int git_futils_mmap_ro_file(git_map *out, const char *path)
{
git_file fd = git_futils_open_ro(path);
git_off_t len;
int result;
if (fd < 0)
return fd;
if ((len = git_futils_filesize(fd)) < 0) {
result = -1;
goto out;
}
if (!git__is_sizet(len)) {
giterr_set(GITERR_OS, "file `%s` too large to mmap", path);
result = -1;
goto out;
}
result = git_futils_mmap_ro(out, fd, 0, (size_t)len);
out:
p_close(fd);
return result;
}
static int write_file_stream(
git_oid *oid, git_odb *odb, const char *path, git_off_t file_size)
{
int fd, error;
char buffer[4096];
git_odb_stream *stream = NULL;
ssize_t read_len = -1, written = 0;
if ((error = git_odb_open_wstream(
&stream, odb, (size_t)file_size, GIT_OBJ_BLOB)) < 0)
return error;
if ((fd = git_futils_open_ro(path)) < 0) {
git_odb_stream_free(stream);
return -1;
}
while (!error && (read_len = p_read(fd, buffer, sizeof(buffer))) > 0) {
error = git_odb_stream_write(stream, buffer, read_len);
written += read_len;
}
p_close(fd);
if (written != file_size || read_len < 0) {
giterr_set(GITERR_OS, "Failed to read file into stream");
error = -1;
}
if (!error)
error = git_odb_stream_finalize_write(oid, stream);
git_odb_stream_free(stream);
return error;
}
static int read_header_loose(git_rawobj *out, git_buf *loc)
{
unsigned char obj[1024];
int fd, obj_len, error;
assert(out && loc);
if (git_buf_oom(loc))
return -1;
out->data = NULL;
if ((error = fd = git_futils_open_ro(loc->ptr)) < 0 ||
(error = obj_len = p_read(fd, obj, sizeof(obj))) < 0)
goto done;
if (!is_zlib_compressed_data(obj, (size_t)obj_len))
error = read_header_loose_packlike(out, obj, (size_t)obj_len);
else
error = read_header_loose_standard(out, obj, (size_t)obj_len);
if (!error && !git_object_typeisloose(out->type)) {
giterr_set(GITERR_ZLIB, "failed to read loose object header");
error = -1;
goto done;
}
done:
if (fd >= 0)
p_close(fd);
return error;
}
static int diff_file_content_load_workdir_file(
git_diff_file_content *fc,
git_buf *path,
git_diff_options *diff_opts)
{
int error = 0;
git_filter_list *fl = NULL;
git_file fd = git_futils_open_ro(git_buf_cstr(path));
git_buf raw = GIT_BUF_INIT;
if (fd < 0)
return fd;
if (!fc->file->size &&
!(fc->file->size = git_futils_filesize(fd)))
goto cleanup;
if ((diff_opts->flags & GIT_DIFF_SHOW_BINARY) == 0 &&
diff_file_content_binary_by_size(fc))
goto cleanup;
if ((error = git_filter_list_load(
&fl, fc->repo, NULL, fc->file->path,
GIT_FILTER_TO_ODB, GIT_FILTER_ALLOW_UNSAFE)) < 0)
goto cleanup;
/* if there are no filters, try to mmap the file */
if (fl == NULL) {
if (!(error = git_futils_mmap_ro(
&fc->map, fd, 0, (size_t)fc->file->size))) {
fc->flags |= GIT_DIFF_FLAG__UNMAP_DATA;
goto cleanup;
}
/* if mmap failed, fall through to try readbuffer below */
giterr_clear();
}
if (!(error = git_futils_readbuffer_fd(&raw, fd, (size_t)fc->file->size))) {
git_buf out = GIT_BUF_INIT;
error = git_filter_list_apply_to_data(&out, fl, &raw);
if (out.ptr != raw.ptr)
git_buf_dispose(&raw);
if (!error) {
fc->map.len = out.size;
fc->map.data = out.ptr;
fc->flags |= GIT_DIFF_FLAG__FREE_DATA;
}
}
cleanup:
git_filter_list_free(fl);
p_close(fd);
return error;
}
static int diff_file_content_load_workdir_file(
git_diff_file_content *fc, git_buf *path)
{
int error = 0;
git_vector filters = GIT_VECTOR_INIT;
git_buf raw = GIT_BUF_INIT, filtered = GIT_BUF_INIT;
git_file fd = git_futils_open_ro(git_buf_cstr(path));
if (fd < 0)
return fd;
if (!fc->file->size &&
!(fc->file->size = git_futils_filesize(fd)))
goto cleanup;
if (diff_file_content_binary_by_size(fc))
goto cleanup;
if ((error = git_filters_load(
&filters, fc->repo, fc->file->path, GIT_FILTER_TO_ODB)) < 0)
goto cleanup;
/* error >= is a filter count */
if (error == 0) {
if (!(error = git_futils_mmap_ro(
&fc->map, fd, 0, (size_t)fc->file->size)))
fc->flags |= GIT_DIFF_FLAG__UNMAP_DATA;
else /* fall through to try readbuffer below */
giterr_clear();
}
if (error != 0) {
error = git_futils_readbuffer_fd(&raw, fd, (size_t)fc->file->size);
if (error < 0)
goto cleanup;
if (!filters.length)
git_buf_swap(&filtered, &raw);
else
error = git_filters_apply(&filtered, &raw, &filters);
if (!error) {
fc->map.len = git_buf_len(&filtered);
fc->map.data = git_buf_detach(&filtered);
fc->flags |= GIT_DIFF_FLAG__FREE_DATA;
}
git_buf_free(&raw);
git_buf_free(&filtered);
}
cleanup:
git_filters_free(&filters);
p_close(fd);
return error;
}
int git_futils_readbuffer_updated(
git_buf *buf, const char *path, time_t *mtime, size_t *size, int *updated)
{
git_file fd;
struct stat st;
bool changed = false;
assert(buf && path && *path);
if (updated != NULL)
*updated = 0;
if ((fd = git_futils_open_ro(path)) < 0)
return fd;
if (p_fstat(fd, &st) < 0 || S_ISDIR(st.st_mode) || !git__is_sizet(st.st_size+1)) {
p_close(fd);
giterr_set(GITERR_OS, "Invalid regular file stat for '%s'", path);
return -1;
}
/*
* If we were given a time and/or a size, we only want to read the file
* if it has been modified.
*/
if (size && *size != (size_t)st.st_size)
changed = true;
if (mtime && *mtime != st.st_mtime)
changed = true;
if (!size && !mtime)
changed = true;
if (!changed) {
p_close(fd);
return 0;
}
if (mtime != NULL)
*mtime = st.st_mtime;
if (size != NULL)
*size = (size_t)st.st_size;
if (git_futils_readbuffer_fd(buf, fd, (size_t)st.st_size) < 0) {
p_close(fd);
return -1;
}
p_close(fd);
if (updated != NULL)
*updated = 1;
return 0;
}
int git_odb_hashfile(git_oid *out, const char *path, git_otype type)
{
git_off_t size;
int result, fd = git_futils_open_ro(path);
if (fd < 0)
return fd;
if ((size = git_futils_filesize(fd)) < 0 || !git__is_sizet(size)) {
giterr_set(GITERR_OS, "File size overflow for 32-bit systems");
p_close(fd);
return -1;
}
result = git_odb__hashfd(out, fd, (size_t)size, type);
p_close(fd);
return result;
}
int git_odb__hashlink(git_oid *out, const char *path)
{
struct stat st;
git_off_t size;
int result;
if (git_path_lstat(path, &st) < 0)
return -1;
size = st.st_size;
if (!git__is_sizet(size)) {
giterr_set(GITERR_OS, "File size overflow for 32-bit systems");
return -1;
}
if (S_ISLNK(st.st_mode)) {
char *link_data;
ssize_t read_len;
link_data = (char*) git__malloc((size_t)(size + 1));
GITERR_CHECK_ALLOC(link_data);
read_len = p_readlink(path, link_data, (size_t)size);
link_data[size] = '\0';
if (read_len != (ssize_t)size) {
giterr_set(GITERR_OS, "Failed to read symlink data for '%s'", path);
git__free(link_data);
return -1;
}
result = git_odb_hash(out, link_data, (size_t)size, GIT_OBJ_BLOB);
git__free(link_data);
} else {
int fd = git_futils_open_ro(path);
if (fd < 0)
return -1;
result = git_odb__hashfd(out, fd, (size_t)size, GIT_OBJ_BLOB);
p_close(fd);
}
return result;
}
int git_hashsig_create_fromfile(
git_hashsig **out,
const char *path,
git_hashsig_option_t opts)
{
char buf[4096];
ssize_t buflen = 0;
int error = 0, fd;
hashsig_in_progress prog = HASHSIG_IN_PROGRESS_INIT;
git_hashsig *sig = hashsig_alloc(opts);
GITERR_CHECK_ALLOC(sig);
if ((fd = git_futils_open_ro(path)) < 0) {
git__free(sig);
return fd;
}
while (!error) {
if ((buflen = p_read(fd, buf, sizeof(buf))) <= 0) {
if ((error = (int)buflen) < 0)
giterr_set(GITERR_OS,
"Read error on '%s' calculating similarity hashes", path);
break;
}
error = hashsig_add_hashes(sig, buf, buflen, &prog);
}
p_close(fd);
if (!error)
error = hashsig_finalize_hashes(sig);
if (!error)
*out = sig;
else
git_hashsig_free(sig);
return error;
}
static int write_file_stream(
git_oid *oid, git_odb *odb, const char *path, git_off_t file_size)
{
int fd, error;
char buffer[4096];
git_odb_stream *stream = NULL;
if ((error = git_odb_open_wstream(
&stream, odb, (size_t)file_size, GIT_OBJ_BLOB)) < 0)
return error;
if ((fd = git_futils_open_ro(path)) < 0) {
stream->free(stream);
return -1;
}
while (!error && file_size > 0) {
ssize_t read_len = p_read(fd, buffer, sizeof(buffer));
if (read_len < 0) {
giterr_set(
GITERR_OS, "Failed to create blob. Can't read whole file");
error = -1;
}
else if (!(error = stream->write(stream, buffer, read_len)))
file_size -= read_len;
}
p_close(fd);
if (!error)
error = stream->finalize_write(oid, stream);
stream->free(stream);
return error;
}
int git_futils_readbuffer_updated(
git_buf *out, const char *path, git_oid *checksum, int *updated)
{
int error;
git_file fd;
struct stat st;
git_buf buf = GIT_BUF_INIT;
git_oid checksum_new;
assert(out && path && *path);
if (updated != NULL)
*updated = 0;
if (p_stat(path, &st) < 0)
return git_path_set_error(errno, path, "stat");
if (S_ISDIR(st.st_mode)) {
giterr_set(GITERR_INVALID, "requested file is a directory");
return GIT_ENOTFOUND;
}
if (!git__is_sizet(st.st_size+1)) {
giterr_set(GITERR_OS, "invalid regular file stat for '%s'", path);
return -1;
}
if ((fd = git_futils_open_ro(path)) < 0)
return fd;
if (git_futils_readbuffer_fd(&buf, fd, (size_t)st.st_size) < 0) {
p_close(fd);
return -1;
}
p_close(fd);
if (checksum) {
if ((error = git_hash_buf(&checksum_new, buf.ptr, buf.size)) < 0) {
git_buf_free(&buf);
return error;
}
/*
* If we were given a checksum, we only want to use it if it's different
*/
if (!git_oid__cmp(checksum, &checksum_new)) {
git_buf_free(&buf);
if (updated)
*updated = 0;
return 0;
}
git_oid_cpy(checksum, &checksum_new);
}
/*
* If we're here, the file did change, or the user didn't have an old version
*/
if (updated != NULL)
*updated = 1;
git_buf_swap(out, &buf);
git_buf_free(&buf);
return 0;
}
int git_attr_file__load(
git_attr_file **out,
git_repository *repo,
git_attr_file_entry *entry,
git_attr_file_source source,
git_attr_file_parser parser)
{
int error = 0;
git_blob *blob = NULL;
git_buf content = GIT_BUF_INIT;
const char *data = NULL;
git_attr_file *file;
struct stat st;
*out = NULL;
switch (source) {
case GIT_ATTR_FILE__IN_MEMORY:
/* in-memory attribute file doesn't need data */
break;
case GIT_ATTR_FILE__FROM_INDEX: {
git_oid id;
if ((error = attr_file_oid_from_index(&id, repo, entry->path)) < 0 ||
(error = git_blob_lookup(&blob, repo, &id)) < 0)
return error;
data = git_blob_rawcontent(blob);
break;
}
case GIT_ATTR_FILE__FROM_FILE: {
int fd;
if (p_stat(entry->fullpath, &st) < 0)
return git_path_set_error(errno, entry->fullpath, "stat");
if (S_ISDIR(st.st_mode))
return GIT_ENOTFOUND;
/* For open or read errors, return ENOTFOUND to skip item */
/* TODO: issue warning when warning API is available */
if ((fd = git_futils_open_ro(entry->fullpath)) < 0)
return GIT_ENOTFOUND;
error = git_futils_readbuffer_fd(&content, fd, (size_t)st.st_size);
p_close(fd);
if (error < 0)
return GIT_ENOTFOUND;
data = content.ptr;
break;
}
default:
giterr_set(GITERR_INVALID, "Unknown file source %d", source);
return -1;
}
if ((error = git_attr_file__new(&file, entry, source)) < 0)
goto cleanup;
if (parser && (error = parser(repo, file, data)) < 0) {
git_attr_file__free(file);
goto cleanup;
}
/* write cache breaker */
if (source == GIT_ATTR_FILE__FROM_INDEX)
git_oid_cpy(&file->cache_data.oid, git_blob_id(blob));
else if (source == GIT_ATTR_FILE__FROM_FILE)
git_futils_filestamp_set_from_stat(&file->cache_data.stamp, &st);
/* else always cacheable */
*out = file;
cleanup:
git_blob_free(blob);
git_buf_free(&content);
return error;
}
int git_attr_file__load(
git_attr_file **out,
git_repository *repo,
git_attr_session *attr_session,
git_attr_file_entry *entry,
git_attr_file_source source,
git_attr_file_parser parser)
{
int error = 0;
git_blob *blob = NULL;
git_buf content = GIT_BUF_INIT;
git_attr_file *file;
struct stat st;
bool nonexistent = false;
*out = NULL;
switch (source) {
case GIT_ATTR_FILE__IN_MEMORY:
/* in-memory attribute file doesn't need data */
break;
case GIT_ATTR_FILE__FROM_INDEX: {
git_oid id;
if ((error = attr_file_oid_from_index(&id, repo, entry->path)) < 0 ||
(error = git_blob_lookup(&blob, repo, &id)) < 0)
return error;
/* Do not assume that data straight from the ODB is NULL-terminated;
* copy the contents of a file to a buffer to work on */
git_buf_put(&content, git_blob_rawcontent(blob), git_blob_rawsize(blob));
break;
}
case GIT_ATTR_FILE__FROM_FILE: {
int fd = -1;
/* For open or read errors, pretend that we got ENOTFOUND. */
/* TODO: issue warning when warning API is available */
if (p_stat(entry->fullpath, &st) < 0 ||
S_ISDIR(st.st_mode) ||
(fd = git_futils_open_ro(entry->fullpath)) < 0 ||
(error = git_futils_readbuffer_fd(&content, fd, (size_t)st.st_size)) < 0)
nonexistent = true;
if (fd >= 0)
p_close(fd);
break;
}
default:
giterr_set(GITERR_INVALID, "Unknown file source %d", source);
return -1;
}
if ((error = git_attr_file__new(&file, entry, source)) < 0)
goto cleanup;
/* store the key of the attr_reader; don't bother with cache
* invalidation during the same attr reader session.
*/
if (attr_session)
file->session_key = attr_session->key;
if (parser && (error = parser(repo, file, git_buf_cstr(&content))) < 0) {
git_attr_file__free(file);
goto cleanup;
}
/* write cache breakers */
if (nonexistent)
file->nonexistent = 1;
else if (source == GIT_ATTR_FILE__FROM_INDEX)
git_oid_cpy(&file->cache_data.oid, git_blob_id(blob));
else if (source == GIT_ATTR_FILE__FROM_FILE)
git_futils_filestamp_set_from_stat(&file->cache_data.stamp, &st);
/* else always cacheable */
*out = file;
cleanup:
git_blob_free(blob);
git_buf_free(&content);
return error;
}
int git_futils_readbuffer_updated(git_buf *buf, const char *path, time_t *mtime, int *updated)
{
git_file fd;
size_t len;
struct stat st;
assert(buf && path && *path);
if (updated != NULL)
*updated = 0;
if ((fd = git_futils_open_ro(path)) < 0)
return fd;
if (p_fstat(fd, &st) < 0 || S_ISDIR(st.st_mode) || !git__is_sizet(st.st_size+1)) {
p_close(fd);
giterr_set(GITERR_OS, "Invalid regular file stat for '%s'", path);
return -1;
}
/*
* If we were given a time, we only want to read the file if it
* has been modified.
*/
if (mtime != NULL && *mtime >= st.st_mtime) {
p_close(fd);
return 0;
}
if (mtime != NULL)
*mtime = st.st_mtime;
len = (size_t) st.st_size;
git_buf_clear(buf);
if (git_buf_grow(buf, len + 1) < 0) {
p_close(fd);
return -1;
}
buf->ptr[len] = '\0';
while (len > 0) {
ssize_t read_size = p_read(fd, buf->ptr, len);
if (read_size < 0) {
p_close(fd);
giterr_set(GITERR_OS, "Failed to read descriptor for '%s'", path);
return -1;
}
len -= read_size;
buf->size += read_size;
}
p_close(fd);
if (updated != NULL)
*updated = 1;
return 0;
}