[pacman-dev] [PATCH] libalpm: Use archive_read_extract2
archive_read_extract() forces resolution of uid/gid to names when extracting the tarball. This can lead to wrong file ownership when using pacman with -r option and when uid/gid differ in the host and in the chroot. archive_read_extract2() uses uid's and gid's only. See also: https://lists.archlinux.org/pipermail/pacman-dev/2017-March/021912.html Signed-off-by: Armin K <krejzi@email.com> --- lib/libalpm/add.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) diff --git a/lib/libalpm/add.c b/lib/libalpm/add.c index c5bed817..0beed01c 100644 --- a/lib/libalpm/add.c +++ b/lib/libalpm/add.c @@ -110,6 +110,7 @@ static int perform_extraction(alpm_handle_t *handle, struct archive *archive, struct archive_entry *entry, const char *filename) { int ret; + struct archive *archive_writer; const int archive_flags = ARCHIVE_EXTRACT_OWNER | ARCHIVE_EXTRACT_PERM | ARCHIVE_EXTRACT_TIME | @@ -118,7 +119,20 @@ static int perform_extraction(alpm_handle_t *handle, struct archive *archive, archive_entry_set_pathname(entry, filename); - ret = archive_read_extract(archive, entry, archive_flags); + archive_writer = archive_write_disk_new(); + if (archive_writer == NULL) { + _alpm_log(handle, ALPM_LOG_ERROR, _("cannot allocate disk archive object")); + alpm_logaction(handle, ALPM_CALLER_PREFIX, + "error: cannot allocate disk archive object"); + return 1; + } + + archive_write_disk_set_options(archive_writer, archive_flags); + + ret = archive_read_extract2(archive, entry, archive_writer); + + archive_write_free(archive_writer); + if(ret == ARCHIVE_WARN && archive_errno(archive) != ENOSPC) { /* operation succeeded but a "non-critical" error was encountered */ _alpm_log(handle, ALPM_LOG_WARNING, _("warning given when extracting %s (%s)\n"), -- 2.12.0
On 07/03/17 06:15, Armin K wrote:
archive_read_extract() forces resolution of uid/gid to names when extracting the tarball. This can lead to wrong file ownership when using pacman with -r option and when uid/gid differ in the host and in the chroot.
archive_read_extract2() uses uid's and gid's only. See also:
https://lists.archlinux.org/pipermail/pacman-dev/2017-March/021912.html
Signed-off-by: Armin K <krejzi@email.com> ---
@Andrew: Is this patch OK to go in until you finish your work on group/user handling? A
lib/libalpm/add.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-)
diff --git a/lib/libalpm/add.c b/lib/libalpm/add.c index c5bed817..0beed01c 100644 --- a/lib/libalpm/add.c +++ b/lib/libalpm/add.c @@ -110,6 +110,7 @@ static int perform_extraction(alpm_handle_t *handle, struct archive *archive, struct archive_entry *entry, const char *filename) { int ret; + struct archive *archive_writer; const int archive_flags = ARCHIVE_EXTRACT_OWNER | ARCHIVE_EXTRACT_PERM | ARCHIVE_EXTRACT_TIME | @@ -118,7 +119,20 @@ static int perform_extraction(alpm_handle_t *handle, struct archive *archive,
archive_entry_set_pathname(entry, filename);
- ret = archive_read_extract(archive, entry, archive_flags); + archive_writer = archive_write_disk_new(); + if (archive_writer == NULL) { + _alpm_log(handle, ALPM_LOG_ERROR, _("cannot allocate disk archive object")); + alpm_logaction(handle, ALPM_CALLER_PREFIX, + "error: cannot allocate disk archive object"); + return 1; + } + + archive_write_disk_set_options(archive_writer, archive_flags); + + ret = archive_read_extract2(archive, entry, archive_writer); + + archive_write_free(archive_writer); + if(ret == ARCHIVE_WARN && archive_errno(archive) != ENOSPC) { /* operation succeeded but a "non-critical" error was encountered */ _alpm_log(handle, ALPM_LOG_WARNING, _("warning given when extracting %s (%s)\n"),
On 04/04/17 at 12:00pm, Allan McRae wrote:
On 07/03/17 06:15, Armin K wrote:
archive_read_extract() forces resolution of uid/gid to names when extracting the tarball. This can lead to wrong file ownership when using pacman with -r option and when uid/gid differ in the host and in the chroot.
archive_read_extract2() uses uid's and gid's only. See also:
https://lists.archlinux.org/pipermail/pacman-dev/2017-March/021912.html
Signed-off-by: Armin K <krejzi@email.com> ---
@Andrew: Is this patch OK to go in until you finish your work on group/user handling?
ACK. This would need to be done for the group/user handling work anyway. apg
participants (3)
-
Allan McRae
-
Andrew Gregory
-
Armin K