[PATCH] git-apply doesn't handle same name patches well

@@ -185,6 +185,18 @@ struct image {
 	struct line *line;
 };
 
+/*
+ * Caches patch filenames to handle the case where a
+ * patch chunk reuses a filename
+ */
+struct fn_cache {
+	char *name;
+	struct patch *patch;
+	struct fn_cache *next;
+};
+
+struct fn_cache *fn_cache_top = NULL;
+
 static uint32_t hash_line(const char *cp, size_t len)
 {
 	size_t i;

@@ -2176,6 +2188,51 @@ static int read_file_or_gitlink(struct cache_entry *ce, struct strbuf *buf)
 	return 0;
 }
 
+struct patch *in_fn_cache(char *name)
+{
+	struct fn_cache *p;
+
+	for (p=fn_cache_top; p; p=p->next) {
+		if (!strcmp(name, p->name))
+			return p->patch;
+	}
+
+	return NULL;
+}
+
+void add_to_fn_cache(char *name, struct patch *patch)
+{
+	struct fn_cache *fn_cache;
+
+	/* Always add new_name unless patch is a deletion */
+	if (name != NULL) {
+		fn_cache = xmalloc(sizeof(*fn_cache));
+
+		/* assuming the pointer to filename won't disappear */
+		fn_cache->name = name;
+		fn_cache->patch = patch;
+		fn_cache->next = fn_cache_top;
+
+		fn_cache_top = fn_cache;
+	}
+
+	/* skip normal diffs, creations and copies */
+	/*
+	 * store a failure on rename/deletion cases because
+	 * later chunks shouldn't patch old names
+	 */
+	if ((name == NULL) || (patch->is_rename)) {
+		fn_cache = xmalloc(sizeof(*fn_cache));
+
+		/* assuming the pointer to filename won't disappear */
+		fn_cache->name = patch->old_name;
+		fn_cache->patch = (struct patch *) -1;
+		fn_cache->next = fn_cache_top;
+
+		fn_cache_top = fn_cache;
+	}
+}
+
 static int apply_data(struct patch *patch, struct stat *st, struct cache_entry *ce)
 {
 	struct strbuf buf;

@@ -2188,7 +2245,16 @@ static int apply_data(struct patch *patch, struct stat *st, struct cache_entry *
 		if (read_file_or_gitlink(ce, &buf))
 			return error("read of %s failed", patch->old_name);
 	} else if (patch->old_name) {
-		if (S_ISGITLINK(patch->old_mode)) {
+		struct patch *tpatch = in_fn_cache(patch->old_name);
+
+		if (tpatch != NULL) {
+			if (tpatch == (struct patch *) -1) {
+				return error("patch %s has been renamed/deleted",
+					patch->old_name);
+			}
+			/* We have a patched copy in memory use that */
+			strbuf_add(&buf, tpatch->result, tpatch->resultsize);
+		} else if (S_ISGITLINK(patch->old_mode)) {
 			if (ce) {
 				read_file_or_gitlink(ce, &buf);
 			} else {

@@ -2211,6 +2277,7 @@ static int apply_data(struct patch *patch, struct stat *st, struct cache_entry *
 		return -1; /* note with --reject this succeeds. */
 	patch->result = image.buf;
 	patch->resultsize = image.len;
+	add_to_fn_cache(patch->new_name, patch);
 	free(image.line_allocated);
 
 	if (0 < patch->is_delete && patch->resultsize)

@@ -0,0 +1,40 @@
+#!/bin/sh
+
+test_description='apply same filename'
+
+. ./test-lib.sh
+
+test_expect_success setup '
+	for i in a b c d e f g h i j k l m
+	do
+		echo $i
+	done >same_fn &&
+	git add same_fn &&
+	git commit -m initial
+'
+test_expect_success 'apply same filename with independent changes' '
+	sed -i -e "s/^d/z/" same_fn &&
+	git diff > patch0 &&
+	git add same_fn &&
+	sed -i -e "s/^i/y/" same_fn &&
+	git diff >> patch0 &&
+	cp same_fn same_fn2 &&
+	git reset --hard &&
+	git-apply patch0 &&
+	diff same_fn same_fn2
+'
+
+test_expect_success 'apply same filename with overlapping changes' '
+	git reset --hard
+	sed -i -e "s/^d/z/" same_fn &&
+	git diff > patch0 &&
+	git add same_fn &&
+	sed -i -e "s/^e/y/" same_fn &&
+	git diff >> patch0 &&
+	cp same_fn same_fn2 &&
+	git reset --hard &&
+	git-apply patch0 &&
+	diff same_fn same_fn2
+'
+
+test_done

@@ -1,9 +1,93 @@
 path-list API
 =============
 
-Talk about <path-list.h>, things like
+The path_list API offers a data structure and functions to handle sorted
+and unsorted string lists.
 
-* it is not just paths but strings in general;
-* the calling sequence.
+The name is a bit misleading, a path_list may store not only paths but
+strings in general.
 
-(Dscho)
+The caller:
+
+. Allocates and clears (`memset(&list, '0', sizeof(path_list));`) a
+  `struct path_list` variable.
+
+. Initializes the members. You can manually set the `items` member, but
+  then you have to set `nr`, accordingly. Also don't forget to set
+  `strdup_paths` if you need it.
+
+. Adds new items to the list, using `path_list_append` or `path_list_insert`.
+
+. Can check if a string is in the list using `path_list_has_path` or
+  `unsorted_path_list_has_path` and get it from the list using
+  `path_list_lookup` for sorted lists.
+
+. Can sort an unsorted list using `sort_path_list`.
+
+. Finally it should free the list using `path_list_clear`.
+
+Functions
+---------
+
+* General ones (works with sorted and unsorted lists as well)
+
+`print_path_list`::
+
+	Dump a path_list to stdout, useful mainly for debugging purposes. It
+	can take an optional header argument and it writes out the
+	string-pointer pairs of the path_list, each one in its own line.
+
+`path_list_clear`::
+
+	Free a path_list. The `path` pointer of the items will be freed in case
+	the `strdup_paths` member of the path_list is set. The second parameter
+	controls if the `util` pointer of the items should be freed or not.
+
+* Functions for sorted lists only
+
+`path_list_has_path`::
+
+	Determine if the path_list has a given string or not.
+
+`path_list_insert`::
+
+	Insert a new element to the path_list. The returned pointer can be handy
+	if you want to write something to the `util` pointer of the
+	path_list_item containing the just added string.
+
+`path_list_lookup`::
+
+	Look up a given string in the path_list, returning the containing
+	path_list_item. If the string is not found, NULL is returned.
+
+* Functions for unsorted lists only
+
+`path_list_append`::
+
+	Append a new string to the end of the path_list.
+
+`sort_path_list`::
+
+	Make an unsorted list sorted.
+
+`unsorted_path_list_has_path`::
+
+	It's like `path_list_has_path()` but for unsorted lists.
+
+Data structures
+---------------
+
+* `struct path_list_item`
+
+Represent an item of the list. The `path` member is a pointer to the
+string, and you may use the `util` member for any purpose, if you want.
+
+* `struct path_list`
+
+Represents the list itself.
+
+. The array of items are available via the `items` member.
+. The `nr` member contains the number of items stored in the list.
+. The `alloc` member is used for `ALLOC_GROW()`.
+. Setting the `strdup_paths` member to 1 means that the added paths are
+  copied to the path list and not just a pointer to them is stored.

@@ -1,9 +1,93 @@
 path-list API
 =============
 
-Talk about <path-list.h>, things like
+The path_list API offers a data structure and functions to handle sorted
+and unsorted string lists.
 
-* it is not just paths but strings in general;
-* the calling sequence.
+The name is a bit misleading, a path_list may store not only paths but
+strings in general.
 
-(Dscho)
+The caller:
+
+. Allocates and clears (`memset(&list, \'0', sizeof(list));`) a
+  `struct path_list` variable.
+
+. Initializes the members. You can manually set the `items` member, but
+  then you have to set `nr`, accordingly. Also don't forget to set
+  `strdup_paths` if you need it.
+
+. Adds new items to the list, using `path_list_append` or `path_list_insert`.
+
+. Can check if a string is in the list using `path_list_has_path` or
+  `unsorted_path_list_has_path` and get it from the list using
+  `path_list_lookup` for sorted lists.
+
+. Can sort an unsorted list using `sort_path_list`.
+
+. Finally it should free the list using `path_list_clear`.
+
+Functions
+---------
+
+* General ones (works with sorted and unsorted lists as well)
+
+`print_path_list`::
+
+	Dump a path_list to stdout, useful mainly for debugging purposes. It
+	can take an optional header argument and it writes out the
+	string-pointer pairs of the path_list, each one in its own line.
+
+`path_list_clear`::
+
+	Free a path_list. The `path` pointer of the items will be freed in case
+	the `strdup_paths` member of the path_list is set. The second parameter
+	controls if the `util` pointer of the items should be freed or not.
+
+* Functions for sorted lists only
+
+`path_list_has_path`::
+
+	Determine if the path_list has a given string or not.
+
+`path_list_insert`::
+
+	Insert a new element to the path_list. The returned pointer can be handy
+	if you want to write something to the `util` pointer of the
+	path_list_item containing the just added string.
+
+`path_list_lookup`::
+
+	Look up a given string in the path_list, returning the containing
+	path_list_item. If the string is not found, NULL is returned.
+
+* Functions for unsorted lists only
+
+`path_list_append`::
+
+	Append a new string to the end of the path_list.
+
+`sort_path_list`::
+
+	Make an unsorted list sorted.
+
+`unsorted_path_list_has_path`::
+
+	It's like `path_list_has_path()` but for unsorted lists.
+
+Data structures
+---------------
+
+* `struct path_list_item`
+
+Represent an item of the list. The `path` member is a pointer to the
+string, and you may use the `util` member for any purpose, if you want.
+
+* `struct path_list`
+
+Represents the list itself.
+
+. The array of items are available via the `items` member.
+. The `nr` member contains the number of items stored in the list.
+. The `alloc` member is used for `ALLOC_GROW()`.
+. Setting the `strdup_paths` member to 1 means that the added paths are
+  copied to the path list and not just a pointer to them is stored.

@@ -1,9 +1,93 @@
 path-list API
 =============
 
-Talk about <path-list.h>, things like
+The path_list API offers a data structure and functions to handle sorted
+and unsorted string lists.
 
-* it is not just paths but strings in general;
-* the calling sequence.
+The name is a bit misleading, a path_list may store not only paths but
+strings in general.
 
-(Dscho)
+The caller:
+
+. Allocates and clears (`memset(&list, \'\0', sizeof(list));`) a
+  `struct path_list` variable.
+
+. Initializes the members. You can manually set the `items` member, but
+  then you have to set `nr`, accordingly. Also don't forget to set
+  `strdup_paths` if you need it.
+
+. Adds new items to the list, using `path_list_append` or `path_list_insert`.
+
+. Can check if a string is in the list using `path_list_has_path` or
+  `unsorted_path_list_has_path` and get it from the list using
+  `path_list_lookup` for sorted lists.
+
+. Can sort an unsorted list using `sort_path_list`.
+
+. Finally it should free the list using `path_list_clear`.
+
+Functions
+---------
+
+* General ones (works with sorted and unsorted lists as well)
+
+`print_path_list`::
+
+	Dump a path_list to stdout, useful mainly for debugging purposes. It
+	can take an optional header argument and it writes out the
+	string-pointer pairs of the path_list, each one in its own line.
+
+`path_list_clear`::
+
+	Free a path_list. The `path` pointer of the items will be freed in case
+	the `strdup_paths` member of the path_list is set. The second parameter
+	controls if the `util` pointer of the items should be freed or not.
+
+* Functions for sorted lists only
+
+`path_list_has_path`::
+
+	Determine if the path_list has a given string or not.
+
+`path_list_insert`::
+
+	Insert a new element to the path_list. The returned pointer can be handy
+	if you want to write something to the `util` pointer of the
+	path_list_item containing the just added string.
+
+`path_list_lookup`::
+
+	Look up a given string in the path_list, returning the containing
+	path_list_item. If the string is not found, NULL is returned.
+
+* Functions for unsorted lists only
+
+`path_list_append`::
+
+	Append a new string to the end of the path_list.
+
+`sort_path_list`::
+
+	Make an unsorted list sorted.
+
+`unsorted_path_list_has_path`::
+
+	It's like `path_list_has_path()` but for unsorted lists.
+
+Data structures
+---------------
+
+* `struct path_list_item`
+
+Represent an item of the list. The `path` member is a pointer to the
+string, and you may use the `util` member for any purpose, if you want.
+
+* `struct path_list`
+
+Represents the list itself.
+
+. The array of items are available via the `items` member.
+. The `nr` member contains the number of items stored in the list.
+. The `alloc` member is used for `ALLOC_GROW()`.
+. Setting the `strdup_paths` member to 1 means that the added paths are
+  copied to the path list and not just a pointer to them is stored.

@@ -1,9 +1,126 @@
 path-list API
 =============
 
-Talk about <path-list.h>, things like
+The path_list API offers a data structure and functions to handle sorted
+and unsorted string lists.
 
-* it is not just paths but strings in general;
-* the calling sequence.
+The name is a bit misleading, a path_list may store not only paths but
+strings in general.
 
-(Dscho)
+The caller:
+
+. Allocates and clears a `struct path_list` variable.
+
+. Initializes the members. You might want to set the flag `strdup_paths`
+  if the strings should be strdup()ed. For example, this is necessary
+  when you add something like git_path("..."), since that function returns
+  a static buffer that will change with the next call to git_path().
++
+If you need something advanced, you can manually malloc() the `items`
+member (you need this if you add things later) and you should set the
+`nr` and `alloc` members in that case, too.
+
+. Adds new items to the list, using `path_list_append` or `path_list_insert`.
+
+. Can check if a string is in the list using `path_list_has_path` or
+  `unsorted_path_list_has_path` and get it from the list using
+  `path_list_lookup` for sorted lists.
+
+. Can sort an unsorted list using `sort_path_list`.
+
+. Finally it should free the list using `path_list_clear`.
+
+Example:
+
+----
+struct path_list list;
+int i;
+
+memset(&list, 0, sizeof(struct path_list));
+path_list_append("foo", &list);
+path_list_append("bar", &list);
+for (i = 0; i < list.nr; i++)
+	printf("%s\n", list.items[i].path)
+----
+
+NOTE: It is more efficient to build an unsorted list and sort it
+afterwards, instead of building a sorted list `(O(n log n)` instead of
+`O(n^2))`.
++
+However, if you use the list to check if a certain string was added
+already, you should not do that (using unsorted_path_list_has_path()),
+because the complexity would be quadratic again (but with a worse factor).
+
+Functions
+---------
+
+* General ones (works with sorted and unsorted lists as well)
+
+`print_path_list`::
+
+	Dump a path_list to stdout, useful mainly for debugging purposes. It
+	can take an optional header argument and it writes out the
+	string-pointer pairs of the path_list, each one in its own line.
+
+`path_list_clear`::
+
+	Free a path_list. The `path` pointer of the items will be freed in case
+	the `strdup_paths` member of the path_list is set. The second parameter
+	controls if the `util` pointer of the items should be freed or not.
+
+* Functions for sorted lists only
+
+`path_list_has_path`::
+
+	Determine if the path_list has a given string or not.
+
+`path_list_insert`::
+
+	Insert a new element to the path_list. The returned pointer can be handy
+	if you want to write something to the `util` pointer of the
+	path_list_item containing the just added string.
++
+Since this function uses xrealloc() (which die()s if it fails) if the
+list needs to grow, it is safe not to check the pointer. I.e. you may
+write `path_list_insert(...)->util = ...;`.
+
+`path_list_lookup`::
+
+	Look up a given string in the path_list, returning the containing
+	path_list_item. If the string is not found, NULL is returned.
+
+* Functions for unsorted lists only
+
+`path_list_append`::
+
+	Append a new string to the end of the path_list.
+
+`sort_path_list`::
+
+	Make an unsorted list sorted.
+
+`unsorted_path_list_has_path`::
+
+	It's like `path_list_has_path()` but for unsorted lists.
++
+This function needs to look through all items, as opposed to its
+counterpart for sorted lists, which performs a binary search.
+
+Data structures
+---------------
+
+* `struct path_list_item`
+
+Represents an item of the list. The `path` member is a pointer to the
+string, and you may use the `util` member for any purpose, if you want.
+
+* `struct path_list`
+
+Represents the list itself.
+
+. The array of items are available via the `items` member.
+. The `nr` member contains the number of items stored in the list.
+. The `alloc` member is used to avoid reallocating at every insertion.
+  You should not tamper with it.
+. Setting the `strdup_paths` member to 1 will strdup() the strings
+  before adding them, see above.

@@ -1,9 +1,126 @@
 path-list API
 =============
 
-Talk about <path-list.h>, things like
+The path_list API offers a data structure and functions to handle sorted
+and unsorted string lists.
 
-* it is not just paths but strings in general;
-* the calling sequence.
+The name is a bit misleading, a path_list may store not only paths but
+strings in general.
 
-(Dscho)
+The caller:
+
+. Allocates and clears a `struct path_list` variable.
+
+. Initializes the members. You might want to set the flag `strdup_paths`
+  if the strings should be strdup()ed. For example, this is necessary
+  when you add something like git_path("..."), since that function returns
+  a static buffer that will change with the next call to git_path().
++
+If you need something advanced, you can manually malloc() the `items`
+member (you need this if you add things later) and you should set the
+`nr` and `alloc` members in that case, too.
+
+. Adds new items to the list, using `path_list_append` or `path_list_insert`.
+
+. Can check if a string is in the list using `path_list_has_path` or
+  `unsorted_path_list_has_path` and get it from the list using
+  `path_list_lookup` for sorted lists.
+
+. Can sort an unsorted list using `sort_path_list`.
+
+. Finally it should free the list using `path_list_clear`.
+
+Example:
+
+----
+struct path_list list;
+int i;
+
+memset(&list, 0, sizeof(struct path_list));
+path_list_append("foo", &list);
+path_list_append("bar", &list);
+for (i = 0; i < list.nr; i++)
+	printf("%s\n", list.items[i].path)
+----
+
+NOTE: It is more efficient to build an unsorted list and sort it
+afterwards, instead of building a sorted list (`O(n log n)` instead of
+`O(n^2)`).
++
+However, if you use the list to check if a certain string was added
+already, you should not do that (using unsorted_path_list_has_path()),
+because the complexity would be quadratic again (but with a worse factor).
+
+Functions
+---------
+
+* General ones (works with sorted and unsorted lists as well)
+
+`print_path_list`::
+
+	Dump a path_list to stdout, useful mainly for debugging purposes. It
+	can take an optional header argument and it writes out the
+	string-pointer pairs of the path_list, each one in its own line.
+
+`path_list_clear`::
+
+	Free a path_list. The `path` pointer of the items will be freed in case
+	the `strdup_paths` member of the path_list is set. The second parameter
+	controls if the `util` pointer of the items should be freed or not.
+
+* Functions for sorted lists only
+
+`path_list_has_path`::
+
+	Determine if the path_list has a given string or not.
+
+`path_list_insert`::
+
+	Insert a new element to the path_list. The returned pointer can be handy
+	if you want to write something to the `util` pointer of the
+	path_list_item containing the just added string.
++
+Since this function uses xrealloc() (which die()s if it fails) if the
+list needs to grow, it is safe not to check the pointer. I.e. you may
+write `path_list_insert(...)->util = ...;`.
+
+`path_list_lookup`::
+
+	Look up a given string in the path_list, returning the containing
+	path_list_item. If the string is not found, NULL is returned.
+
+* Functions for unsorted lists only
+
+`path_list_append`::
+
+	Append a new string to the end of the path_list.
+
+`sort_path_list`::
+
+	Make an unsorted list sorted.
+
+`unsorted_path_list_has_path`::
+
+	It's like `path_list_has_path()` but for unsorted lists.
++
+This function needs to look through all items, as opposed to its
+counterpart for sorted lists, which performs a binary search.
+
+Data structures
+---------------
+
+* `struct path_list_item`
+
+Represents an item of the list. The `path` member is a pointer to the
+string, and you may use the `util` member for any purpose, if you want.
+
+* `struct path_list`
+
+Represents the list itself.
+
+. The array of items are available via the `items` member.
+. The `nr` member contains the number of items stored in the list.
+. The `alloc` member is used to avoid reallocating at every insertion.
+  You should not tamper with it.
+. Setting the `strdup_paths` member to 1 will strdup() the strings
+  before adding them, see above.