git/path.c

1480 lines
35 KiB
C
Raw Normal View History

/*
* Utilities for paths and pathnames
*/
#include "cache.h"
#include "repository.h"
#include "strbuf.h"
#include "string-list.h"
#include "dir.h"
#include "worktree.h"
allow do_submodule_path to work even if submodule isn't checked out Currently, do_submodule_path will attempt locating the .git directory by using read_gitfile on <path>/.git. If this fails it just assumes the <path>/.git is actually a git directory. This is good because it allows for handling submodules which were cloned in a regular manner first before being added to the superproject. Unfortunately this fails if the <path> is not actually checked out any longer, such as by removing the directory. Fix this by checking if the directory we found is actually a gitdir. In the case it is not, attempt to lookup the submodule configuration and find the name of where it is stored in the .git/modules/ directory of the superproject. If we can't locate the submodule configuration, this might occur because for example a submodule gitlink was added but the corresponding .gitmodules file was not properly updated. A die() here would not be pleasant to the users of submodule diff formats, so instead, modify do_submodule_path() to return an error code: - git_pathdup_submodule() returns NULL when we fail to find a path. - strbuf_git_path_submodule() propagates the error code to the caller. Modify the callers of these functions to check the error code and fail properly. This ensures we don't attempt to use a bad path that doesn't match the corresponding submodule. Because this change fixes add_submodule_odb() to work even if the submodule is not checked out, update the wording of the submodule log diff format to correctly display that the submodule is "not initialized" instead of "not checked out" Add tests to ensure this change works as expected. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2016-08-31 23:27:22 +00:00
#include "submodule-config.h"
#include "path.h"
#include "packfile.h"
#include "object-store.h"
static int get_st_mode_bits(const char *path, int *mode)
{
struct stat st;
if (lstat(path, &st) < 0)
return -1;
*mode = st.st_mode;
return 0;
}
static char bad_path[] = "/bad-path/";
static struct strbuf *get_pathname(void)
{
static struct strbuf pathname_array[4] = {
STRBUF_INIT, STRBUF_INIT, STRBUF_INIT, STRBUF_INIT
};
static int index;
struct strbuf *sb = &pathname_array[index];
index = (index + 1) % ARRAY_SIZE(pathname_array);
strbuf_reset(sb);
return sb;
}
path.c: fix uninitialized memory access In cleanup_path we're passing in a char array, run a memcmp on it, and run through it without ever checking if something is in the array in the first place. This can lead us to access uninitialized memory, for example in t5541-http-push-smart.sh test 7, when run under valgrind: ==4423== Conditional jump or move depends on uninitialised value(s) ==4423== at 0x242FA9: cleanup_path (path.c:35) ==4423== by 0x242FA9: mkpath (path.c:456) ==4423== by 0x256CC7: refname_match (refs.c:364) ==4423== by 0x26C181: count_refspec_match (remote.c:1015) ==4423== by 0x26C181: match_explicit_lhs (remote.c:1126) ==4423== by 0x26C181: check_push_refs (remote.c:1409) ==4423== by 0x2ABB4D: transport_push (transport.c:870) ==4423== by 0x186703: push_with_options (push.c:332) ==4423== by 0x18746D: do_push (push.c:409) ==4423== by 0x18746D: cmd_push (push.c:566) ==4423== by 0x1183E0: run_builtin (git.c:352) ==4423== by 0x11973E: handle_builtin (git.c:539) ==4423== by 0x11973E: run_argv (git.c:593) ==4423== by 0x11973E: main (git.c:698) ==4423== Uninitialised value was created by a heap allocation ==4423== at 0x4C2CD8F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) ==4423== by 0x4C2F195: realloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) ==4423== by 0x2C196B: xrealloc (wrapper.c:137) ==4423== by 0x29A30B: strbuf_grow (strbuf.c:66) ==4423== by 0x29A30B: strbuf_vaddf (strbuf.c:277) ==4423== by 0x242F9F: mkpath (path.c:454) ==4423== by 0x256CC7: refname_match (refs.c:364) ==4423== by 0x26C181: count_refspec_match (remote.c:1015) ==4423== by 0x26C181: match_explicit_lhs (remote.c:1126) ==4423== by 0x26C181: check_push_refs (remote.c:1409) ==4423== by 0x2ABB4D: transport_push (transport.c:870) ==4423== by 0x186703: push_with_options (push.c:332) ==4423== by 0x18746D: do_push (push.c:409) ==4423== by 0x18746D: cmd_push (push.c:566) ==4423== by 0x1183E0: run_builtin (git.c:352) ==4423== by 0x11973E: handle_builtin (git.c:539) ==4423== by 0x11973E: run_argv (git.c:593) ==4423== by 0x11973E: main (git.c:698) ==4423== Avoid this by using skip_prefix(), which knows not to go beyond the end of the string. Reported-by: Thomas Gummerer <t.gummerer@gmail.com> Signed-off-by: Jeff King <peff@peff.net> Reviewed-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-10-03 23:30:40 +00:00
static const char *cleanup_path(const char *path)
{
/* Clean it up */
path.c: fix uninitialized memory access In cleanup_path we're passing in a char array, run a memcmp on it, and run through it without ever checking if something is in the array in the first place. This can lead us to access uninitialized memory, for example in t5541-http-push-smart.sh test 7, when run under valgrind: ==4423== Conditional jump or move depends on uninitialised value(s) ==4423== at 0x242FA9: cleanup_path (path.c:35) ==4423== by 0x242FA9: mkpath (path.c:456) ==4423== by 0x256CC7: refname_match (refs.c:364) ==4423== by 0x26C181: count_refspec_match (remote.c:1015) ==4423== by 0x26C181: match_explicit_lhs (remote.c:1126) ==4423== by 0x26C181: check_push_refs (remote.c:1409) ==4423== by 0x2ABB4D: transport_push (transport.c:870) ==4423== by 0x186703: push_with_options (push.c:332) ==4423== by 0x18746D: do_push (push.c:409) ==4423== by 0x18746D: cmd_push (push.c:566) ==4423== by 0x1183E0: run_builtin (git.c:352) ==4423== by 0x11973E: handle_builtin (git.c:539) ==4423== by 0x11973E: run_argv (git.c:593) ==4423== by 0x11973E: main (git.c:698) ==4423== Uninitialised value was created by a heap allocation ==4423== at 0x4C2CD8F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) ==4423== by 0x4C2F195: realloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) ==4423== by 0x2C196B: xrealloc (wrapper.c:137) ==4423== by 0x29A30B: strbuf_grow (strbuf.c:66) ==4423== by 0x29A30B: strbuf_vaddf (strbuf.c:277) ==4423== by 0x242F9F: mkpath (path.c:454) ==4423== by 0x256CC7: refname_match (refs.c:364) ==4423== by 0x26C181: count_refspec_match (remote.c:1015) ==4423== by 0x26C181: match_explicit_lhs (remote.c:1126) ==4423== by 0x26C181: check_push_refs (remote.c:1409) ==4423== by 0x2ABB4D: transport_push (transport.c:870) ==4423== by 0x186703: push_with_options (push.c:332) ==4423== by 0x18746D: do_push (push.c:409) ==4423== by 0x18746D: cmd_push (push.c:566) ==4423== by 0x1183E0: run_builtin (git.c:352) ==4423== by 0x11973E: handle_builtin (git.c:539) ==4423== by 0x11973E: run_argv (git.c:593) ==4423== by 0x11973E: main (git.c:698) ==4423== Avoid this by using skip_prefix(), which knows not to go beyond the end of the string. Reported-by: Thomas Gummerer <t.gummerer@gmail.com> Signed-off-by: Jeff King <peff@peff.net> Reviewed-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-10-03 23:30:40 +00:00
if (skip_prefix(path, "./", &path)) {
while (*path == '/')
path++;
}
return path;
}
static void strbuf_cleanup_path(struct strbuf *sb)
{
path.c: fix uninitialized memory access In cleanup_path we're passing in a char array, run a memcmp on it, and run through it without ever checking if something is in the array in the first place. This can lead us to access uninitialized memory, for example in t5541-http-push-smart.sh test 7, when run under valgrind: ==4423== Conditional jump or move depends on uninitialised value(s) ==4423== at 0x242FA9: cleanup_path (path.c:35) ==4423== by 0x242FA9: mkpath (path.c:456) ==4423== by 0x256CC7: refname_match (refs.c:364) ==4423== by 0x26C181: count_refspec_match (remote.c:1015) ==4423== by 0x26C181: match_explicit_lhs (remote.c:1126) ==4423== by 0x26C181: check_push_refs (remote.c:1409) ==4423== by 0x2ABB4D: transport_push (transport.c:870) ==4423== by 0x186703: push_with_options (push.c:332) ==4423== by 0x18746D: do_push (push.c:409) ==4423== by 0x18746D: cmd_push (push.c:566) ==4423== by 0x1183E0: run_builtin (git.c:352) ==4423== by 0x11973E: handle_builtin (git.c:539) ==4423== by 0x11973E: run_argv (git.c:593) ==4423== by 0x11973E: main (git.c:698) ==4423== Uninitialised value was created by a heap allocation ==4423== at 0x4C2CD8F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) ==4423== by 0x4C2F195: realloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) ==4423== by 0x2C196B: xrealloc (wrapper.c:137) ==4423== by 0x29A30B: strbuf_grow (strbuf.c:66) ==4423== by 0x29A30B: strbuf_vaddf (strbuf.c:277) ==4423== by 0x242F9F: mkpath (path.c:454) ==4423== by 0x256CC7: refname_match (refs.c:364) ==4423== by 0x26C181: count_refspec_match (remote.c:1015) ==4423== by 0x26C181: match_explicit_lhs (remote.c:1126) ==4423== by 0x26C181: check_push_refs (remote.c:1409) ==4423== by 0x2ABB4D: transport_push (transport.c:870) ==4423== by 0x186703: push_with_options (push.c:332) ==4423== by 0x18746D: do_push (push.c:409) ==4423== by 0x18746D: cmd_push (push.c:566) ==4423== by 0x1183E0: run_builtin (git.c:352) ==4423== by 0x11973E: handle_builtin (git.c:539) ==4423== by 0x11973E: run_argv (git.c:593) ==4423== by 0x11973E: main (git.c:698) ==4423== Avoid this by using skip_prefix(), which knows not to go beyond the end of the string. Reported-by: Thomas Gummerer <t.gummerer@gmail.com> Signed-off-by: Jeff King <peff@peff.net> Reviewed-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-10-03 23:30:40 +00:00
const char *path = cleanup_path(sb->buf);
if (path > sb->buf)
strbuf_remove(sb, 0, path - sb->buf);
}
char *mksnpath(char *buf, size_t n, const char *fmt, ...)
{
va_list args;
unsigned len;
va_start(args, fmt);
len = vsnprintf(buf, n, fmt, args);
va_end(args);
if (len >= n) {
strlcpy(buf, bad_path, n);
return buf;
}
path.c: fix uninitialized memory access In cleanup_path we're passing in a char array, run a memcmp on it, and run through it without ever checking if something is in the array in the first place. This can lead us to access uninitialized memory, for example in t5541-http-push-smart.sh test 7, when run under valgrind: ==4423== Conditional jump or move depends on uninitialised value(s) ==4423== at 0x242FA9: cleanup_path (path.c:35) ==4423== by 0x242FA9: mkpath (path.c:456) ==4423== by 0x256CC7: refname_match (refs.c:364) ==4423== by 0x26C181: count_refspec_match (remote.c:1015) ==4423== by 0x26C181: match_explicit_lhs (remote.c:1126) ==4423== by 0x26C181: check_push_refs (remote.c:1409) ==4423== by 0x2ABB4D: transport_push (transport.c:870) ==4423== by 0x186703: push_with_options (push.c:332) ==4423== by 0x18746D: do_push (push.c:409) ==4423== by 0x18746D: cmd_push (push.c:566) ==4423== by 0x1183E0: run_builtin (git.c:352) ==4423== by 0x11973E: handle_builtin (git.c:539) ==4423== by 0x11973E: run_argv (git.c:593) ==4423== by 0x11973E: main (git.c:698) ==4423== Uninitialised value was created by a heap allocation ==4423== at 0x4C2CD8F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) ==4423== by 0x4C2F195: realloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) ==4423== by 0x2C196B: xrealloc (wrapper.c:137) ==4423== by 0x29A30B: strbuf_grow (strbuf.c:66) ==4423== by 0x29A30B: strbuf_vaddf (strbuf.c:277) ==4423== by 0x242F9F: mkpath (path.c:454) ==4423== by 0x256CC7: refname_match (refs.c:364) ==4423== by 0x26C181: count_refspec_match (remote.c:1015) ==4423== by 0x26C181: match_explicit_lhs (remote.c:1126) ==4423== by 0x26C181: check_push_refs (remote.c:1409) ==4423== by 0x2ABB4D: transport_push (transport.c:870) ==4423== by 0x186703: push_with_options (push.c:332) ==4423== by 0x18746D: do_push (push.c:409) ==4423== by 0x18746D: cmd_push (push.c:566) ==4423== by 0x1183E0: run_builtin (git.c:352) ==4423== by 0x11973E: handle_builtin (git.c:539) ==4423== by 0x11973E: run_argv (git.c:593) ==4423== by 0x11973E: main (git.c:698) ==4423== Avoid this by using skip_prefix(), which knows not to go beyond the end of the string. Reported-by: Thomas Gummerer <t.gummerer@gmail.com> Signed-off-by: Jeff King <peff@peff.net> Reviewed-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-10-03 23:30:40 +00:00
return (char *)cleanup_path(buf);
}
2014-11-30 08:24:31 +00:00
static int dir_prefix(const char *buf, const char *dir)
{
2014-11-30 08:24:31 +00:00
int len = strlen(dir);
return !strncmp(buf, dir, len) &&
(is_dir_sep(buf[len]) || buf[len] == '\0');
}
2014-11-30 08:24:31 +00:00
/* $buf =~ m|$dir/+$file| but without regex */
static int is_dir_file(const char *buf, const char *dir, const char *file)
{
int len = strlen(dir);
if (strncmp(buf, dir, len) || !is_dir_sep(buf[len]))
return 0;
while (is_dir_sep(buf[len]))
len++;
return !strcmp(buf + len, file);
}
static void replace_dir(struct strbuf *buf, int len, const char *newdir)
{
int newlen = strlen(newdir);
int need_sep = (buf->buf[len] && !is_dir_sep(buf->buf[len])) &&
!is_dir_sep(newdir[newlen - 1]);
if (need_sep)
len--; /* keep one char, to be replaced with '/' */
strbuf_splice(buf, 0, len, newdir, newlen);
if (need_sep)
buf->buf[newlen] = '/';
}
struct common_dir {
/* Not considered garbage for report_linked_checkout_garbage */
unsigned ignore_garbage:1;
unsigned is_dir:1;
/* Not common even though its parent is */
unsigned exclude:1;
const char *dirname;
};
static struct common_dir common_list[] = {
{ 0, 1, 0, "branches" },
{ 0, 1, 0, "common" },
{ 0, 1, 0, "hooks" },
{ 0, 1, 0, "info" },
{ 0, 0, 1, "info/sparse-checkout" },
{ 1, 1, 0, "logs" },
{ 1, 1, 1, "logs/HEAD" },
{ 0, 1, 1, "logs/refs/bisect" },
{ 0, 1, 1, "logs/refs/rewritten" },
{ 0, 1, 1, "logs/refs/worktree" },
{ 0, 1, 0, "lost-found" },
{ 0, 1, 0, "objects" },
{ 0, 1, 0, "refs" },
{ 0, 1, 1, "refs/bisect" },
{ 0, 1, 1, "refs/rewritten" },
{ 0, 1, 1, "refs/worktree" },
{ 0, 1, 0, "remotes" },
{ 0, 1, 0, "worktrees" },
{ 0, 1, 0, "rr-cache" },
{ 0, 1, 0, "svn" },
{ 0, 0, 0, "config" },
{ 1, 0, 0, "gc.pid" },
{ 0, 0, 0, "packed-refs" },
{ 0, 0, 0, "shallow" },
{ 0, 0, 0, NULL }
$GIT_COMMON_DIR: a new environment variable This variable is intended to support multiple working directories attached to a repository. Such a repository may have a main working directory, created by either "git init" or "git clone" and one or more linked working directories. These working directories and the main repository share the same repository directory. In linked working directories, $GIT_COMMON_DIR must be defined to point to the real repository directory and $GIT_DIR points to an unused subdirectory inside $GIT_COMMON_DIR. File locations inside the repository are reorganized from the linked worktree view point: - worktree-specific such as HEAD, logs/HEAD, index, other top-level refs and unrecognized files are from $GIT_DIR. - the rest like objects, refs, info, hooks, packed-refs, shallow... are from $GIT_COMMON_DIR (except info/sparse-checkout, but that's a separate patch) Scripts are supposed to retrieve paths in $GIT_DIR with "git rev-parse --git-path", which will take care of "$GIT_DIR vs $GIT_COMMON_DIR" business. The redirection is done by git_path(), git_pathdup() and strbuf_git_path(). The selected list of paths goes to $GIT_COMMON_DIR, not the other way around in case a developer adds a new worktree-specific file and it's accidentally promoted to be shared across repositories (this includes unknown files added by third party commands) The list of known files that belong to $GIT_DIR are: ADD_EDIT.patch BISECT_ANCESTORS_OK BISECT_EXPECTED_REV BISECT_LOG BISECT_NAMES CHERRY_PICK_HEAD COMMIT_MSG FETCH_HEAD HEAD MERGE_HEAD MERGE_MODE MERGE_RR NOTES_EDITMSG NOTES_MERGE_WORKTREE ORIG_HEAD REVERT_HEAD SQUASH_MSG TAG_EDITMSG fast_import_crash_* logs/HEAD next-index-* rebase-apply rebase-merge rsync-refs-* sequencer/* shallow_* Path mapping is NOT done for git_path_submodule(). Multi-checkouts are not supported as submodules. Helped-by: Jens Lehmann <Jens.Lehmann@web.de> Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-11-30 08:24:36 +00:00
};
/*
* A compressed trie. A trie node consists of zero or more characters that
* are common to all elements with this prefix, optionally followed by some
* children. If value is not NULL, the trie node is a terminal node.
*
* For example, consider the following set of strings:
* abc
* def
* definite
* definition
*
* The trie would look like:
* root: len = 0, children a and d non-NULL, value = NULL.
* a: len = 2, contents = bc, value = (data for "abc")
* d: len = 2, contents = ef, children i non-NULL, value = (data for "def")
* i: len = 3, contents = nit, children e and i non-NULL, value = NULL
* e: len = 0, children all NULL, value = (data for "definite")
* i: len = 2, contents = on, children all NULL,
* value = (data for "definition")
*/
struct trie {
struct trie *children[256];
int len;
char *contents;
void *value;
};
static struct trie *make_trie_node(const char *key, void *value)
$GIT_COMMON_DIR: a new environment variable This variable is intended to support multiple working directories attached to a repository. Such a repository may have a main working directory, created by either "git init" or "git clone" and one or more linked working directories. These working directories and the main repository share the same repository directory. In linked working directories, $GIT_COMMON_DIR must be defined to point to the real repository directory and $GIT_DIR points to an unused subdirectory inside $GIT_COMMON_DIR. File locations inside the repository are reorganized from the linked worktree view point: - worktree-specific such as HEAD, logs/HEAD, index, other top-level refs and unrecognized files are from $GIT_DIR. - the rest like objects, refs, info, hooks, packed-refs, shallow... are from $GIT_COMMON_DIR (except info/sparse-checkout, but that's a separate patch) Scripts are supposed to retrieve paths in $GIT_DIR with "git rev-parse --git-path", which will take care of "$GIT_DIR vs $GIT_COMMON_DIR" business. The redirection is done by git_path(), git_pathdup() and strbuf_git_path(). The selected list of paths goes to $GIT_COMMON_DIR, not the other way around in case a developer adds a new worktree-specific file and it's accidentally promoted to be shared across repositories (this includes unknown files added by third party commands) The list of known files that belong to $GIT_DIR are: ADD_EDIT.patch BISECT_ANCESTORS_OK BISECT_EXPECTED_REV BISECT_LOG BISECT_NAMES CHERRY_PICK_HEAD COMMIT_MSG FETCH_HEAD HEAD MERGE_HEAD MERGE_MODE MERGE_RR NOTES_EDITMSG NOTES_MERGE_WORKTREE ORIG_HEAD REVERT_HEAD SQUASH_MSG TAG_EDITMSG fast_import_crash_* logs/HEAD next-index-* rebase-apply rebase-merge rsync-refs-* sequencer/* shallow_* Path mapping is NOT done for git_path_submodule(). Multi-checkouts are not supported as submodules. Helped-by: Jens Lehmann <Jens.Lehmann@web.de> Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-11-30 08:24:36 +00:00
{
struct trie *new_node = xcalloc(1, sizeof(*new_node));
new_node->len = strlen(key);
if (new_node->len) {
new_node->contents = xmalloc(new_node->len);
memcpy(new_node->contents, key, new_node->len);
}
new_node->value = value;
return new_node;
}
$GIT_COMMON_DIR: a new environment variable This variable is intended to support multiple working directories attached to a repository. Such a repository may have a main working directory, created by either "git init" or "git clone" and one or more linked working directories. These working directories and the main repository share the same repository directory. In linked working directories, $GIT_COMMON_DIR must be defined to point to the real repository directory and $GIT_DIR points to an unused subdirectory inside $GIT_COMMON_DIR. File locations inside the repository are reorganized from the linked worktree view point: - worktree-specific such as HEAD, logs/HEAD, index, other top-level refs and unrecognized files are from $GIT_DIR. - the rest like objects, refs, info, hooks, packed-refs, shallow... are from $GIT_COMMON_DIR (except info/sparse-checkout, but that's a separate patch) Scripts are supposed to retrieve paths in $GIT_DIR with "git rev-parse --git-path", which will take care of "$GIT_DIR vs $GIT_COMMON_DIR" business. The redirection is done by git_path(), git_pathdup() and strbuf_git_path(). The selected list of paths goes to $GIT_COMMON_DIR, not the other way around in case a developer adds a new worktree-specific file and it's accidentally promoted to be shared across repositories (this includes unknown files added by third party commands) The list of known files that belong to $GIT_DIR are: ADD_EDIT.patch BISECT_ANCESTORS_OK BISECT_EXPECTED_REV BISECT_LOG BISECT_NAMES CHERRY_PICK_HEAD COMMIT_MSG FETCH_HEAD HEAD MERGE_HEAD MERGE_MODE MERGE_RR NOTES_EDITMSG NOTES_MERGE_WORKTREE ORIG_HEAD REVERT_HEAD SQUASH_MSG TAG_EDITMSG fast_import_crash_* logs/HEAD next-index-* rebase-apply rebase-merge rsync-refs-* sequencer/* shallow_* Path mapping is NOT done for git_path_submodule(). Multi-checkouts are not supported as submodules. Helped-by: Jens Lehmann <Jens.Lehmann@web.de> Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-11-30 08:24:36 +00:00
/*
* Add a key/value pair to a trie. The key is assumed to be \0-terminated.
* If there was an existing value for this key, return it.
*/
static void *add_to_trie(struct trie *root, const char *key, void *value)
{
struct trie *child;
void *old;
int i;
if (!*key) {
/* we have reached the end of the key */
old = root->value;
root->value = value;
return old;
}
for (i = 0; i < root->len; i++) {
if (root->contents[i] == key[i])
continue;
/*
* Split this node: child will contain this node's
* existing children.
*/
child = xmalloc(sizeof(*child));
memcpy(child->children, root->children, sizeof(root->children));
child->len = root->len - i - 1;
if (child->len) {
child->contents = xstrndup(root->contents + i + 1,
child->len);
$GIT_COMMON_DIR: a new environment variable This variable is intended to support multiple working directories attached to a repository. Such a repository may have a main working directory, created by either "git init" or "git clone" and one or more linked working directories. These working directories and the main repository share the same repository directory. In linked working directories, $GIT_COMMON_DIR must be defined to point to the real repository directory and $GIT_DIR points to an unused subdirectory inside $GIT_COMMON_DIR. File locations inside the repository are reorganized from the linked worktree view point: - worktree-specific such as HEAD, logs/HEAD, index, other top-level refs and unrecognized files are from $GIT_DIR. - the rest like objects, refs, info, hooks, packed-refs, shallow... are from $GIT_COMMON_DIR (except info/sparse-checkout, but that's a separate patch) Scripts are supposed to retrieve paths in $GIT_DIR with "git rev-parse --git-path", which will take care of "$GIT_DIR vs $GIT_COMMON_DIR" business. The redirection is done by git_path(), git_pathdup() and strbuf_git_path(). The selected list of paths goes to $GIT_COMMON_DIR, not the other way around in case a developer adds a new worktree-specific file and it's accidentally promoted to be shared across repositories (this includes unknown files added by third party commands) The list of known files that belong to $GIT_DIR are: ADD_EDIT.patch BISECT_ANCESTORS_OK BISECT_EXPECTED_REV BISECT_LOG BISECT_NAMES CHERRY_PICK_HEAD COMMIT_MSG FETCH_HEAD HEAD MERGE_HEAD MERGE_MODE MERGE_RR NOTES_EDITMSG NOTES_MERGE_WORKTREE ORIG_HEAD REVERT_HEAD SQUASH_MSG TAG_EDITMSG fast_import_crash_* logs/HEAD next-index-* rebase-apply rebase-merge rsync-refs-* sequencer/* shallow_* Path mapping is NOT done for git_path_submodule(). Multi-checkouts are not supported as submodules. Helped-by: Jens Lehmann <Jens.Lehmann@web.de> Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-11-30 08:24:36 +00:00
}
child->value = root->value;
root->value = NULL;
root->len = i;
memset(root->children, 0, sizeof(root->children));
root->children[(unsigned char)root->contents[i]] = child;
/* This is the newly-added child. */
root->children[(unsigned char)key[i]] =
make_trie_node(key + i + 1, value);
return NULL;
}
/* We have matched the entire compressed section */
if (key[i]) {
child = root->children[(unsigned char)key[root->len]];
if (child) {
return add_to_trie(child, key + root->len + 1, value);
} else {
child = make_trie_node(key + root->len + 1, value);
root->children[(unsigned char)key[root->len]] = child;
return NULL;
$GIT_COMMON_DIR: a new environment variable This variable is intended to support multiple working directories attached to a repository. Such a repository may have a main working directory, created by either "git init" or "git clone" and one or more linked working directories. These working directories and the main repository share the same repository directory. In linked working directories, $GIT_COMMON_DIR must be defined to point to the real repository directory and $GIT_DIR points to an unused subdirectory inside $GIT_COMMON_DIR. File locations inside the repository are reorganized from the linked worktree view point: - worktree-specific such as HEAD, logs/HEAD, index, other top-level refs and unrecognized files are from $GIT_DIR. - the rest like objects, refs, info, hooks, packed-refs, shallow... are from $GIT_COMMON_DIR (except info/sparse-checkout, but that's a separate patch) Scripts are supposed to retrieve paths in $GIT_DIR with "git rev-parse --git-path", which will take care of "$GIT_DIR vs $GIT_COMMON_DIR" business. The redirection is done by git_path(), git_pathdup() and strbuf_git_path(). The selected list of paths goes to $GIT_COMMON_DIR, not the other way around in case a developer adds a new worktree-specific file and it's accidentally promoted to be shared across repositories (this includes unknown files added by third party commands) The list of known files that belong to $GIT_DIR are: ADD_EDIT.patch BISECT_ANCESTORS_OK BISECT_EXPECTED_REV BISECT_LOG BISECT_NAMES CHERRY_PICK_HEAD COMMIT_MSG FETCH_HEAD HEAD MERGE_HEAD MERGE_MODE MERGE_RR NOTES_EDITMSG NOTES_MERGE_WORKTREE ORIG_HEAD REVERT_HEAD SQUASH_MSG TAG_EDITMSG fast_import_crash_* logs/HEAD next-index-* rebase-apply rebase-merge rsync-refs-* sequencer/* shallow_* Path mapping is NOT done for git_path_submodule(). Multi-checkouts are not supported as submodules. Helped-by: Jens Lehmann <Jens.Lehmann@web.de> Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-11-30 08:24:36 +00:00
}
}
old = root->value;
root->value = value;
return old;
}
typedef int (*match_fn)(const char *unmatched, void *data, void *baton);
/*
* Search a trie for some key. Find the longest /-or-\0-terminated
* prefix of the key for which the trie contains a value. Call fn
* with the unmatched portion of the key and the found value, and
* return its return value. If there is no such prefix, return -1.
*
* The key is partially normalized: consecutive slashes are skipped.
*
* For example, consider the trie containing only [refs,
* refs/worktree] (both with values).
*
* | key | unmatched | val from node | return value |
* |-----------------|------------|---------------|--------------|
* | a | not called | n/a | -1 |
* | refs | \0 | refs | as per fn |
* | refs/ | / | refs | as per fn |
* | refs/w | /w | refs | as per fn |
* | refs/worktree | \0 | refs/worktree | as per fn |
* | refs/worktree/ | / | refs/worktree | as per fn |
* | refs/worktree/a | /a | refs/worktree | as per fn |
* |-----------------|------------|---------------|--------------|
*
*/
static int trie_find(struct trie *root, const char *key, match_fn fn,
void *baton)
{
int i;
int result;
struct trie *child;
if (!*key) {
/* we have reached the end of the key */
if (root->value && !root->len)
return fn(key, root->value, baton);
else
return -1;
}
for (i = 0; i < root->len; i++) {
/* Partial path normalization: skip consecutive slashes. */
if (key[i] == '/' && key[i+1] == '/') {
key++;
continue;
}
if (root->contents[i] != key[i])
return -1;
}
/* Matched the entire compressed section */
key += i;
if (!*key)
/* End of key */
return fn(key, root->value, baton);
/* Partial path normalization: skip consecutive slashes */
while (key[0] == '/' && key[1] == '/')
key++;
child = root->children[(unsigned char)*key];
if (child)
result = trie_find(child, key + 1, fn, baton);
else
result = -1;
if (result >= 0 || (*key != '/' && *key != 0))
return result;
if (root->value)
return fn(key, root->value, baton);
else
return -1;
}
static struct trie common_trie;
static int common_trie_done_setup;
static void init_common_trie(void)
{
struct common_dir *p;
if (common_trie_done_setup)
return;
for (p = common_list; p->dirname; p++)
add_to_trie(&common_trie, p->dirname, p);
common_trie_done_setup = 1;
}
/*
* Helper function for update_common_dir: returns 1 if the dir
* prefix is common.
*/
static int check_common(const char *unmatched, void *value, void *baton)
{
struct common_dir *dir = value;
if (!dir)
return 0;
if (dir->is_dir && (unmatched[0] == 0 || unmatched[0] == '/'))
return !dir->exclude;
if (!dir->is_dir && unmatched[0] == 0)
return !dir->exclude;
return 0;
}
static void update_common_dir(struct strbuf *buf, int git_dir_len,
const char *common_dir)
{
char *base = buf->buf + git_dir_len;
init_common_trie();
if (trie_find(&common_trie, base, check_common, NULL) > 0)
replace_dir(buf, git_dir_len, common_dir);
$GIT_COMMON_DIR: a new environment variable This variable is intended to support multiple working directories attached to a repository. Such a repository may have a main working directory, created by either "git init" or "git clone" and one or more linked working directories. These working directories and the main repository share the same repository directory. In linked working directories, $GIT_COMMON_DIR must be defined to point to the real repository directory and $GIT_DIR points to an unused subdirectory inside $GIT_COMMON_DIR. File locations inside the repository are reorganized from the linked worktree view point: - worktree-specific such as HEAD, logs/HEAD, index, other top-level refs and unrecognized files are from $GIT_DIR. - the rest like objects, refs, info, hooks, packed-refs, shallow... are from $GIT_COMMON_DIR (except info/sparse-checkout, but that's a separate patch) Scripts are supposed to retrieve paths in $GIT_DIR with "git rev-parse --git-path", which will take care of "$GIT_DIR vs $GIT_COMMON_DIR" business. The redirection is done by git_path(), git_pathdup() and strbuf_git_path(). The selected list of paths goes to $GIT_COMMON_DIR, not the other way around in case a developer adds a new worktree-specific file and it's accidentally promoted to be shared across repositories (this includes unknown files added by third party commands) The list of known files that belong to $GIT_DIR are: ADD_EDIT.patch BISECT_ANCESTORS_OK BISECT_EXPECTED_REV BISECT_LOG BISECT_NAMES CHERRY_PICK_HEAD COMMIT_MSG FETCH_HEAD HEAD MERGE_HEAD MERGE_MODE MERGE_RR NOTES_EDITMSG NOTES_MERGE_WORKTREE ORIG_HEAD REVERT_HEAD SQUASH_MSG TAG_EDITMSG fast_import_crash_* logs/HEAD next-index-* rebase-apply rebase-merge rsync-refs-* sequencer/* shallow_* Path mapping is NOT done for git_path_submodule(). Multi-checkouts are not supported as submodules. Helped-by: Jens Lehmann <Jens.Lehmann@web.de> Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-11-30 08:24:36 +00:00
}
void report_linked_checkout_garbage(void)
{
struct strbuf sb = STRBUF_INIT;
const struct common_dir *p;
int len;
if (!the_repository->different_commondir)
return;
strbuf_addf(&sb, "%s/", get_git_dir());
len = sb.len;
for (p = common_list; p->dirname; p++) {
const char *path = p->dirname;
if (p->ignore_garbage)
continue;
strbuf_setlen(&sb, len);
strbuf_addstr(&sb, path);
if (file_exists(sb.buf))
report_garbage(PACKDIR_FILE_GARBAGE, sb.buf);
}
strbuf_release(&sb);
}
static void adjust_git_path(const struct repository *repo,
struct strbuf *buf, int git_dir_len)
2014-11-30 08:24:31 +00:00
{
const char *base = buf->buf + git_dir_len;
if (is_dir_file(base, "info", "grafts"))
2014-11-30 08:24:31 +00:00
strbuf_splice(buf, 0, buf->len,
repo->graft_file, strlen(repo->graft_file));
else if (!strcmp(base, "index"))
2014-11-30 08:24:31 +00:00
strbuf_splice(buf, 0, buf->len,
repo->index_file, strlen(repo->index_file));
else if (dir_prefix(base, "objects"))
sha1-file: use an object_directory for the main object dir Our handling of alternate object directories is needlessly different from the main object directory. As a result, many places in the code basically look like this: do_something(r->objects->objdir); for (odb = r->objects->alt_odb_list; odb; odb = odb->next) do_something(odb->path); That gets annoying when do_something() is non-trivial, and we've resorted to gross hacks like creating fake alternates (see find_short_object_filename()). Instead, let's give each raw_object_store a unified list of object_directory structs. The first will be the main store, and everything after is an alternate. Very few callers even care about the distinction, and can just loop over the whole list (and those who care can just treat the first element differently). A few observations: - we don't need r->objects->objectdir anymore, and can just mechanically convert that to r->objects->odb->path - object_directory's path field needs to become a real pointer rather than a FLEX_ARRAY, in order to fill it with expand_base_dir() - we'll call prepare_alt_odb() earlier in many functions (i.e., outside of the loop). This may result in us calling it even when our function would be satisfied looking only at the main odb. But this doesn't matter in practice. It's not a very expensive operation in the first place, and in the majority of cases it will be a noop. We call it already (and cache its results) in prepare_packed_git(), and we'll generally check packs before loose objects. So essentially every program is going to call it immediately once per program. Arguably we should just prepare_alt_odb() immediately upon setting up the repository's object directory, which would save us sprinkling calls throughout the code base (and forgetting to do so has been a source of subtle bugs in the past). But I've stopped short of that here, since there are already a lot of other moving parts in this patch. - Most call sites just get shorter. The check_and_freshen() functions are an exception, because they have entry points to handle local and nonlocal directories separately. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-11-12 14:50:39 +00:00
replace_dir(buf, git_dir_len + 7, repo->objects->odb->path);
else if (git_hooks_path && dir_prefix(base, "hooks"))
replace_dir(buf, git_dir_len + 5, git_hooks_path);
else if (repo->different_commondir)
update_common_dir(buf, git_dir_len, repo->commondir);
2014-11-30 08:24:31 +00:00
}
static void strbuf_worktree_gitdir(struct strbuf *buf,
const struct repository *repo,
const struct worktree *wt)
{
if (!wt)
strbuf_addstr(buf, repo->gitdir);
else if (!wt->id)
strbuf_addstr(buf, repo->commondir);
else
strbuf_git_common_path(buf, repo, "worktrees/%s", wt->id);
}
static void do_git_path(const struct repository *repo,
const struct worktree *wt, struct strbuf *buf,
const char *fmt, va_list args)
{
2014-11-30 08:24:31 +00:00
int gitdir_len;
strbuf_worktree_gitdir(buf, repo, wt);
if (buf->len && !is_dir_sep(buf->buf[buf->len - 1]))
strbuf_addch(buf, '/');
2014-11-30 08:24:31 +00:00
gitdir_len = buf->len;
strbuf_vaddf(buf, fmt, args);
path: worktree_git_path() should not use file relocation git_path is a convenience function that usually produces a string $GIT_DIR/<path>. Since v2.5.0-rc0~143^2~35 (git_path(): be aware of file relocation in $GIT_DIR, 2014-11-30), as a side benefit callers get support for path relocation variables like $GIT_OBJECT_DIRECTORY: - git_path("index") is $GIT_INDEX_FILE when set - git_path("info/grafts") is $GIT_GRAFTS_FILE when set - git_path("objects/<foo>") is $GIT_OBJECT_DIRECTORY/<foo> when set - git_path("hooks/<foo>") is <foo> under core.hookspath when set - git_path("refs/<foo>") etc (see path.c::common_list) is relative to $GIT_COMMON_DIR instead of $GIT_DIR worktree_git_path, by comparison, is designed to resolve files in a specific worktree's git dir. Unfortunately, it shares code with git_path and performs the same relocation. The result is that paths that are meant to be relative to the specified worktree's git dir end up replaced by paths from environment variables within the current git dir. Luckily, no current callers pass such arguments. The relocation was noticed when testing the result of merging two patches under review, one of which introduces a caller: * The first patch made git prune check the index file in each worktree's git dir (using worktree_git_path(wt, "index")) for objects not to prune. This would trigger the unwanted relocation when GIT_INDEX_FILE is set, causing objects reachable from the index to be pruned. * The second patch simplified the relocation logic for index, info/grafts, objects, and hooks to happen unconditionally instead of based on whether environment or configuration variables are set. This caused the relocation to trigger even when GIT_INDEX_FILE is not set. [jn: rewrote commit message; skipping all relocation instead of just GIT_INDEX_FILE] Signed-off-by: Brandon Williams <bmwill@google.com> Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-06-22 18:43:39 +00:00
if (!wt)
adjust_git_path(repo, buf, gitdir_len);
strbuf_cleanup_path(buf);
}
char *repo_git_path(const struct repository *repo,
const char *fmt, ...)
{
struct strbuf path = STRBUF_INIT;
va_list args;
va_start(args, fmt);
do_git_path(repo, NULL, &path, fmt, args);
va_end(args);
return strbuf_detach(&path, NULL);
}
void strbuf_repo_git_path(struct strbuf *sb,
const struct repository *repo,
const char *fmt, ...)
{
va_list args;
va_start(args, fmt);
do_git_path(repo, NULL, sb, fmt, args);
va_end(args);
}
char *git_path_buf(struct strbuf *buf, const char *fmt, ...)
{
va_list args;
strbuf_reset(buf);
va_start(args, fmt);
do_git_path(the_repository, NULL, buf, fmt, args);
va_end(args);
return buf->buf;
}
void strbuf_git_path(struct strbuf *sb, const char *fmt, ...)
{
va_list args;
va_start(args, fmt);
do_git_path(the_repository, NULL, sb, fmt, args);
va_end(args);
}
const char *git_path(const char *fmt, ...)
{
struct strbuf *pathname = get_pathname();
va_list args;
va_start(args, fmt);
do_git_path(the_repository, NULL, pathname, fmt, args);
va_end(args);
return pathname->buf;
}
char *git_pathdup(const char *fmt, ...)
{
struct strbuf path = STRBUF_INIT;
va_list args;
va_start(args, fmt);
do_git_path(the_repository, NULL, &path, fmt, args);
va_end(args);
return strbuf_detach(&path, NULL);
}
char *mkpathdup(const char *fmt, ...)
{
struct strbuf sb = STRBUF_INIT;
va_list args;
va_start(args, fmt);
strbuf_vaddf(&sb, fmt, args);
va_end(args);
strbuf_cleanup_path(&sb);
return strbuf_detach(&sb, NULL);
}
const char *mkpath(const char *fmt, ...)
{
va_list args;
struct strbuf *pathname = get_pathname();
va_start(args, fmt);
strbuf_vaddf(pathname, fmt, args);
va_end(args);
return cleanup_path(pathname->buf);
}
const char *worktree_git_path(const struct worktree *wt, const char *fmt, ...)
{
struct strbuf *pathname = get_pathname();
va_list args;
va_start(args, fmt);
do_git_path(the_repository, wt, pathname, fmt, args);
va_end(args);
return pathname->buf;
}
static void do_worktree_path(const struct repository *repo,
struct strbuf *buf,
const char *fmt, va_list args)
{
strbuf_addstr(buf, repo->worktree);
if(buf->len && !is_dir_sep(buf->buf[buf->len - 1]))
strbuf_addch(buf, '/');
strbuf_vaddf(buf, fmt, args);
strbuf_cleanup_path(buf);
}
char *repo_worktree_path(const struct repository *repo, const char *fmt, ...)
{
struct strbuf path = STRBUF_INIT;
va_list args;
if (!repo->worktree)
return NULL;
va_start(args, fmt);
do_worktree_path(repo, &path, fmt, args);
va_end(args);
return strbuf_detach(&path, NULL);
}
void strbuf_repo_worktree_path(struct strbuf *sb,
const struct repository *repo,
const char *fmt, ...)
{
va_list args;
if (!repo->worktree)
return;
va_start(args, fmt);
do_worktree_path(repo, sb, fmt, args);
va_end(args);
}
allow do_submodule_path to work even if submodule isn't checked out Currently, do_submodule_path will attempt locating the .git directory by using read_gitfile on <path>/.git. If this fails it just assumes the <path>/.git is actually a git directory. This is good because it allows for handling submodules which were cloned in a regular manner first before being added to the superproject. Unfortunately this fails if the <path> is not actually checked out any longer, such as by removing the directory. Fix this by checking if the directory we found is actually a gitdir. In the case it is not, attempt to lookup the submodule configuration and find the name of where it is stored in the .git/modules/ directory of the superproject. If we can't locate the submodule configuration, this might occur because for example a submodule gitlink was added but the corresponding .gitmodules file was not properly updated. A die() here would not be pleasant to the users of submodule diff formats, so instead, modify do_submodule_path() to return an error code: - git_pathdup_submodule() returns NULL when we fail to find a path. - strbuf_git_path_submodule() propagates the error code to the caller. Modify the callers of these functions to check the error code and fail properly. This ensures we don't attempt to use a bad path that doesn't match the corresponding submodule. Because this change fixes add_submodule_odb() to work even if the submodule is not checked out, update the wording of the submodule log diff format to correctly display that the submodule is "not initialized" instead of "not checked out" Add tests to ensure this change works as expected. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2016-08-31 23:27:22 +00:00
/* Returns 0 on success, negative on failure. */
static int do_submodule_path(struct strbuf *buf, const char *path,
const char *fmt, va_list args)
{
struct strbuf git_submodule_common_dir = STRBUF_INIT;
struct strbuf git_submodule_dir = STRBUF_INIT;
int ret;
ret = submodule_to_gitdir(&git_submodule_dir, path);
if (ret)
goto cleanup;
strbuf_complete(&git_submodule_dir, '/');
strbuf_addbuf(buf, &git_submodule_dir);
strbuf_vaddf(buf, fmt, args);
if (get_common_dir_noenv(&git_submodule_common_dir, git_submodule_dir.buf))
update_common_dir(buf, git_submodule_dir.len, git_submodule_common_dir.buf);
strbuf_cleanup_path(buf);
allow do_submodule_path to work even if submodule isn't checked out Currently, do_submodule_path will attempt locating the .git directory by using read_gitfile on <path>/.git. If this fails it just assumes the <path>/.git is actually a git directory. This is good because it allows for handling submodules which were cloned in a regular manner first before being added to the superproject. Unfortunately this fails if the <path> is not actually checked out any longer, such as by removing the directory. Fix this by checking if the directory we found is actually a gitdir. In the case it is not, attempt to lookup the submodule configuration and find the name of where it is stored in the .git/modules/ directory of the superproject. If we can't locate the submodule configuration, this might occur because for example a submodule gitlink was added but the corresponding .gitmodules file was not properly updated. A die() here would not be pleasant to the users of submodule diff formats, so instead, modify do_submodule_path() to return an error code: - git_pathdup_submodule() returns NULL when we fail to find a path. - strbuf_git_path_submodule() propagates the error code to the caller. Modify the callers of these functions to check the error code and fail properly. This ensures we don't attempt to use a bad path that doesn't match the corresponding submodule. Because this change fixes add_submodule_odb() to work even if the submodule is not checked out, update the wording of the submodule log diff format to correctly display that the submodule is "not initialized" instead of "not checked out" Add tests to ensure this change works as expected. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2016-08-31 23:27:22 +00:00
cleanup:
strbuf_release(&git_submodule_dir);
strbuf_release(&git_submodule_common_dir);
return ret;
}
char *git_pathdup_submodule(const char *path, const char *fmt, ...)
{
allow do_submodule_path to work even if submodule isn't checked out Currently, do_submodule_path will attempt locating the .git directory by using read_gitfile on <path>/.git. If this fails it just assumes the <path>/.git is actually a git directory. This is good because it allows for handling submodules which were cloned in a regular manner first before being added to the superproject. Unfortunately this fails if the <path> is not actually checked out any longer, such as by removing the directory. Fix this by checking if the directory we found is actually a gitdir. In the case it is not, attempt to lookup the submodule configuration and find the name of where it is stored in the .git/modules/ directory of the superproject. If we can't locate the submodule configuration, this might occur because for example a submodule gitlink was added but the corresponding .gitmodules file was not properly updated. A die() here would not be pleasant to the users of submodule diff formats, so instead, modify do_submodule_path() to return an error code: - git_pathdup_submodule() returns NULL when we fail to find a path. - strbuf_git_path_submodule() propagates the error code to the caller. Modify the callers of these functions to check the error code and fail properly. This ensures we don't attempt to use a bad path that doesn't match the corresponding submodule. Because this change fixes add_submodule_odb() to work even if the submodule is not checked out, update the wording of the submodule log diff format to correctly display that the submodule is "not initialized" instead of "not checked out" Add tests to ensure this change works as expected. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2016-08-31 23:27:22 +00:00
int err;
va_list args;
struct strbuf buf = STRBUF_INIT;
va_start(args, fmt);
allow do_submodule_path to work even if submodule isn't checked out Currently, do_submodule_path will attempt locating the .git directory by using read_gitfile on <path>/.git. If this fails it just assumes the <path>/.git is actually a git directory. This is good because it allows for handling submodules which were cloned in a regular manner first before being added to the superproject. Unfortunately this fails if the <path> is not actually checked out any longer, such as by removing the directory. Fix this by checking if the directory we found is actually a gitdir. In the case it is not, attempt to lookup the submodule configuration and find the name of where it is stored in the .git/modules/ directory of the superproject. If we can't locate the submodule configuration, this might occur because for example a submodule gitlink was added but the corresponding .gitmodules file was not properly updated. A die() here would not be pleasant to the users of submodule diff formats, so instead, modify do_submodule_path() to return an error code: - git_pathdup_submodule() returns NULL when we fail to find a path. - strbuf_git_path_submodule() propagates the error code to the caller. Modify the callers of these functions to check the error code and fail properly. This ensures we don't attempt to use a bad path that doesn't match the corresponding submodule. Because this change fixes add_submodule_odb() to work even if the submodule is not checked out, update the wording of the submodule log diff format to correctly display that the submodule is "not initialized" instead of "not checked out" Add tests to ensure this change works as expected. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2016-08-31 23:27:22 +00:00
err = do_submodule_path(&buf, path, fmt, args);
va_end(args);
allow do_submodule_path to work even if submodule isn't checked out Currently, do_submodule_path will attempt locating the .git directory by using read_gitfile on <path>/.git. If this fails it just assumes the <path>/.git is actually a git directory. This is good because it allows for handling submodules which were cloned in a regular manner first before being added to the superproject. Unfortunately this fails if the <path> is not actually checked out any longer, such as by removing the directory. Fix this by checking if the directory we found is actually a gitdir. In the case it is not, attempt to lookup the submodule configuration and find the name of where it is stored in the .git/modules/ directory of the superproject. If we can't locate the submodule configuration, this might occur because for example a submodule gitlink was added but the corresponding .gitmodules file was not properly updated. A die() here would not be pleasant to the users of submodule diff formats, so instead, modify do_submodule_path() to return an error code: - git_pathdup_submodule() returns NULL when we fail to find a path. - strbuf_git_path_submodule() propagates the error code to the caller. Modify the callers of these functions to check the error code and fail properly. This ensures we don't attempt to use a bad path that doesn't match the corresponding submodule. Because this change fixes add_submodule_odb() to work even if the submodule is not checked out, update the wording of the submodule log diff format to correctly display that the submodule is "not initialized" instead of "not checked out" Add tests to ensure this change works as expected. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2016-08-31 23:27:22 +00:00
if (err) {
strbuf_release(&buf);
return NULL;
}
return strbuf_detach(&buf, NULL);
}
allow do_submodule_path to work even if submodule isn't checked out Currently, do_submodule_path will attempt locating the .git directory by using read_gitfile on <path>/.git. If this fails it just assumes the <path>/.git is actually a git directory. This is good because it allows for handling submodules which were cloned in a regular manner first before being added to the superproject. Unfortunately this fails if the <path> is not actually checked out any longer, such as by removing the directory. Fix this by checking if the directory we found is actually a gitdir. In the case it is not, attempt to lookup the submodule configuration and find the name of where it is stored in the .git/modules/ directory of the superproject. If we can't locate the submodule configuration, this might occur because for example a submodule gitlink was added but the corresponding .gitmodules file was not properly updated. A die() here would not be pleasant to the users of submodule diff formats, so instead, modify do_submodule_path() to return an error code: - git_pathdup_submodule() returns NULL when we fail to find a path. - strbuf_git_path_submodule() propagates the error code to the caller. Modify the callers of these functions to check the error code and fail properly. This ensures we don't attempt to use a bad path that doesn't match the corresponding submodule. Because this change fixes add_submodule_odb() to work even if the submodule is not checked out, update the wording of the submodule log diff format to correctly display that the submodule is "not initialized" instead of "not checked out" Add tests to ensure this change works as expected. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2016-08-31 23:27:22 +00:00
int strbuf_git_path_submodule(struct strbuf *buf, const char *path,
const char *fmt, ...)
{
allow do_submodule_path to work even if submodule isn't checked out Currently, do_submodule_path will attempt locating the .git directory by using read_gitfile on <path>/.git. If this fails it just assumes the <path>/.git is actually a git directory. This is good because it allows for handling submodules which were cloned in a regular manner first before being added to the superproject. Unfortunately this fails if the <path> is not actually checked out any longer, such as by removing the directory. Fix this by checking if the directory we found is actually a gitdir. In the case it is not, attempt to lookup the submodule configuration and find the name of where it is stored in the .git/modules/ directory of the superproject. If we can't locate the submodule configuration, this might occur because for example a submodule gitlink was added but the corresponding .gitmodules file was not properly updated. A die() here would not be pleasant to the users of submodule diff formats, so instead, modify do_submodule_path() to return an error code: - git_pathdup_submodule() returns NULL when we fail to find a path. - strbuf_git_path_submodule() propagates the error code to the caller. Modify the callers of these functions to check the error code and fail properly. This ensures we don't attempt to use a bad path that doesn't match the corresponding submodule. Because this change fixes add_submodule_odb() to work even if the submodule is not checked out, update the wording of the submodule log diff format to correctly display that the submodule is "not initialized" instead of "not checked out" Add tests to ensure this change works as expected. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2016-08-31 23:27:22 +00:00
int err;
va_list args;
va_start(args, fmt);
allow do_submodule_path to work even if submodule isn't checked out Currently, do_submodule_path will attempt locating the .git directory by using read_gitfile on <path>/.git. If this fails it just assumes the <path>/.git is actually a git directory. This is good because it allows for handling submodules which were cloned in a regular manner first before being added to the superproject. Unfortunately this fails if the <path> is not actually checked out any longer, such as by removing the directory. Fix this by checking if the directory we found is actually a gitdir. In the case it is not, attempt to lookup the submodule configuration and find the name of where it is stored in the .git/modules/ directory of the superproject. If we can't locate the submodule configuration, this might occur because for example a submodule gitlink was added but the corresponding .gitmodules file was not properly updated. A die() here would not be pleasant to the users of submodule diff formats, so instead, modify do_submodule_path() to return an error code: - git_pathdup_submodule() returns NULL when we fail to find a path. - strbuf_git_path_submodule() propagates the error code to the caller. Modify the callers of these functions to check the error code and fail properly. This ensures we don't attempt to use a bad path that doesn't match the corresponding submodule. Because this change fixes add_submodule_odb() to work even if the submodule is not checked out, update the wording of the submodule log diff format to correctly display that the submodule is "not initialized" instead of "not checked out" Add tests to ensure this change works as expected. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2016-08-31 23:27:22 +00:00
err = do_submodule_path(buf, path, fmt, args);
va_end(args);
allow do_submodule_path to work even if submodule isn't checked out Currently, do_submodule_path will attempt locating the .git directory by using read_gitfile on <path>/.git. If this fails it just assumes the <path>/.git is actually a git directory. This is good because it allows for handling submodules which were cloned in a regular manner first before being added to the superproject. Unfortunately this fails if the <path> is not actually checked out any longer, such as by removing the directory. Fix this by checking if the directory we found is actually a gitdir. In the case it is not, attempt to lookup the submodule configuration and find the name of where it is stored in the .git/modules/ directory of the superproject. If we can't locate the submodule configuration, this might occur because for example a submodule gitlink was added but the corresponding .gitmodules file was not properly updated. A die() here would not be pleasant to the users of submodule diff formats, so instead, modify do_submodule_path() to return an error code: - git_pathdup_submodule() returns NULL when we fail to find a path. - strbuf_git_path_submodule() propagates the error code to the caller. Modify the callers of these functions to check the error code and fail properly. This ensures we don't attempt to use a bad path that doesn't match the corresponding submodule. Because this change fixes add_submodule_odb() to work even if the submodule is not checked out, update the wording of the submodule log diff format to correctly display that the submodule is "not initialized" instead of "not checked out" Add tests to ensure this change works as expected. Signed-off-by: Jacob Keller <jacob.keller@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2016-08-31 23:27:22 +00:00
return err;
}
static void do_git_common_path(const struct repository *repo,
struct strbuf *buf,
const char *fmt,
va_list args)
{
strbuf_addstr(buf, repo->commondir);
if (buf->len && !is_dir_sep(buf->buf[buf->len - 1]))
strbuf_addch(buf, '/');
strbuf_vaddf(buf, fmt, args);
strbuf_cleanup_path(buf);
}
const char *git_common_path(const char *fmt, ...)
{
struct strbuf *pathname = get_pathname();
va_list args;
va_start(args, fmt);
do_git_common_path(the_repository, pathname, fmt, args);
va_end(args);
return pathname->buf;
}
void strbuf_git_common_path(struct strbuf *sb,
const struct repository *repo,
const char *fmt, ...)
{
va_list args;
va_start(args, fmt);
do_git_common_path(repo, sb, fmt, args);
va_end(args);
}
int validate_headref(const char *path)
{
struct stat st;
char buffer[256];
const char *refname;
struct object_id oid;
int fd;
ssize_t len;
if (lstat(path, &st) < 0)
return -1;
/* Make sure it is a "refs/.." symlink */
if (S_ISLNK(st.st_mode)) {
len = readlink(path, buffer, sizeof(buffer)-1);
if (len >= 5 && !memcmp("refs/", buffer, 5))
return 0;
return -1;
}
/*
* Anything else, just open it and try to see if it is a symbolic ref.
*/
fd = open(path, O_RDONLY);
if (fd < 0)
return -1;
len = read_in_full(fd, buffer, sizeof(buffer)-1);
close(fd);
if (len < 0)
return -1;
buffer[len] = '\0';
/*
* Is it a symbolic ref?
*/
if (skip_prefix(buffer, "ref:", &refname)) {
while (isspace(*refname))
refname++;
if (starts_with(refname, "refs/"))
return 0;
}
/*
* Is this a detached HEAD?
*/
if (!get_oid_hex(buffer, &oid))
return 0;
return -1;
}
static struct passwd *getpw_str(const char *username, size_t len)
{
[PATCH] daemon.c and path.enter_repo(): revamp path validation. The whitelist of git-daemon is checked against return value from enter_repo(), and enter_repo() used to return the value obtained from getcwd() to avoid directory aliasing issues as discussed earier (mid October 2005). Unfortunately, it did not go well as we hoped. For example, /pub on a kernel.org public machine is a symlink to its real mountpoint, and it is understandable that the administrator does not want to adjust the whitelist every time /pub needs to point at a different partition for storage allcation or whatever reasons. Being able to keep using /pub/scm as the whitelist is a desirable property. So this version of enter_repo() reports what it used to chdir() and validate, but does not use getcwd() to canonicalize the directory name. When it sees a user relative path ~user/path, it internally resolves it to try chdir() there, but it still reports ~user/path (possibly after appending .git if allowed to do so, in which case it would report ~user/path.git). What this means is that if a whitelist wants to allow a user relative path, it needs to say "~" (for all users) or list user home directories like "~alice" "~bob". And no, you cannot say /home if the advertised way to access user home directories are ~alice,~bob, etc. The whole point of this is to avoid unnecessary aliasing issues. Anyway, because of this, daemon needs to do a bit more work to guard itself. Namely, it needs to make sure that the accessor does not try to exploit its leading path match rule by inserting /../ in the middle or hanging /.. at the end. I resurrected the belts and suspender paranoia code HPA did for this purpose. This check cannot be done in the enter_repo() unconditionally, because there are valid callers of enter_repo() that want to honor /../; authorized users coming over ssh to run send-pack and fetch-pack should be allowed to do so. Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-12-03 09:45:57 +00:00
struct passwd *pw;
char *username_z = xmemdupz(username, len);
pw = getpwnam(username_z);
free(username_z);
return pw;
}
/*
* Return a string with ~ and ~user expanded via getpw*. If buf != NULL,
* then it is a newly allocated string. Returns NULL on getpw failure or
* if path is NULL.
*
* If real_home is true, real_path($HOME) is used in the expansion.
*/
char *expand_user_path(const char *path, int real_home)
{
struct strbuf user_path = STRBUF_INIT;
const char *to_copy = path;
if (path == NULL)
goto return_null;
if (path[0] == '~') {
const char *first_slash = strchrnul(path, '/');
const char *username = path + 1;
size_t username_len = first_slash - username;
if (username_len == 0) {
const char *home = getenv("HOME");
if (!home)
goto return_null;
if (real_home)
strbuf_add_real_path(&user_path, home);
else
strbuf_addstr(&user_path, home);
#ifdef GIT_WINDOWS_NATIVE
convert_slashes(user_path.buf);
#endif
} else {
struct passwd *pw = getpw_str(username, username_len);
if (!pw)
goto return_null;
strbuf_addstr(&user_path, pw->pw_dir);
}
to_copy = first_slash;
[PATCH] daemon.c and path.enter_repo(): revamp path validation. The whitelist of git-daemon is checked against return value from enter_repo(), and enter_repo() used to return the value obtained from getcwd() to avoid directory aliasing issues as discussed earier (mid October 2005). Unfortunately, it did not go well as we hoped. For example, /pub on a kernel.org public machine is a symlink to its real mountpoint, and it is understandable that the administrator does not want to adjust the whitelist every time /pub needs to point at a different partition for storage allcation or whatever reasons. Being able to keep using /pub/scm as the whitelist is a desirable property. So this version of enter_repo() reports what it used to chdir() and validate, but does not use getcwd() to canonicalize the directory name. When it sees a user relative path ~user/path, it internally resolves it to try chdir() there, but it still reports ~user/path (possibly after appending .git if allowed to do so, in which case it would report ~user/path.git). What this means is that if a whitelist wants to allow a user relative path, it needs to say "~" (for all users) or list user home directories like "~alice" "~bob". And no, you cannot say /home if the advertised way to access user home directories are ~alice,~bob, etc. The whole point of this is to avoid unnecessary aliasing issues. Anyway, because of this, daemon needs to do a bit more work to guard itself. Namely, it needs to make sure that the accessor does not try to exploit its leading path match rule by inserting /../ in the middle or hanging /.. at the end. I resurrected the belts and suspender paranoia code HPA did for this purpose. This check cannot be done in the enter_repo() unconditionally, because there are valid callers of enter_repo() that want to honor /../; authorized users coming over ssh to run send-pack and fetch-pack should be allowed to do so. Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-12-03 09:45:57 +00:00
}
strbuf_addstr(&user_path, to_copy);
return strbuf_detach(&user_path, NULL);
return_null:
strbuf_release(&user_path);
return NULL;
}
[PATCH] daemon.c and path.enter_repo(): revamp path validation. The whitelist of git-daemon is checked against return value from enter_repo(), and enter_repo() used to return the value obtained from getcwd() to avoid directory aliasing issues as discussed earier (mid October 2005). Unfortunately, it did not go well as we hoped. For example, /pub on a kernel.org public machine is a symlink to its real mountpoint, and it is understandable that the administrator does not want to adjust the whitelist every time /pub needs to point at a different partition for storage allcation or whatever reasons. Being able to keep using /pub/scm as the whitelist is a desirable property. So this version of enter_repo() reports what it used to chdir() and validate, but does not use getcwd() to canonicalize the directory name. When it sees a user relative path ~user/path, it internally resolves it to try chdir() there, but it still reports ~user/path (possibly after appending .git if allowed to do so, in which case it would report ~user/path.git). What this means is that if a whitelist wants to allow a user relative path, it needs to say "~" (for all users) or list user home directories like "~alice" "~bob". And no, you cannot say /home if the advertised way to access user home directories are ~alice,~bob, etc. The whole point of this is to avoid unnecessary aliasing issues. Anyway, because of this, daemon needs to do a bit more work to guard itself. Namely, it needs to make sure that the accessor does not try to exploit its leading path match rule by inserting /../ in the middle or hanging /.. at the end. I resurrected the belts and suspender paranoia code HPA did for this purpose. This check cannot be done in the enter_repo() unconditionally, because there are valid callers of enter_repo() that want to honor /../; authorized users coming over ssh to run send-pack and fetch-pack should be allowed to do so. Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-12-03 09:45:57 +00:00
/*
* First, one directory to try is determined by the following algorithm.
*
* (0) If "strict" is given, the path is used as given and no DWIM is
* done. Otherwise:
* (1) "~/path" to mean path under the running user's home directory;
* (2) "~user/path" to mean path under named user's home directory;
* (3) "relative/path" to mean cwd relative directory; or
* (4) "/absolute/path" to mean absolute directory.
*
* Unless "strict" is given, we check "%s/.git", "%s", "%s.git/.git", "%s.git"
* in this order. We select the first one that is a valid git repository, and
* chdir() to it. If none match, or we fail to chdir, we return NULL.
[PATCH] daemon.c and path.enter_repo(): revamp path validation. The whitelist of git-daemon is checked against return value from enter_repo(), and enter_repo() used to return the value obtained from getcwd() to avoid directory aliasing issues as discussed earier (mid October 2005). Unfortunately, it did not go well as we hoped. For example, /pub on a kernel.org public machine is a symlink to its real mountpoint, and it is understandable that the administrator does not want to adjust the whitelist every time /pub needs to point at a different partition for storage allcation or whatever reasons. Being able to keep using /pub/scm as the whitelist is a desirable property. So this version of enter_repo() reports what it used to chdir() and validate, but does not use getcwd() to canonicalize the directory name. When it sees a user relative path ~user/path, it internally resolves it to try chdir() there, but it still reports ~user/path (possibly after appending .git if allowed to do so, in which case it would report ~user/path.git). What this means is that if a whitelist wants to allow a user relative path, it needs to say "~" (for all users) or list user home directories like "~alice" "~bob". And no, you cannot say /home if the advertised way to access user home directories are ~alice,~bob, etc. The whole point of this is to avoid unnecessary aliasing issues. Anyway, because of this, daemon needs to do a bit more work to guard itself. Namely, it needs to make sure that the accessor does not try to exploit its leading path match rule by inserting /../ in the middle or hanging /.. at the end. I resurrected the belts and suspender paranoia code HPA did for this purpose. This check cannot be done in the enter_repo() unconditionally, because there are valid callers of enter_repo() that want to honor /../; authorized users coming over ssh to run send-pack and fetch-pack should be allowed to do so. Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-12-03 09:45:57 +00:00
*
* If all goes well, we return the directory we used to chdir() (but
* before ~user is expanded), avoiding getcwd() resolving symbolic
* links. User relative paths are also returned as they are given,
* except DWIM suffixing.
*/
const char *enter_repo(const char *path, int strict)
{
static struct strbuf validated_path = STRBUF_INIT;
static struct strbuf used_path = STRBUF_INIT;
[PATCH] daemon.c and path.enter_repo(): revamp path validation. The whitelist of git-daemon is checked against return value from enter_repo(), and enter_repo() used to return the value obtained from getcwd() to avoid directory aliasing issues as discussed earier (mid October 2005). Unfortunately, it did not go well as we hoped. For example, /pub on a kernel.org public machine is a symlink to its real mountpoint, and it is understandable that the administrator does not want to adjust the whitelist every time /pub needs to point at a different partition for storage allcation or whatever reasons. Being able to keep using /pub/scm as the whitelist is a desirable property. So this version of enter_repo() reports what it used to chdir() and validate, but does not use getcwd() to canonicalize the directory name. When it sees a user relative path ~user/path, it internally resolves it to try chdir() there, but it still reports ~user/path (possibly after appending .git if allowed to do so, in which case it would report ~user/path.git). What this means is that if a whitelist wants to allow a user relative path, it needs to say "~" (for all users) or list user home directories like "~alice" "~bob". And no, you cannot say /home if the advertised way to access user home directories are ~alice,~bob, etc. The whole point of this is to avoid unnecessary aliasing issues. Anyway, because of this, daemon needs to do a bit more work to guard itself. Namely, it needs to make sure that the accessor does not try to exploit its leading path match rule by inserting /../ in the middle or hanging /.. at the end. I resurrected the belts and suspender paranoia code HPA did for this purpose. This check cannot be done in the enter_repo() unconditionally, because there are valid callers of enter_repo() that want to honor /../; authorized users coming over ssh to run send-pack and fetch-pack should be allowed to do so. Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-12-03 09:45:57 +00:00
if (!path)
return NULL;
[PATCH] daemon.c and path.enter_repo(): revamp path validation. The whitelist of git-daemon is checked against return value from enter_repo(), and enter_repo() used to return the value obtained from getcwd() to avoid directory aliasing issues as discussed earier (mid October 2005). Unfortunately, it did not go well as we hoped. For example, /pub on a kernel.org public machine is a symlink to its real mountpoint, and it is understandable that the administrator does not want to adjust the whitelist every time /pub needs to point at a different partition for storage allcation or whatever reasons. Being able to keep using /pub/scm as the whitelist is a desirable property. So this version of enter_repo() reports what it used to chdir() and validate, but does not use getcwd() to canonicalize the directory name. When it sees a user relative path ~user/path, it internally resolves it to try chdir() there, but it still reports ~user/path (possibly after appending .git if allowed to do so, in which case it would report ~user/path.git). What this means is that if a whitelist wants to allow a user relative path, it needs to say "~" (for all users) or list user home directories like "~alice" "~bob". And no, you cannot say /home if the advertised way to access user home directories are ~alice,~bob, etc. The whole point of this is to avoid unnecessary aliasing issues. Anyway, because of this, daemon needs to do a bit more work to guard itself. Namely, it needs to make sure that the accessor does not try to exploit its leading path match rule by inserting /../ in the middle or hanging /.. at the end. I resurrected the belts and suspender paranoia code HPA did for this purpose. This check cannot be done in the enter_repo() unconditionally, because there are valid callers of enter_repo() that want to honor /../; authorized users coming over ssh to run send-pack and fetch-pack should be allowed to do so. Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-12-03 09:45:57 +00:00
if (!strict) {
static const char *suffix[] = {
standardize and improve lookup rules for external local repos When you specify a local repository on the command line of clone, ls-remote, upload-pack, receive-pack, or upload-archive, or in a request to git-daemon, we perform a little bit of lookup magic, doing things like looking in working trees for .git directories and appending ".git" for bare repos. For clone, this magic happens in get_repo_path. For everything else, it happens in enter_repo. In both cases, there are some ambiguous or confusing cases that aren't handled well, and there is one case that is not handled the same by both methods. This patch tries to provide (and test!) standard, sensible lookup rules for both code paths. The intended changes are: 1. When looking up "foo", we have always preferred a working tree "foo" (containing "foo/.git" over the bare "foo.git". But we did not prefer a bare "foo" over "foo.git". With this patch, we do so. 2. We would select directories that existed but didn't actually look like git repositories. With this patch, we make sure a selected directory looks like a git repo. Not only is this more sensible in general, but it will help anybody who is negatively affected by change (1) negatively (e.g., if they had "foo.git" next to its separate work tree "foo", and expect to keep finding "foo.git" when they reference "foo"). 3. The enter_repo code path would, given "foo", look for "foo.git/.git" (i.e., do the ".git" append magic even for a repo with working tree). The clone code path did not; with this patch, they now behave the same. In the unlikely case of a working tree overlaying a bare repo (i.e., a ".git" directory _inside_ a bare repo), we continue to treat it as a working tree (prefering the "inner" .git over the bare repo). This is mainly because the combination seems nonsensical, and I'd rather stick with existing behavior on the off chance that somebody is relying on it. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-02-02 21:59:13 +00:00
"/.git", "", ".git/.git", ".git", NULL,
[PATCH] daemon.c and path.enter_repo(): revamp path validation. The whitelist of git-daemon is checked against return value from enter_repo(), and enter_repo() used to return the value obtained from getcwd() to avoid directory aliasing issues as discussed earier (mid October 2005). Unfortunately, it did not go well as we hoped. For example, /pub on a kernel.org public machine is a symlink to its real mountpoint, and it is understandable that the administrator does not want to adjust the whitelist every time /pub needs to point at a different partition for storage allcation or whatever reasons. Being able to keep using /pub/scm as the whitelist is a desirable property. So this version of enter_repo() reports what it used to chdir() and validate, but does not use getcwd() to canonicalize the directory name. When it sees a user relative path ~user/path, it internally resolves it to try chdir() there, but it still reports ~user/path (possibly after appending .git if allowed to do so, in which case it would report ~user/path.git). What this means is that if a whitelist wants to allow a user relative path, it needs to say "~" (for all users) or list user home directories like "~alice" "~bob". And no, you cannot say /home if the advertised way to access user home directories are ~alice,~bob, etc. The whole point of this is to avoid unnecessary aliasing issues. Anyway, because of this, daemon needs to do a bit more work to guard itself. Namely, it needs to make sure that the accessor does not try to exploit its leading path match rule by inserting /../ in the middle or hanging /.. at the end. I resurrected the belts and suspender paranoia code HPA did for this purpose. This check cannot be done in the enter_repo() unconditionally, because there are valid callers of enter_repo() that want to honor /../; authorized users coming over ssh to run send-pack and fetch-pack should be allowed to do so. Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-12-03 09:45:57 +00:00
};
const char *gitfile;
[PATCH] daemon.c and path.enter_repo(): revamp path validation. The whitelist of git-daemon is checked against return value from enter_repo(), and enter_repo() used to return the value obtained from getcwd() to avoid directory aliasing issues as discussed earier (mid October 2005). Unfortunately, it did not go well as we hoped. For example, /pub on a kernel.org public machine is a symlink to its real mountpoint, and it is understandable that the administrator does not want to adjust the whitelist every time /pub needs to point at a different partition for storage allcation or whatever reasons. Being able to keep using /pub/scm as the whitelist is a desirable property. So this version of enter_repo() reports what it used to chdir() and validate, but does not use getcwd() to canonicalize the directory name. When it sees a user relative path ~user/path, it internally resolves it to try chdir() there, but it still reports ~user/path (possibly after appending .git if allowed to do so, in which case it would report ~user/path.git). What this means is that if a whitelist wants to allow a user relative path, it needs to say "~" (for all users) or list user home directories like "~alice" "~bob". And no, you cannot say /home if the advertised way to access user home directories are ~alice,~bob, etc. The whole point of this is to avoid unnecessary aliasing issues. Anyway, because of this, daemon needs to do a bit more work to guard itself. Namely, it needs to make sure that the accessor does not try to exploit its leading path match rule by inserting /../ in the middle or hanging /.. at the end. I resurrected the belts and suspender paranoia code HPA did for this purpose. This check cannot be done in the enter_repo() unconditionally, because there are valid callers of enter_repo() that want to honor /../; authorized users coming over ssh to run send-pack and fetch-pack should be allowed to do so. Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-12-03 09:45:57 +00:00
int len = strlen(path);
int i;
while ((1 < len) && (path[len-1] == '/'))
[PATCH] daemon.c and path.enter_repo(): revamp path validation. The whitelist of git-daemon is checked against return value from enter_repo(), and enter_repo() used to return the value obtained from getcwd() to avoid directory aliasing issues as discussed earier (mid October 2005). Unfortunately, it did not go well as we hoped. For example, /pub on a kernel.org public machine is a symlink to its real mountpoint, and it is understandable that the administrator does not want to adjust the whitelist every time /pub needs to point at a different partition for storage allcation or whatever reasons. Being able to keep using /pub/scm as the whitelist is a desirable property. So this version of enter_repo() reports what it used to chdir() and validate, but does not use getcwd() to canonicalize the directory name. When it sees a user relative path ~user/path, it internally resolves it to try chdir() there, but it still reports ~user/path (possibly after appending .git if allowed to do so, in which case it would report ~user/path.git). What this means is that if a whitelist wants to allow a user relative path, it needs to say "~" (for all users) or list user home directories like "~alice" "~bob". And no, you cannot say /home if the advertised way to access user home directories are ~alice,~bob, etc. The whole point of this is to avoid unnecessary aliasing issues. Anyway, because of this, daemon needs to do a bit more work to guard itself. Namely, it needs to make sure that the accessor does not try to exploit its leading path match rule by inserting /../ in the middle or hanging /.. at the end. I resurrected the belts and suspender paranoia code HPA did for this purpose. This check cannot be done in the enter_repo() unconditionally, because there are valid callers of enter_repo() that want to honor /../; authorized users coming over ssh to run send-pack and fetch-pack should be allowed to do so. Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-12-03 09:45:57 +00:00
len--;
/*
* We can handle arbitrary-sized buffers, but this remains as a
* sanity check on untrusted input.
*/
[PATCH] daemon.c and path.enter_repo(): revamp path validation. The whitelist of git-daemon is checked against return value from enter_repo(), and enter_repo() used to return the value obtained from getcwd() to avoid directory aliasing issues as discussed earier (mid October 2005). Unfortunately, it did not go well as we hoped. For example, /pub on a kernel.org public machine is a symlink to its real mountpoint, and it is understandable that the administrator does not want to adjust the whitelist every time /pub needs to point at a different partition for storage allcation or whatever reasons. Being able to keep using /pub/scm as the whitelist is a desirable property. So this version of enter_repo() reports what it used to chdir() and validate, but does not use getcwd() to canonicalize the directory name. When it sees a user relative path ~user/path, it internally resolves it to try chdir() there, but it still reports ~user/path (possibly after appending .git if allowed to do so, in which case it would report ~user/path.git). What this means is that if a whitelist wants to allow a user relative path, it needs to say "~" (for all users) or list user home directories like "~alice" "~bob". And no, you cannot say /home if the advertised way to access user home directories are ~alice,~bob, etc. The whole point of this is to avoid unnecessary aliasing issues. Anyway, because of this, daemon needs to do a bit more work to guard itself. Namely, it needs to make sure that the accessor does not try to exploit its leading path match rule by inserting /../ in the middle or hanging /.. at the end. I resurrected the belts and suspender paranoia code HPA did for this purpose. This check cannot be done in the enter_repo() unconditionally, because there are valid callers of enter_repo() that want to honor /../; authorized users coming over ssh to run send-pack and fetch-pack should be allowed to do so. Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-12-03 09:45:57 +00:00
if (PATH_MAX <= len)
return NULL;
strbuf_reset(&used_path);
strbuf_reset(&validated_path);
strbuf_add(&used_path, path, len);
strbuf_add(&validated_path, path, len);
if (used_path.buf[0] == '~') {
char *newpath = expand_user_path(used_path.buf, 0);
if (!newpath)
[PATCH] daemon.c and path.enter_repo(): revamp path validation. The whitelist of git-daemon is checked against return value from enter_repo(), and enter_repo() used to return the value obtained from getcwd() to avoid directory aliasing issues as discussed earier (mid October 2005). Unfortunately, it did not go well as we hoped. For example, /pub on a kernel.org public machine is a symlink to its real mountpoint, and it is understandable that the administrator does not want to adjust the whitelist every time /pub needs to point at a different partition for storage allcation or whatever reasons. Being able to keep using /pub/scm as the whitelist is a desirable property. So this version of enter_repo() reports what it used to chdir() and validate, but does not use getcwd() to canonicalize the directory name. When it sees a user relative path ~user/path, it internally resolves it to try chdir() there, but it still reports ~user/path (possibly after appending .git if allowed to do so, in which case it would report ~user/path.git). What this means is that if a whitelist wants to allow a user relative path, it needs to say "~" (for all users) or list user home directories like "~alice" "~bob". And no, you cannot say /home if the advertised way to access user home directories are ~alice,~bob, etc. The whole point of this is to avoid unnecessary aliasing issues. Anyway, because of this, daemon needs to do a bit more work to guard itself. Namely, it needs to make sure that the accessor does not try to exploit its leading path match rule by inserting /../ in the middle or hanging /.. at the end. I resurrected the belts and suspender paranoia code HPA did for this purpose. This check cannot be done in the enter_repo() unconditionally, because there are valid callers of enter_repo() that want to honor /../; authorized users coming over ssh to run send-pack and fetch-pack should be allowed to do so. Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-12-03 09:45:57 +00:00
return NULL;
strbuf_attach(&used_path, newpath, strlen(newpath),
strlen(newpath));
[PATCH] daemon.c and path.enter_repo(): revamp path validation. The whitelist of git-daemon is checked against return value from enter_repo(), and enter_repo() used to return the value obtained from getcwd() to avoid directory aliasing issues as discussed earier (mid October 2005). Unfortunately, it did not go well as we hoped. For example, /pub on a kernel.org public machine is a symlink to its real mountpoint, and it is understandable that the administrator does not want to adjust the whitelist every time /pub needs to point at a different partition for storage allcation or whatever reasons. Being able to keep using /pub/scm as the whitelist is a desirable property. So this version of enter_repo() reports what it used to chdir() and validate, but does not use getcwd() to canonicalize the directory name. When it sees a user relative path ~user/path, it internally resolves it to try chdir() there, but it still reports ~user/path (possibly after appending .git if allowed to do so, in which case it would report ~user/path.git). What this means is that if a whitelist wants to allow a user relative path, it needs to say "~" (for all users) or list user home directories like "~alice" "~bob". And no, you cannot say /home if the advertised way to access user home directories are ~alice,~bob, etc. The whole point of this is to avoid unnecessary aliasing issues. Anyway, because of this, daemon needs to do a bit more work to guard itself. Namely, it needs to make sure that the accessor does not try to exploit its leading path match rule by inserting /../ in the middle or hanging /.. at the end. I resurrected the belts and suspender paranoia code HPA did for this purpose. This check cannot be done in the enter_repo() unconditionally, because there are valid callers of enter_repo() that want to honor /../; authorized users coming over ssh to run send-pack and fetch-pack should be allowed to do so. Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-12-03 09:45:57 +00:00
}
for (i = 0; suffix[i]; i++) {
standardize and improve lookup rules for external local repos When you specify a local repository on the command line of clone, ls-remote, upload-pack, receive-pack, or upload-archive, or in a request to git-daemon, we perform a little bit of lookup magic, doing things like looking in working trees for .git directories and appending ".git" for bare repos. For clone, this magic happens in get_repo_path. For everything else, it happens in enter_repo. In both cases, there are some ambiguous or confusing cases that aren't handled well, and there is one case that is not handled the same by both methods. This patch tries to provide (and test!) standard, sensible lookup rules for both code paths. The intended changes are: 1. When looking up "foo", we have always preferred a working tree "foo" (containing "foo/.git" over the bare "foo.git". But we did not prefer a bare "foo" over "foo.git". With this patch, we do so. 2. We would select directories that existed but didn't actually look like git repositories. With this patch, we make sure a selected directory looks like a git repo. Not only is this more sensible in general, but it will help anybody who is negatively affected by change (1) negatively (e.g., if they had "foo.git" next to its separate work tree "foo", and expect to keep finding "foo.git" when they reference "foo"). 3. The enter_repo code path would, given "foo", look for "foo.git/.git" (i.e., do the ".git" append magic even for a repo with working tree). The clone code path did not; with this patch, they now behave the same. In the unlikely case of a working tree overlaying a bare repo (i.e., a ".git" directory _inside_ a bare repo), we continue to treat it as a working tree (prefering the "inner" .git over the bare repo). This is mainly because the combination seems nonsensical, and I'd rather stick with existing behavior on the off chance that somebody is relying on it. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-02-02 21:59:13 +00:00
struct stat st;
size_t baselen = used_path.len;
strbuf_addstr(&used_path, suffix[i]);
if (!stat(used_path.buf, &st) &&
standardize and improve lookup rules for external local repos When you specify a local repository on the command line of clone, ls-remote, upload-pack, receive-pack, or upload-archive, or in a request to git-daemon, we perform a little bit of lookup magic, doing things like looking in working trees for .git directories and appending ".git" for bare repos. For clone, this magic happens in get_repo_path. For everything else, it happens in enter_repo. In both cases, there are some ambiguous or confusing cases that aren't handled well, and there is one case that is not handled the same by both methods. This patch tries to provide (and test!) standard, sensible lookup rules for both code paths. The intended changes are: 1. When looking up "foo", we have always preferred a working tree "foo" (containing "foo/.git" over the bare "foo.git". But we did not prefer a bare "foo" over "foo.git". With this patch, we do so. 2. We would select directories that existed but didn't actually look like git repositories. With this patch, we make sure a selected directory looks like a git repo. Not only is this more sensible in general, but it will help anybody who is negatively affected by change (1) negatively (e.g., if they had "foo.git" next to its separate work tree "foo", and expect to keep finding "foo.git" when they reference "foo"). 3. The enter_repo code path would, given "foo", look for "foo.git/.git" (i.e., do the ".git" append magic even for a repo with working tree). The clone code path did not; with this patch, they now behave the same. In the unlikely case of a working tree overlaying a bare repo (i.e., a ".git" directory _inside_ a bare repo), we continue to treat it as a working tree (prefering the "inner" .git over the bare repo). This is mainly because the combination seems nonsensical, and I'd rather stick with existing behavior on the off chance that somebody is relying on it. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-02-02 21:59:13 +00:00
(S_ISREG(st.st_mode) ||
(S_ISDIR(st.st_mode) && is_git_directory(used_path.buf)))) {
strbuf_addstr(&validated_path, suffix[i]);
[PATCH] daemon.c and path.enter_repo(): revamp path validation. The whitelist of git-daemon is checked against return value from enter_repo(), and enter_repo() used to return the value obtained from getcwd() to avoid directory aliasing issues as discussed earier (mid October 2005). Unfortunately, it did not go well as we hoped. For example, /pub on a kernel.org public machine is a symlink to its real mountpoint, and it is understandable that the administrator does not want to adjust the whitelist every time /pub needs to point at a different partition for storage allcation or whatever reasons. Being able to keep using /pub/scm as the whitelist is a desirable property. So this version of enter_repo() reports what it used to chdir() and validate, but does not use getcwd() to canonicalize the directory name. When it sees a user relative path ~user/path, it internally resolves it to try chdir() there, but it still reports ~user/path (possibly after appending .git if allowed to do so, in which case it would report ~user/path.git). What this means is that if a whitelist wants to allow a user relative path, it needs to say "~" (for all users) or list user home directories like "~alice" "~bob". And no, you cannot say /home if the advertised way to access user home directories are ~alice,~bob, etc. The whole point of this is to avoid unnecessary aliasing issues. Anyway, because of this, daemon needs to do a bit more work to guard itself. Namely, it needs to make sure that the accessor does not try to exploit its leading path match rule by inserting /../ in the middle or hanging /.. at the end. I resurrected the belts and suspender paranoia code HPA did for this purpose. This check cannot be done in the enter_repo() unconditionally, because there are valid callers of enter_repo() that want to honor /../; authorized users coming over ssh to run send-pack and fetch-pack should be allowed to do so. Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-12-03 09:45:57 +00:00
break;
}
strbuf_setlen(&used_path, baselen);
[PATCH] daemon.c and path.enter_repo(): revamp path validation. The whitelist of git-daemon is checked against return value from enter_repo(), and enter_repo() used to return the value obtained from getcwd() to avoid directory aliasing issues as discussed earier (mid October 2005). Unfortunately, it did not go well as we hoped. For example, /pub on a kernel.org public machine is a symlink to its real mountpoint, and it is understandable that the administrator does not want to adjust the whitelist every time /pub needs to point at a different partition for storage allcation or whatever reasons. Being able to keep using /pub/scm as the whitelist is a desirable property. So this version of enter_repo() reports what it used to chdir() and validate, but does not use getcwd() to canonicalize the directory name. When it sees a user relative path ~user/path, it internally resolves it to try chdir() there, but it still reports ~user/path (possibly after appending .git if allowed to do so, in which case it would report ~user/path.git). What this means is that if a whitelist wants to allow a user relative path, it needs to say "~" (for all users) or list user home directories like "~alice" "~bob". And no, you cannot say /home if the advertised way to access user home directories are ~alice,~bob, etc. The whole point of this is to avoid unnecessary aliasing issues. Anyway, because of this, daemon needs to do a bit more work to guard itself. Namely, it needs to make sure that the accessor does not try to exploit its leading path match rule by inserting /../ in the middle or hanging /.. at the end. I resurrected the belts and suspender paranoia code HPA did for this purpose. This check cannot be done in the enter_repo() unconditionally, because there are valid callers of enter_repo() that want to honor /../; authorized users coming over ssh to run send-pack and fetch-pack should be allowed to do so. Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-12-03 09:45:57 +00:00
}
if (!suffix[i])
return NULL;
gitfile = read_gitfile(used_path.buf);
if (gitfile) {
strbuf_reset(&used_path);
strbuf_addstr(&used_path, gitfile);
}
if (chdir(used_path.buf))
return NULL;
path = validated_path.buf;
}
else {
const char *gitfile = read_gitfile(path);
if (gitfile)
path = gitfile;
if (chdir(path))
return NULL;
}
if (is_git_directory(".")) {
set_git_dir(".");
check_repository_format();
[PATCH] daemon.c and path.enter_repo(): revamp path validation. The whitelist of git-daemon is checked against return value from enter_repo(), and enter_repo() used to return the value obtained from getcwd() to avoid directory aliasing issues as discussed earier (mid October 2005). Unfortunately, it did not go well as we hoped. For example, /pub on a kernel.org public machine is a symlink to its real mountpoint, and it is understandable that the administrator does not want to adjust the whitelist every time /pub needs to point at a different partition for storage allcation or whatever reasons. Being able to keep using /pub/scm as the whitelist is a desirable property. So this version of enter_repo() reports what it used to chdir() and validate, but does not use getcwd() to canonicalize the directory name. When it sees a user relative path ~user/path, it internally resolves it to try chdir() there, but it still reports ~user/path (possibly after appending .git if allowed to do so, in which case it would report ~user/path.git). What this means is that if a whitelist wants to allow a user relative path, it needs to say "~" (for all users) or list user home directories like "~alice" "~bob". And no, you cannot say /home if the advertised way to access user home directories are ~alice,~bob, etc. The whole point of this is to avoid unnecessary aliasing issues. Anyway, because of this, daemon needs to do a bit more work to guard itself. Namely, it needs to make sure that the accessor does not try to exploit its leading path match rule by inserting /../ in the middle or hanging /.. at the end. I resurrected the belts and suspender paranoia code HPA did for this purpose. This check cannot be done in the enter_repo() unconditionally, because there are valid callers of enter_repo() that want to honor /../; authorized users coming over ssh to run send-pack and fetch-pack should be allowed to do so. Signed-off-by: Junio C Hamano <junkio@cox.net>
2005-12-03 09:45:57 +00:00
return path;
}
return NULL;
}
static int calc_shared_perm(int mode)
{
int tweak;
if (get_shared_repository() < 0)
tweak = -get_shared_repository();
else
tweak = get_shared_repository();
if (!(mode & S_IWUSR))
tweak &= ~0222;
if (mode & S_IXUSR)
/* Copy read bits to execute bits */
tweak |= (tweak & 0444) >> 2;
if (get_shared_repository() < 0)
mode = (mode & ~0777) | tweak;
else
mode |= tweak;
return mode;
}
int adjust_shared_perm(const char *path)
{
int old_mode, new_mode;
if (!get_shared_repository())
return 0;
if (get_st_mode_bits(path, &old_mode) < 0)
return -1;
new_mode = calc_shared_perm(old_mode);
if (S_ISDIR(old_mode)) {
/* Copy read bits to execute bits */
new_mode |= (new_mode & 0444) >> 2;
new_mode |= FORCE_DIR_SET_GID;
}
if (((old_mode ^ new_mode) & ~S_IFMT) &&
chmod(path, (new_mode & ~S_IFMT)) < 0)
return -2;
return 0;
}
void safe_create_dir(const char *dir, int share)
{
if (mkdir(dir, 0777) < 0) {
if (errno != EEXIST) {
perror(dir);
exit(1);
}
}
else if (share && adjust_shared_perm(dir))
die(_("Could not make %s writable by group"), dir);
}
static int have_same_root(const char *path1, const char *path2)
{
int is_abs1, is_abs2;
is_abs1 = is_absolute_path(path1);
is_abs2 = is_absolute_path(path2);
return (is_abs1 && is_abs2 && tolower(path1[0]) == tolower(path2[0])) ||
(!is_abs1 && !is_abs2);
}
path.c: refactor relative_path(), not only strip prefix Original design of relative_path() is simple, just strip the prefix (*base) from the absolute path (*abs). In most cases, we need a real relative path, such as: ../foo, ../../bar. That's why there is another reimplementation (path_relative()) in quote.c. Borrow some codes from path_relative() in quote.c to refactor relative_path() in path.c, so that it could return real relative path, and user can reuse this function without reimplementing his/her own. The function path_relative() in quote.c will be substituted, and I would use the new relative_path() function when implementing the interactive git-clean later. Different results for relative_path() before and after this refactor: abs path base path relative (original) relative (refactor) ======== ========= =================== =================== /a/b /a/b . ./ /a/b/ /a/b . ./ /a /a/b/ /a ../ / /a/b/ / ../../ /a/c /a/b/ /a/c ../c /x/y /a/b/ /x/y ../../x/y a/b/ a/b/ . ./ a/b/ a/b . ./ a a/b a ../ x/y a/b/ x/y ../../x/y a/c a/b a/c ../c (empty) (null) (empty) ./ (empty) (empty) (empty) ./ (empty) /a/b (empty) ./ (null) (null) (null) ./ (null) (empty) (null) ./ (null) /a/b (segfault) ./ You may notice that return value "." has been changed to "./". It is because: * Function quote_path_relative() in quote.c will show the relative path as "./" if abs(in) and base(prefix) are the same. * Function relative_path() is called only once (in setup.c), and it will be OK for the return value as "./" instead of ".". Signed-off-by: Jiang Xin <worldhello.net@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-06-25 15:53:43 +00:00
/*
* Give path as relative to prefix.
*
* The strbuf may or may not be used, so do not assume it contains the
* returned path.
*/
const char *relative_path(const char *in, const char *prefix,
struct strbuf *sb)
Make git_dir a path relative to work_tree in setup_work_tree() Once we find the absolute paths for git_dir and work_tree, we can make git_dir a relative path since we know pwd will be work_tree. This should save the kernel some time traversing the path to work_tree all the time if git_dir is inside work_tree. Daniel's patch didn't apply for me as-is, so I recreated it with some differences, and here are the numbers from ten runs each. There is some IO for me - probably due to more-or-less random flushing of the journal - so the variation is bigger than I'd like, but whatever: Before: real 0m8.135s real 0m7.933s real 0m8.080s real 0m7.954s real 0m7.949s real 0m8.112s real 0m7.934s real 0m8.059s real 0m7.979s real 0m8.038s After: real 0m7.685s real 0m7.968s real 0m7.703s real 0m7.850s real 0m7.995s real 0m7.817s real 0m7.963s real 0m7.955s real 0m7.848s real 0m7.969s Now, going by "best of ten" (on the assumption that the longer numbers are all due to IO), I'm saying a 7.933s -> 7.685s reduction, and it does seem to be outside of the noise (ie the "after" case never broke 8s, while the "before" case did so half the time). So looks like about 3% to me. Doing it for a slightly smaller test-case (just the "arch" subdirectory) gets more stable numbers probably due to not filling the journal with metadata updates, so we have: Before: real 0m1.633s real 0m1.633s real 0m1.633s real 0m1.632s real 0m1.632s real 0m1.630s real 0m1.634s real 0m1.631s real 0m1.632s real 0m1.632s After: real 0m1.610s real 0m1.609s real 0m1.610s real 0m1.608s real 0m1.607s real 0m1.610s real 0m1.609s real 0m1.611s real 0m1.608s real 0m1.611s where I'ld just take the averages and say 1.632 vs 1.610, which is just over 1% peformance improvement. So it's not in the noise, but it's not as big as I initially thought and measured. (That said, it obviously depends on how deep the working directory path is too, and whether it is behind NFS or something else that might need to cause more work to look up). Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-06-19 19:34:06 +00:00
{
path.c: refactor relative_path(), not only strip prefix Original design of relative_path() is simple, just strip the prefix (*base) from the absolute path (*abs). In most cases, we need a real relative path, such as: ../foo, ../../bar. That's why there is another reimplementation (path_relative()) in quote.c. Borrow some codes from path_relative() in quote.c to refactor relative_path() in path.c, so that it could return real relative path, and user can reuse this function without reimplementing his/her own. The function path_relative() in quote.c will be substituted, and I would use the new relative_path() function when implementing the interactive git-clean later. Different results for relative_path() before and after this refactor: abs path base path relative (original) relative (refactor) ======== ========= =================== =================== /a/b /a/b . ./ /a/b/ /a/b . ./ /a /a/b/ /a ../ / /a/b/ / ../../ /a/c /a/b/ /a/c ../c /x/y /a/b/ /x/y ../../x/y a/b/ a/b/ . ./ a/b/ a/b . ./ a a/b a ../ x/y a/b/ x/y ../../x/y a/c a/b a/c ../c (empty) (null) (empty) ./ (empty) (empty) (empty) ./ (empty) /a/b (empty) ./ (null) (null) (null) ./ (null) (empty) (null) ./ (null) /a/b (segfault) ./ You may notice that return value "." has been changed to "./". It is because: * Function quote_path_relative() in quote.c will show the relative path as "./" if abs(in) and base(prefix) are the same. * Function relative_path() is called only once (in setup.c), and it will be OK for the return value as "./" instead of ".". Signed-off-by: Jiang Xin <worldhello.net@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-06-25 15:53:43 +00:00
int in_len = in ? strlen(in) : 0;
int prefix_len = prefix ? strlen(prefix) : 0;
int in_off = 0;
int prefix_off = 0;
int i = 0, j = 0;
path.c: refactor relative_path(), not only strip prefix Original design of relative_path() is simple, just strip the prefix (*base) from the absolute path (*abs). In most cases, we need a real relative path, such as: ../foo, ../../bar. That's why there is another reimplementation (path_relative()) in quote.c. Borrow some codes from path_relative() in quote.c to refactor relative_path() in path.c, so that it could return real relative path, and user can reuse this function without reimplementing his/her own. The function path_relative() in quote.c will be substituted, and I would use the new relative_path() function when implementing the interactive git-clean later. Different results for relative_path() before and after this refactor: abs path base path relative (original) relative (refactor) ======== ========= =================== =================== /a/b /a/b . ./ /a/b/ /a/b . ./ /a /a/b/ /a ../ / /a/b/ / ../../ /a/c /a/b/ /a/c ../c /x/y /a/b/ /x/y ../../x/y a/b/ a/b/ . ./ a/b/ a/b . ./ a a/b a ../ x/y a/b/ x/y ../../x/y a/c a/b a/c ../c (empty) (null) (empty) ./ (empty) (empty) (empty) ./ (empty) /a/b (empty) ./ (null) (null) (null) ./ (null) (empty) (null) ./ (null) /a/b (segfault) ./ You may notice that return value "." has been changed to "./". It is because: * Function quote_path_relative() in quote.c will show the relative path as "./" if abs(in) and base(prefix) are the same. * Function relative_path() is called only once (in setup.c), and it will be OK for the return value as "./" instead of ".". Signed-off-by: Jiang Xin <worldhello.net@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-06-25 15:53:43 +00:00
if (!in_len)
return "./";
else if (!prefix_len)
return in;
if (have_same_root(in, prefix))
/* bypass dos_drive, for "c:" is identical to "C:" */
i = j = has_dos_drive_prefix(in);
else {
return in;
}
path.c: refactor relative_path(), not only strip prefix Original design of relative_path() is simple, just strip the prefix (*base) from the absolute path (*abs). In most cases, we need a real relative path, such as: ../foo, ../../bar. That's why there is another reimplementation (path_relative()) in quote.c. Borrow some codes from path_relative() in quote.c to refactor relative_path() in path.c, so that it could return real relative path, and user can reuse this function without reimplementing his/her own. The function path_relative() in quote.c will be substituted, and I would use the new relative_path() function when implementing the interactive git-clean later. Different results for relative_path() before and after this refactor: abs path base path relative (original) relative (refactor) ======== ========= =================== =================== /a/b /a/b . ./ /a/b/ /a/b . ./ /a /a/b/ /a ../ / /a/b/ / ../../ /a/c /a/b/ /a/c ../c /x/y /a/b/ /x/y ../../x/y a/b/ a/b/ . ./ a/b/ a/b . ./ a a/b a ../ x/y a/b/ x/y ../../x/y a/c a/b a/c ../c (empty) (null) (empty) ./ (empty) (empty) (empty) ./ (empty) /a/b (empty) ./ (null) (null) (null) ./ (null) (empty) (null) ./ (null) /a/b (segfault) ./ You may notice that return value "." has been changed to "./". It is because: * Function quote_path_relative() in quote.c will show the relative path as "./" if abs(in) and base(prefix) are the same. * Function relative_path() is called only once (in setup.c), and it will be OK for the return value as "./" instead of ".". Signed-off-by: Jiang Xin <worldhello.net@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-06-25 15:53:43 +00:00
while (i < prefix_len && j < in_len && prefix[i] == in[j]) {
if (is_dir_sep(prefix[i])) {
while (is_dir_sep(prefix[i]))
i++;
path.c: refactor relative_path(), not only strip prefix Original design of relative_path() is simple, just strip the prefix (*base) from the absolute path (*abs). In most cases, we need a real relative path, such as: ../foo, ../../bar. That's why there is another reimplementation (path_relative()) in quote.c. Borrow some codes from path_relative() in quote.c to refactor relative_path() in path.c, so that it could return real relative path, and user can reuse this function without reimplementing his/her own. The function path_relative() in quote.c will be substituted, and I would use the new relative_path() function when implementing the interactive git-clean later. Different results for relative_path() before and after this refactor: abs path base path relative (original) relative (refactor) ======== ========= =================== =================== /a/b /a/b . ./ /a/b/ /a/b . ./ /a /a/b/ /a ../ / /a/b/ / ../../ /a/c /a/b/ /a/c ../c /x/y /a/b/ /x/y ../../x/y a/b/ a/b/ . ./ a/b/ a/b . ./ a a/b a ../ x/y a/b/ x/y ../../x/y a/c a/b a/c ../c (empty) (null) (empty) ./ (empty) (empty) (empty) ./ (empty) /a/b (empty) ./ (null) (null) (null) ./ (null) (empty) (null) ./ (null) /a/b (segfault) ./ You may notice that return value "." has been changed to "./". It is because: * Function quote_path_relative() in quote.c will show the relative path as "./" if abs(in) and base(prefix) are the same. * Function relative_path() is called only once (in setup.c), and it will be OK for the return value as "./" instead of ".". Signed-off-by: Jiang Xin <worldhello.net@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-06-25 15:53:43 +00:00
while (is_dir_sep(in[j]))
j++;
prefix_off = i;
in_off = j;
} else {
i++;
j++;
}
}
if (
/* "prefix" seems like prefix of "in" */
i >= prefix_len &&
/*
* but "/foo" is not a prefix of "/foobar"
* (i.e. prefix not end with '/')
*/
prefix_off < prefix_len) {
if (j >= in_len) {
/* in="/a/b", prefix="/a/b" */
in_off = in_len;
} else if (is_dir_sep(in[j])) {
/* in="/a/b/c", prefix="/a/b" */
while (is_dir_sep(in[j]))
j++;
path.c: refactor relative_path(), not only strip prefix Original design of relative_path() is simple, just strip the prefix (*base) from the absolute path (*abs). In most cases, we need a real relative path, such as: ../foo, ../../bar. That's why there is another reimplementation (path_relative()) in quote.c. Borrow some codes from path_relative() in quote.c to refactor relative_path() in path.c, so that it could return real relative path, and user can reuse this function without reimplementing his/her own. The function path_relative() in quote.c will be substituted, and I would use the new relative_path() function when implementing the interactive git-clean later. Different results for relative_path() before and after this refactor: abs path base path relative (original) relative (refactor) ======== ========= =================== =================== /a/b /a/b . ./ /a/b/ /a/b . ./ /a /a/b/ /a ../ / /a/b/ / ../../ /a/c /a/b/ /a/c ../c /x/y /a/b/ /x/y ../../x/y a/b/ a/b/ . ./ a/b/ a/b . ./ a a/b a ../ x/y a/b/ x/y ../../x/y a/c a/b a/c ../c (empty) (null) (empty) ./ (empty) (empty) (empty) ./ (empty) /a/b (empty) ./ (null) (null) (null) ./ (null) (empty) (null) ./ (null) /a/b (segfault) ./ You may notice that return value "." has been changed to "./". It is because: * Function quote_path_relative() in quote.c will show the relative path as "./" if abs(in) and base(prefix) are the same. * Function relative_path() is called only once (in setup.c), and it will be OK for the return value as "./" instead of ".". Signed-off-by: Jiang Xin <worldhello.net@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-06-25 15:53:43 +00:00
in_off = j;
} else {
/* in="/a/bbb/c", prefix="/a/b" */
i = prefix_off;
}
} else if (
/* "in" is short than "prefix" */
j >= in_len &&
/* "in" not end with '/' */
in_off < in_len) {
if (is_dir_sep(prefix[i])) {
/* in="/a/b", prefix="/a/b/c/" */
while (is_dir_sep(prefix[i]))
i++;
in_off = in_len;
}
}
in += in_off;
in_len -= in_off;
if (i >= prefix_len) {
if (!in_len)
return "./";
else
return in;
}
strbuf_reset(sb);
strbuf_grow(sb, in_len);
while (i < prefix_len) {
if (is_dir_sep(prefix[i])) {
strbuf_addstr(sb, "../");
while (is_dir_sep(prefix[i]))
i++;
continue;
}
i++;
}
path.c: refactor relative_path(), not only strip prefix Original design of relative_path() is simple, just strip the prefix (*base) from the absolute path (*abs). In most cases, we need a real relative path, such as: ../foo, ../../bar. That's why there is another reimplementation (path_relative()) in quote.c. Borrow some codes from path_relative() in quote.c to refactor relative_path() in path.c, so that it could return real relative path, and user can reuse this function without reimplementing his/her own. The function path_relative() in quote.c will be substituted, and I would use the new relative_path() function when implementing the interactive git-clean later. Different results for relative_path() before and after this refactor: abs path base path relative (original) relative (refactor) ======== ========= =================== =================== /a/b /a/b . ./ /a/b/ /a/b . ./ /a /a/b/ /a ../ / /a/b/ / ../../ /a/c /a/b/ /a/c ../c /x/y /a/b/ /x/y ../../x/y a/b/ a/b/ . ./ a/b/ a/b . ./ a a/b a ../ x/y a/b/ x/y ../../x/y a/c a/b a/c ../c (empty) (null) (empty) ./ (empty) (empty) (empty) ./ (empty) /a/b (empty) ./ (null) (null) (null) ./ (null) (empty) (null) ./ (null) /a/b (segfault) ./ You may notice that return value "." has been changed to "./". It is because: * Function quote_path_relative() in quote.c will show the relative path as "./" if abs(in) and base(prefix) are the same. * Function relative_path() is called only once (in setup.c), and it will be OK for the return value as "./" instead of ".". Signed-off-by: Jiang Xin <worldhello.net@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-06-25 15:53:43 +00:00
if (!is_dir_sep(prefix[prefix_len - 1]))
strbuf_addstr(sb, "../");
strbuf_addstr(sb, in);
return sb->buf;
Make git_dir a path relative to work_tree in setup_work_tree() Once we find the absolute paths for git_dir and work_tree, we can make git_dir a relative path since we know pwd will be work_tree. This should save the kernel some time traversing the path to work_tree all the time if git_dir is inside work_tree. Daniel's patch didn't apply for me as-is, so I recreated it with some differences, and here are the numbers from ten runs each. There is some IO for me - probably due to more-or-less random flushing of the journal - so the variation is bigger than I'd like, but whatever: Before: real 0m8.135s real 0m7.933s real 0m8.080s real 0m7.954s real 0m7.949s real 0m8.112s real 0m7.934s real 0m8.059s real 0m7.979s real 0m8.038s After: real 0m7.685s real 0m7.968s real 0m7.703s real 0m7.850s real 0m7.995s real 0m7.817s real 0m7.963s real 0m7.955s real 0m7.848s real 0m7.969s Now, going by "best of ten" (on the assumption that the longer numbers are all due to IO), I'm saying a 7.933s -> 7.685s reduction, and it does seem to be outside of the noise (ie the "after" case never broke 8s, while the "before" case did so half the time). So looks like about 3% to me. Doing it for a slightly smaller test-case (just the "arch" subdirectory) gets more stable numbers probably due to not filling the journal with metadata updates, so we have: Before: real 0m1.633s real 0m1.633s real 0m1.633s real 0m1.632s real 0m1.632s real 0m1.630s real 0m1.634s real 0m1.631s real 0m1.632s real 0m1.632s After: real 0m1.610s real 0m1.609s real 0m1.610s real 0m1.608s real 0m1.607s real 0m1.610s real 0m1.609s real 0m1.611s real 0m1.608s real 0m1.611s where I'ld just take the averages and say 1.632 vs 1.610, which is just over 1% peformance improvement. So it's not in the noise, but it's not as big as I initially thought and measured. (That said, it obviously depends on how deep the working directory path is too, and whether it is behind NFS or something else that might need to cause more work to look up). Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-06-19 19:34:06 +00:00
}
/*
* A simpler implementation of relative_path
*
* Get relative path by removing "prefix" from "in". This function
* first appears in v1.5.6-1-g044bbbc, and makes git_dir shorter
* to increase performance when traversing the path to work_tree.
*/
const char *remove_leading_path(const char *in, const char *prefix)
{
static struct strbuf buf = STRBUF_INIT;
int i = 0, j = 0;
if (!prefix || !prefix[0])
return in;
while (prefix[i]) {
if (is_dir_sep(prefix[i])) {
if (!is_dir_sep(in[j]))
return in;
while (is_dir_sep(prefix[i]))
i++;
while (is_dir_sep(in[j]))
j++;
continue;
} else if (in[j] != prefix[i]) {
return in;
}
i++;
j++;
}
if (
/* "/foo" is a prefix of "/foo" */
in[j] &&
/* "/foo" is not a prefix of "/foobar" */
!is_dir_sep(prefix[i-1]) && !is_dir_sep(in[j])
)
return in;
while (is_dir_sep(in[j]))
j++;
strbuf_reset(&buf);
if (!in[j])
strbuf_addstr(&buf, ".");
else
strbuf_addstr(&buf, in + j);
return buf.buf;
}
/*
* It is okay if dst == src, but they should not overlap otherwise.
*
* Performs the following normalizations on src, storing the result in dst:
* - Ensures that components are separated by '/' (Windows only)
* - Squashes sequences of '/' except "//server/share" on Windows
* - Removes "." components.
* - Removes ".." components, and the components the precede them.
* Returns failure (non-zero) if a ".." component appears as first path
* component anytime during the normalization. Otherwise, returns success (0).
*
* Note that this function is purely textual. It does not follow symlinks,
* verify the existence of the path, or make any system calls.
*
* prefix_len != NULL is for a specific case of prefix_pathspec():
* assume that src == dst and src[0..prefix_len-1] is already
* normalized, any time "../" eats up to the prefix_len part,
* prefix_len is reduced. In the end prefix_len is the remaining
* prefix that has not been overridden by user pathspec.
*
* NEEDSWORK: This function doesn't perform normalization w.r.t. trailing '/'.
* For everything but the root folder itself, the normalized path should not
* end with a '/', then the callers need to be fixed up accordingly.
*
*/
int normalize_path_copy_len(char *dst, const char *src, int *prefix_len)
{
char *dst0;
const char *end;
/*
* Copy initial part of absolute path: "/", "C:/", "//server/share/".
*/
end = src + offset_1st_component(src);
while (src < end) {
char c = *src++;
if (is_dir_sep(c))
c = '/';
*dst++ = c;
}
dst0 = dst;
while (is_dir_sep(*src))
src++;
for (;;) {
char c = *src;
/*
* A path component that begins with . could be
* special:
* (1) "." and ends -- ignore and terminate.
* (2) "./" -- ignore them, eat slash and continue.
* (3) ".." and ends -- strip one and terminate.
* (4) "../" -- strip one, eat slash and continue.
*/
if (c == '.') {
if (!src[1]) {
/* (1) */
src++;
} else if (is_dir_sep(src[1])) {
/* (2) */
src += 2;
while (is_dir_sep(*src))
src++;
continue;
} else if (src[1] == '.') {
if (!src[2]) {
/* (3) */
src += 2;
goto up_one;
} else if (is_dir_sep(src[2])) {
/* (4) */
src += 3;
while (is_dir_sep(*src))
src++;
goto up_one;
}
}
}
/* copy up to the next '/', and eat all '/' */
while ((c = *src++) != '\0' && !is_dir_sep(c))
*dst++ = c;
if (is_dir_sep(c)) {
*dst++ = '/';
while (is_dir_sep(c))
c = *src++;
src--;
} else if (!c)
break;
continue;
up_one:
/*
* dst0..dst is prefix portion, and dst[-1] is '/';
* go up one level.
*/
dst--; /* go to trailing '/' */
if (dst <= dst0)
return -1;
/* Windows: dst[-1] cannot be backslash anymore */
while (dst0 < dst && dst[-1] != '/')
dst--;
if (prefix_len && *prefix_len > dst - dst0)
*prefix_len = dst - dst0;
}
*dst = '\0';
return 0;
}
int normalize_path_copy(char *dst, const char *src)
{
return normalize_path_copy_len(dst, src, NULL);
}
/*
* path = Canonical absolute path
longest_ancestor_length(): require prefix list entries to be normalized Move the responsibility for normalizing prefixes from longest_ancestor_length() to its callers. Use slightly different normalizations at the two callers: In setup_git_directory_gently_1(), use the old normalization, which ignores paths that are not usable. In the next commit we will change this caller to also resolve symlinks in the paths from GIT_CEILING_DIRECTORIES as part of the normalization. In "test-path-utils longest_ancestor_length", use the old normalization, but die() if any paths are unusable. Also change t0060 to only pass normalized paths to the test program (no empty entries or non-absolute paths, strip trailing slashes from the paths, and remove tests that thereby become redundant). The point of this change is to reduce the scope of the ancestor_length tests in t0060 from testing normalization+longest_prefix to testing only mostly longest_prefix. This is necessary because when setup_git_directory_gently_1() starts resolving symlinks as part of its normalization, it will not be reasonable to do the same in the test suite, because that would make the test results depend on the contents of the root directory of the filesystem on which the test is run. HOWEVER: under Windows, bash mangles arguments that look like absolute POSIX paths into DOS paths. So we have to retain the level of normalization done by normalize_path_copy() to convert the bash-mangled DOS paths (which contain backslashes) into paths that use forward slashes. Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Jeff King <peff@peff.net>
2012-10-28 16:16:25 +00:00
* prefixes = string_list containing normalized, absolute paths without
* trailing slashes (except for the root directory, which is denoted by "/").
*
longest_ancestor_length(): require prefix list entries to be normalized Move the responsibility for normalizing prefixes from longest_ancestor_length() to its callers. Use slightly different normalizations at the two callers: In setup_git_directory_gently_1(), use the old normalization, which ignores paths that are not usable. In the next commit we will change this caller to also resolve symlinks in the paths from GIT_CEILING_DIRECTORIES as part of the normalization. In "test-path-utils longest_ancestor_length", use the old normalization, but die() if any paths are unusable. Also change t0060 to only pass normalized paths to the test program (no empty entries or non-absolute paths, strip trailing slashes from the paths, and remove tests that thereby become redundant). The point of this change is to reduce the scope of the ancestor_length tests in t0060 from testing normalization+longest_prefix to testing only mostly longest_prefix. This is necessary because when setup_git_directory_gently_1() starts resolving symlinks as part of its normalization, it will not be reasonable to do the same in the test suite, because that would make the test results depend on the contents of the root directory of the filesystem on which the test is run. HOWEVER: under Windows, bash mangles arguments that look like absolute POSIX paths into DOS paths. So we have to retain the level of normalization done by normalize_path_copy() to convert the bash-mangled DOS paths (which contain backslashes) into paths that use forward slashes. Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Jeff King <peff@peff.net>
2012-10-28 16:16:25 +00:00
* Determines, for each path in prefixes, whether the "prefix"
* is an ancestor directory of path. Returns the length of the longest
* ancestor directory, excluding any trailing slashes, or -1 if no prefix
* is an ancestor. (Note that this means 0 is returned if prefixes is
* ["/"].) "/foo" is not considered an ancestor of "/foobar". Directories
* are not considered to be their own ancestors. path must be in a
* canonical form: empty components, or "." or ".." components are not
longest_ancestor_length(): require prefix list entries to be normalized Move the responsibility for normalizing prefixes from longest_ancestor_length() to its callers. Use slightly different normalizations at the two callers: In setup_git_directory_gently_1(), use the old normalization, which ignores paths that are not usable. In the next commit we will change this caller to also resolve symlinks in the paths from GIT_CEILING_DIRECTORIES as part of the normalization. In "test-path-utils longest_ancestor_length", use the old normalization, but die() if any paths are unusable. Also change t0060 to only pass normalized paths to the test program (no empty entries or non-absolute paths, strip trailing slashes from the paths, and remove tests that thereby become redundant). The point of this change is to reduce the scope of the ancestor_length tests in t0060 from testing normalization+longest_prefix to testing only mostly longest_prefix. This is necessary because when setup_git_directory_gently_1() starts resolving symlinks as part of its normalization, it will not be reasonable to do the same in the test suite, because that would make the test results depend on the contents of the root directory of the filesystem on which the test is run. HOWEVER: under Windows, bash mangles arguments that look like absolute POSIX paths into DOS paths. So we have to retain the level of normalization done by normalize_path_copy() to convert the bash-mangled DOS paths (which contain backslashes) into paths that use forward slashes. Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Jeff King <peff@peff.net>
2012-10-28 16:16:25 +00:00
* allowed.
*/
int longest_ancestor_length(const char *path, struct string_list *prefixes)
{
int i, max_len = -1;
if (!strcmp(path, "/"))
return -1;
for (i = 0; i < prefixes->nr; i++) {
const char *ceil = prefixes->items[i].string;
int len = strlen(ceil);
longest_ancestor_length(): require prefix list entries to be normalized Move the responsibility for normalizing prefixes from longest_ancestor_length() to its callers. Use slightly different normalizations at the two callers: In setup_git_directory_gently_1(), use the old normalization, which ignores paths that are not usable. In the next commit we will change this caller to also resolve symlinks in the paths from GIT_CEILING_DIRECTORIES as part of the normalization. In "test-path-utils longest_ancestor_length", use the old normalization, but die() if any paths are unusable. Also change t0060 to only pass normalized paths to the test program (no empty entries or non-absolute paths, strip trailing slashes from the paths, and remove tests that thereby become redundant). The point of this change is to reduce the scope of the ancestor_length tests in t0060 from testing normalization+longest_prefix to testing only mostly longest_prefix. This is necessary because when setup_git_directory_gently_1() starts resolving symlinks as part of its normalization, it will not be reasonable to do the same in the test suite, because that would make the test results depend on the contents of the root directory of the filesystem on which the test is run. HOWEVER: under Windows, bash mangles arguments that look like absolute POSIX paths into DOS paths. So we have to retain the level of normalization done by normalize_path_copy() to convert the bash-mangled DOS paths (which contain backslashes) into paths that use forward slashes. Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Jeff King <peff@peff.net>
2012-10-28 16:16:25 +00:00
if (len == 1 && ceil[0] == '/')
len = 0; /* root matches anything, with length 0 */
else if (!strncmp(path, ceil, len) && path[len] == '/')
; /* match of length len */
else
continue; /* no match */
longest_ancestor_length(): require prefix list entries to be normalized Move the responsibility for normalizing prefixes from longest_ancestor_length() to its callers. Use slightly different normalizations at the two callers: In setup_git_directory_gently_1(), use the old normalization, which ignores paths that are not usable. In the next commit we will change this caller to also resolve symlinks in the paths from GIT_CEILING_DIRECTORIES as part of the normalization. In "test-path-utils longest_ancestor_length", use the old normalization, but die() if any paths are unusable. Also change t0060 to only pass normalized paths to the test program (no empty entries or non-absolute paths, strip trailing slashes from the paths, and remove tests that thereby become redundant). The point of this change is to reduce the scope of the ancestor_length tests in t0060 from testing normalization+longest_prefix to testing only mostly longest_prefix. This is necessary because when setup_git_directory_gently_1() starts resolving symlinks as part of its normalization, it will not be reasonable to do the same in the test suite, because that would make the test results depend on the contents of the root directory of the filesystem on which the test is run. HOWEVER: under Windows, bash mangles arguments that look like absolute POSIX paths into DOS paths. So we have to retain the level of normalization done by normalize_path_copy() to convert the bash-mangled DOS paths (which contain backslashes) into paths that use forward slashes. Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu> Signed-off-by: Jeff King <peff@peff.net>
2012-10-28 16:16:25 +00:00
if (len > max_len)
max_len = len;
}
return max_len;
}
/* strip arbitrary amount of directory separators at end of path */
static inline int chomp_trailing_dir_sep(const char *path, int len)
{
while (len && is_dir_sep(path[len - 1]))
len--;
return len;
}
/*
* If path ends with suffix (complete path components), returns the offset of
* the last character in the path before the suffix (sans trailing directory
* separators), and -1 otherwise.
*/
static ssize_t stripped_path_suffix_offset(const char *path, const char *suffix)
{
int path_len = strlen(path), suffix_len = strlen(suffix);
while (suffix_len) {
if (!path_len)
return -1;
if (is_dir_sep(path[path_len - 1])) {
if (!is_dir_sep(suffix[suffix_len - 1]))
return -1;
path_len = chomp_trailing_dir_sep(path, path_len);
suffix_len = chomp_trailing_dir_sep(suffix, suffix_len);
}
else if (path[--path_len] != suffix[--suffix_len])
return -1;
}
if (path_len && !is_dir_sep(path[path_len - 1]))
return -1;
return chomp_trailing_dir_sep(path, path_len);
}
/*
* Returns true if the path ends with components, considering only complete path
* components, and false otherwise.
*/
int ends_with_path_components(const char *path, const char *components)
{
return stripped_path_suffix_offset(path, components) != -1;
}
/*
* If path ends with suffix (complete path components), returns the
* part before suffix (sans trailing directory separators).
* Otherwise returns NULL.
*/
char *strip_path_suffix(const char *path, const char *suffix)
{
ssize_t offset = stripped_path_suffix_offset(path, suffix);
return offset == -1 ? NULL : xstrndup(path, offset);
}
int daemon_avoid_alias(const char *p)
{
int sl, ndot;
/*
* This resurrects the belts and suspenders paranoia check by HPA
* done in <435560F7.4080006@zytor.com> thread, now enter_repo()
* does not do getcwd() based path canonicalization.
*
* sl becomes true immediately after seeing '/' and continues to
* be true as long as dots continue after that without intervening
* non-dot character.
*/
if (!p || (*p != '/' && *p != '~'))
return -1;
sl = 1; ndot = 0;
p++;
while (1) {
char ch = *p++;
if (sl) {
if (ch == '.')
ndot++;
else if (ch == '/') {
if (ndot < 3)
/* reject //, /./ and /../ */
return -1;
ndot = 0;
}
else if (ch == 0) {
if (0 < ndot && ndot < 3)
/* reject /.$ and /..$ */
return -1;
return 0;
}
else
sl = ndot = 0;
}
else if (ch == 0)
return 0;
else if (ch == '/') {
sl = 1;
ndot = 0;
}
}
}
path: add is_ntfs_dotgit() helper We do not allow paths with a ".git" component to be added to the index, as that would mean repository contents could overwrite our repository files. However, asking "is this path the same as .git" is not as simple as strcmp() on some filesystems. On NTFS (and FAT32), there exist so-called "short names" for backwards-compatibility: 8.3 compliant names that refer to the same files as their long names. As ".git" is not an 8.3 compliant name, a short name is generated automatically, typically "git~1". Depending on the Windows version, any combination of trailing spaces and periods are ignored, too, so that both "git~1." and ".git." still refer to the Git directory. The reason is that 8.3 stores file names shorter than 8 characters with trailing spaces. So literally, it does not matter for the short name whether it is padded with spaces or whether it is shorter than 8 characters, it is considered to be the exact same. The period is the separator between file name and file extension, and again, an empty extension consists just of spaces in 8.3 format. So technically, we would need only take care of the equivalent of this regex: (\.git {0,4}|git~1 {0,3})\. {0,3} However, there are indications that at least some Windows versions might be more lenient and accept arbitrary combinations of trailing spaces and periods and strip them out. So we're playing it real safe here. Besides, there can be little doubt about the intention behind using file names matching even the more lenient pattern specified above, therefore we should be fine with disallowing such patterns. Extra care is taken to catch names such as '.\\.git\\booh' because the backslash is marked as a directory separator only on Windows, and we want to use this new helper function also in fsck on other platforms. A big thank you goes to Ed Thomson and an unnamed Microsoft engineer for the detailed analysis performed to come up with the corresponding fixes for libgit2. This commit adds a function to detect whether a given file name can refer to the Git directory by mistake. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-12-16 22:31:03 +00:00
static int only_spaces_and_periods(const char *path, size_t len, size_t skip)
{
if (len < skip)
return 0;
len -= skip;
path += skip;
while (len-- > 0) {
char c = *(path++);
if (c != ' ' && c != '.')
return 0;
}
return 1;
}
int is_ntfs_dotgit(const char *name)
{
size_t len;
path: add is_ntfs_dotgit() helper We do not allow paths with a ".git" component to be added to the index, as that would mean repository contents could overwrite our repository files. However, asking "is this path the same as .git" is not as simple as strcmp() on some filesystems. On NTFS (and FAT32), there exist so-called "short names" for backwards-compatibility: 8.3 compliant names that refer to the same files as their long names. As ".git" is not an 8.3 compliant name, a short name is generated automatically, typically "git~1". Depending on the Windows version, any combination of trailing spaces and periods are ignored, too, so that both "git~1." and ".git." still refer to the Git directory. The reason is that 8.3 stores file names shorter than 8 characters with trailing spaces. So literally, it does not matter for the short name whether it is padded with spaces or whether it is shorter than 8 characters, it is considered to be the exact same. The period is the separator between file name and file extension, and again, an empty extension consists just of spaces in 8.3 format. So technically, we would need only take care of the equivalent of this regex: (\.git {0,4}|git~1 {0,3})\. {0,3} However, there are indications that at least some Windows versions might be more lenient and accept arbitrary combinations of trailing spaces and periods and strip them out. So we're playing it real safe here. Besides, there can be little doubt about the intention behind using file names matching even the more lenient pattern specified above, therefore we should be fine with disallowing such patterns. Extra care is taken to catch names such as '.\\.git\\booh' because the backslash is marked as a directory separator only on Windows, and we want to use this new helper function also in fsck on other platforms. A big thank you goes to Ed Thomson and an unnamed Microsoft engineer for the detailed analysis performed to come up with the corresponding fixes for libgit2. This commit adds a function to detect whether a given file name can refer to the Git directory by mistake. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-12-16 22:31:03 +00:00
for (len = 0; ; len++)
if (!name[len] || name[len] == '\\' || is_dir_sep(name[len])) {
if (only_spaces_and_periods(name, len, 4) &&
!strncasecmp(name, ".git", 4))
return 1;
if (only_spaces_and_periods(name, len, 5) &&
!strncasecmp(name, "git~1", 5))
return 1;
if (name[len] != '\\')
return 0;
name += len + 1;
len = -1;
}
}
2015-04-21 04:06:27 +00:00
is_ntfs_dotgit: match other .git files When we started to catch NTFS short names that clash with .git, we only looked for GIT~1. This is sufficient because we only ever clone into an empty directory, so .git is guaranteed to be the first subdirectory or file in that directory. However, even with a fresh clone, .gitmodules is *not* necessarily the first file to be written that would want the NTFS short name GITMOD~1: a malicious repository can add .gitmodul0000 and friends, which sorts before `.gitmodules` and is therefore checked out *first*. For that reason, we have to test not only for ~1 short names, but for others, too. It's hard to just adapt the existing checks in is_ntfs_dotgit(): since Windows 2000 (i.e., in all Windows versions still supported by Git), NTFS short names are only generated in the <prefix>~<number> form up to number 4. After that, a *different* prefix is used, calculated from the long file name using an undocumented, but stable algorithm. For example, the short name of .gitmodules would be GITMOD~1, but if it is taken, and all of ~2, ~3 and ~4 are taken, too, the short name GI7EBA~1 will be used. From there, collisions are handled by incrementing the number, shortening the prefix as needed (until ~9999999 is reached, in which case NTFS will not allow the file to be created). We'd also want to handle .gitignore and .gitattributes, which suffer from a similar problem, using the fall-back short names GI250A~1 and GI7D29~1, respectively. To accommodate for that, we could reimplement the hashing algorithm, but it is just safer and simpler to provide the known prefixes. This algorithm has been reverse-engineered and described at https://usn.pw/blog/gen/2015/06/09/filenames/, which is defunct but still available via https://web.archive.org/. These can be recomputed by running the following Perl script: -- snip -- use warnings; use strict; sub compute_short_name_hash ($) { my $checksum = 0; foreach (split('', $_[0])) { $checksum = ($checksum * 0x25 + ord($_)) & 0xffff; } $checksum = ($checksum * 314159269) & 0xffffffff; $checksum = 1 + (~$checksum & 0x7fffffff) if ($checksum & 0x80000000); $checksum -= (($checksum * 1152921497) >> 60) * 1000000007; return scalar reverse sprintf("%x", $checksum & 0xffff); } print compute_short_name_hash($ARGV[0]); -- snap -- E.g., running that with the argument ".gitignore" will result in "250a" (which then becomes "gi250a" in the code). Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Jeff King <peff@peff.net>
2018-05-11 14:03:54 +00:00
static int is_ntfs_dot_generic(const char *name,
const char *dotgit_name,
size_t len,
const char *dotgit_ntfs_shortname_prefix)
{
int saw_tilde;
size_t i;
if ((name[0] == '.' && !strncasecmp(name + 1, dotgit_name, len))) {
i = len + 1;
only_spaces_and_periods:
for (;;) {
char c = name[i++];
if (!c)
return 1;
if (c != ' ' && c != '.')
return 0;
}
}
/*
* Is it a regular NTFS short name, i.e. shortened to 6 characters,
* followed by ~1, ... ~4?
*/
if (!strncasecmp(name, dotgit_name, 6) && name[6] == '~' &&
name[7] >= '1' && name[7] <= '4') {
i = 8;
goto only_spaces_and_periods;
}
/*
* Is it a fall-back NTFS short name (for details, see
* https://en.wikipedia.org/wiki/8.3_filename?
*/
for (i = 0, saw_tilde = 0; i < 8; i++)
if (name[i] == '\0')
return 0;
else if (saw_tilde) {
if (name[i] < '0' || name[i] > '9')
return 0;
} else if (name[i] == '~') {
if (name[++i] < '1' || name[i] > '9')
return 0;
saw_tilde = 1;
} else if (i >= 6)
return 0;
else if (name[i] & 0x80) {
is_ntfs_dotgit: match other .git files When we started to catch NTFS short names that clash with .git, we only looked for GIT~1. This is sufficient because we only ever clone into an empty directory, so .git is guaranteed to be the first subdirectory or file in that directory. However, even with a fresh clone, .gitmodules is *not* necessarily the first file to be written that would want the NTFS short name GITMOD~1: a malicious repository can add .gitmodul0000 and friends, which sorts before `.gitmodules` and is therefore checked out *first*. For that reason, we have to test not only for ~1 short names, but for others, too. It's hard to just adapt the existing checks in is_ntfs_dotgit(): since Windows 2000 (i.e., in all Windows versions still supported by Git), NTFS short names are only generated in the <prefix>~<number> form up to number 4. After that, a *different* prefix is used, calculated from the long file name using an undocumented, but stable algorithm. For example, the short name of .gitmodules would be GITMOD~1, but if it is taken, and all of ~2, ~3 and ~4 are taken, too, the short name GI7EBA~1 will be used. From there, collisions are handled by incrementing the number, shortening the prefix as needed (until ~9999999 is reached, in which case NTFS will not allow the file to be created). We'd also want to handle .gitignore and .gitattributes, which suffer from a similar problem, using the fall-back short names GI250A~1 and GI7D29~1, respectively. To accommodate for that, we could reimplement the hashing algorithm, but it is just safer and simpler to provide the known prefixes. This algorithm has been reverse-engineered and described at https://usn.pw/blog/gen/2015/06/09/filenames/, which is defunct but still available via https://web.archive.org/. These can be recomputed by running the following Perl script: -- snip -- use warnings; use strict; sub compute_short_name_hash ($) { my $checksum = 0; foreach (split('', $_[0])) { $checksum = ($checksum * 0x25 + ord($_)) & 0xffff; } $checksum = ($checksum * 314159269) & 0xffffffff; $checksum = 1 + (~$checksum & 0x7fffffff) if ($checksum & 0x80000000); $checksum -= (($checksum * 1152921497) >> 60) * 1000000007; return scalar reverse sprintf("%x", $checksum & 0xffff); } print compute_short_name_hash($ARGV[0]); -- snap -- E.g., running that with the argument ".gitignore" will result in "250a" (which then becomes "gi250a" in the code). Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Jeff King <peff@peff.net>
2018-05-11 14:03:54 +00:00
/*
* We know our needles contain only ASCII, so we clamp
* here to make the results of tolower() sane.
*/
return 0;
} else if (tolower(name[i]) != dotgit_ntfs_shortname_prefix[i])
return 0;
goto only_spaces_and_periods;
}
/*
* Inline helper to make sure compiler resolves strlen() on literals at
* compile time.
*/
static inline int is_ntfs_dot_str(const char *name, const char *dotgit_name,
const char *dotgit_ntfs_shortname_prefix)
{
return is_ntfs_dot_generic(name, dotgit_name, strlen(dotgit_name),
dotgit_ntfs_shortname_prefix);
}
int is_ntfs_dotgitmodules(const char *name)
{
return is_ntfs_dot_str(name, "gitmodules", "gi7eba");
}
int is_ntfs_dotgitignore(const char *name)
{
return is_ntfs_dot_str(name, "gitignore", "gi250a");
}
int is_ntfs_dotgitattributes(const char *name)
{
return is_ntfs_dot_str(name, "gitattributes", "gi7d29");
}
int looks_like_command_line_option(const char *str)
{
return str && str[0] == '-';
}
2015-04-21 04:06:27 +00:00
char *xdg_config_home(const char *filename)
{
const char *home, *config_home;
assert(filename);
config_home = getenv("XDG_CONFIG_HOME");
if (config_home && *config_home)
return mkpathdup("%s/git/%s", config_home, filename);
home = getenv("HOME");
if (home)
return mkpathdup("%s/.config/git/%s", home, filename);
return NULL;
}
memoize common git-path "constant" files One of the most common uses of git_path() is to pass a constant, like git_path("MERGE_MSG"). This has two drawbacks: 1. The return value is a static buffer, and the lifetime is dependent on other calls to git_path, etc. 2. There's no compile-time checking of the pathname. This is OK for a one-off (after all, we have to spell it correctly at least once), but many of these constant strings appear throughout the code. This patch introduces a series of functions to "memoize" these strings, which are essentially globals for the lifetime of the program. We compute the value once, take ownership of the buffer, and return the cached value for subsequent calls. cache.h provides a helper macro for defining these functions as one-liners, and defines a few common ones for global use. Using a macro is a little bit gross, but it does nicely document the purpose of the functions. If we need to touch them all later (e.g., because we learned how to change the git_dir variable at runtime, and need to invalidate all of the stored values), it will be much easier to have the complete list. Note that the shared-global functions have separate, manual declarations. We could do something clever with the macros (e.g., expand it to a declaration in some places, and a declaration _and_ a definition in path.c). But there aren't that many, and it's probably better to stay away from too-magical macros. Likewise, if we abandon the C preprocessor in favor of generating these with a script, we could get much fancier. E.g., normalizing "FOO/BAR-BAZ" into "git_path_foo_bar_baz". But the small amount of saved typing is probably not worth the resulting confusion to readers who want to grep for the function's definition. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-08-10 09:38:57 +00:00
char *xdg_cache_home(const char *filename)
{
const char *home, *cache_home;
assert(filename);
cache_home = getenv("XDG_CACHE_HOME");
if (cache_home && *cache_home)
return mkpathdup("%s/git/%s", cache_home, filename);
home = getenv("HOME");
if (home)
return mkpathdup("%s/.cache/git/%s", home, filename);
return NULL;
}
REPO_GIT_PATH_FUNC(cherry_pick_head, "CHERRY_PICK_HEAD")
REPO_GIT_PATH_FUNC(revert_head, "REVERT_HEAD")
REPO_GIT_PATH_FUNC(squash_msg, "SQUASH_MSG")
REPO_GIT_PATH_FUNC(merge_msg, "MERGE_MSG")
REPO_GIT_PATH_FUNC(merge_rr, "MERGE_RR")
REPO_GIT_PATH_FUNC(merge_mode, "MERGE_MODE")
REPO_GIT_PATH_FUNC(merge_head, "MERGE_HEAD")
REPO_GIT_PATH_FUNC(fetch_head, "FETCH_HEAD")
REPO_GIT_PATH_FUNC(shallow, "shallow")