image_count)
- return NULL;
- return wim->wim_info->images[image - 1].name;
-}
-
-/* API function documented in wimlib.h */
-WIMLIBAPI const tchar *
-wimlib_get_image_description(const WIMStruct *wim, int image)
-{
- if (image < 1 || image > wim->hdr.image_count)
- return NULL;
- return wim->wim_info->images[image - 1].description;
-}
-
-/* API function documented in wimlib.h */
-WIMLIBAPI bool
-wimlib_image_name_in_use(const WIMStruct *wim, const tchar *name)
-{
- if (!name || !*name)
- return false;
- for (int i = 1; i <= wim->hdr.image_count; i++)
- if (!tstrcmp(wim->wim_info->images[i - 1].name, name))
- return true;
- return false;
-}
-
+/*----------------------------------------------------------------------------*
+ * Library API functions *
+ *----------------------------------------------------------------------------*/
-/* API function documented in wimlib.h */
WIMLIBAPI int
wimlib_get_xml_data(WIMStruct *wim, void **buf_ret, size_t *bufsize_ret)
{
const struct wim_reshdr *xml_reshdr;
if (wim->filename == NULL && filedes_is_seekable(&wim->in_fd))
- return WIMLIB_ERR_INVALID_PARAM;
+ return WIMLIB_ERR_NO_FILENAME;
if (buf_ret == NULL || bufsize_ret == NULL)
return WIMLIB_ERR_INVALID_PARAM;
xml_reshdr = &wim->hdr.xml_data_reshdr;
- DEBUG("Reading XML data.");
*bufsize_ret = xml_reshdr->uncompressed_size;
return wim_reshdr_to_data(xml_reshdr, wim, buf_ret);
}
return ret;
}
-/* API function documented in wimlib.h */
-WIMLIBAPI int
-wimlib_set_image_name(WIMStruct *wim, int image, const tchar *name)
+static bool
+image_name_in_use(const WIMStruct *wim, const tchar *name, int excluded_image)
{
- tchar *p;
- int i;
-
- if (name == NULL)
- name = T("");
+ const struct wim_xml_info *info = wim->xml_info;
+ const tchar *existing_name;
- if (image < 1 || image > wim->hdr.image_count)
- return WIMLIB_ERR_INVALID_IMAGE;
+ /* Any number of images can have "no name". */
+ if (!name || !*name)
+ return false;
- if (*name) {
- for (i = 1; i <= wim->hdr.image_count; i++) {
- if (i == image)
- continue;
- if (!tstrcmp(wim->wim_info->images[i - 1].name, name))
- return WIMLIB_ERR_IMAGE_NAME_COLLISION;
- }
+ /* Check for images that have the specified name. */
+ for (int i = 0; i < info->image_count; i++) {
+ if (i + 1 == excluded_image)
+ continue;
+ existing_name = xml_get_text_by_path(info->images[i],
+ T("NAME"));
+ if (existing_name && !tstrcmp(existing_name, name))
+ return true;
}
+ return false;
+}
- p = TSTRDUP(name);
- if (!p)
- return WIMLIB_ERR_NOMEM;
+WIMLIBAPI bool
+wimlib_image_name_in_use(const WIMStruct *wim, const tchar *name)
+{
+ return image_name_in_use(wim, name, WIMLIB_NO_IMAGE);
+}
- FREE(wim->wim_info->images[image - 1].name);
- wim->wim_info->images[image - 1].name = p;
- return 0;
+WIMLIBAPI const tchar *
+wimlib_get_image_name(const WIMStruct *wim, int image)
+{
+ const struct wim_xml_info *info = wim->xml_info;
+ const tchar *name;
+
+ if (image < 1 || image > info->image_count)
+ return NULL;
+ name = wimlib_get_image_property(wim, image, T("NAME"));
+ return name ? name : T("");
}
-static int
-do_set_image_info_str(WIMStruct *wim, int image, const tchar *tstr,
- size_t offset)
+WIMLIBAPI const tchar *
+wimlib_get_image_description(const WIMStruct *wim, int image)
{
- tchar *tstr_copy;
- tchar **dest_tstr_p;
+ return wimlib_get_image_property(wim, image, T("DESCRIPTION"));
+}
- if (image < 1 || image > wim->hdr.image_count) {
- ERROR("%d is not a valid image", image);
- return WIMLIB_ERR_INVALID_IMAGE;
- }
- if (tstr) {
- tstr_copy = TSTRDUP(tstr);
- if (!tstr_copy)
- return WIMLIB_ERR_NOMEM;
- } else {
- tstr_copy = NULL;
- }
- dest_tstr_p = (tchar**)((void*)&wim->wim_info->images[image - 1] + offset);
+WIMLIBAPI const tchar *
+wimlib_get_image_property(const WIMStruct *wim, int image,
+ const tchar *property_name)
+{
+ const struct wim_xml_info *info = wim->xml_info;
- FREE(*dest_tstr_p);
- *dest_tstr_p = tstr_copy;
- return 0;
+ if (!property_name || !*property_name)
+ return NULL;
+ if (image < 1 || image > info->image_count)
+ return NULL;
+ return xml_get_text_by_path(info->images[image - 1], property_name);
+}
+
+WIMLIBAPI int
+wimlib_set_image_name(WIMStruct *wim, int image, const tchar *name)
+{
+ return wimlib_set_image_property(wim, image, T("NAME"), name);
}
-/* API function documented in wimlib.h */
WIMLIBAPI int
-wimlib_set_image_descripton(WIMStruct *wim, int image,
- const tchar *description)
+wimlib_set_image_descripton(WIMStruct *wim, int image, const tchar *description)
{
- return do_set_image_info_str(wim, image, description,
- offsetof(struct image_info, description));
+ return wimlib_set_image_property(wim, image, T("DESCRIPTION"),
+ description);
}
-/* API function documented in wimlib.h */
WIMLIBAPI int
wimlib_set_image_flags(WIMStruct *wim, int image, const tchar *flags)
{
- return do_set_image_info_str(wim, image, flags,
- offsetof(struct image_info, flags));
+ return wimlib_set_image_property(wim, image, T("FLAGS"), flags);
+}
+
+WIMLIBAPI int
+wimlib_set_image_property(WIMStruct *wim, int image, const tchar *property_name,
+ const tchar *property_value)
+{
+ struct wim_xml_info *info = wim->xml_info;
+
+ if (!property_name || !*property_name)
+ return WIMLIB_ERR_INVALID_PARAM;
+
+ if (!xml_legal_path(property_name)) {
+ ERROR("Property name '%"TS"' is illegal in XML", property_name);
+ return WIMLIB_ERR_INVALID_PARAM;
+ }
+
+ if (property_value && !xml_legal_value(property_value)) {
+ WARNING("Value of property '%"TS"' contains illegal characters",
+ property_name);
+ return WIMLIB_ERR_INVALID_PARAM;
+ }
+
+ if (image < 1 || image > info->image_count)
+ return WIMLIB_ERR_INVALID_IMAGE;
+
+ if (!tstrcmp(property_name, T("NAME")) &&
+ image_name_in_use(wim, property_value, image))
+ return WIMLIB_ERR_IMAGE_NAME_COLLISION;
+
+ return xml_set_text_by_path(info->images[image - 1], property_name,
+ property_value);
}
git://wimlib.net / wimlib / blobdiff