Fix #199

[JRF63]

Root cause is not properly following section 2.e of the cp spec. The bug itself is encountered if the file system lists D/D before D/a.

[jgarzik]

Here is the failure after testing this issue199 branch:

     Running tests/tree-tests.rs (target/release/deps/tree_tests-48a52f9cf418fb37)

running 111 tests
test cp::test_cp_dir_slash ... ok
test cp::test_cp_deref ... ok
test cp::test_cp_dir_vs_file ... ok
test cp::test_cp_childproof ... ok
test cp::test_cp_fail_perm ... FAILED
test cp::test_cp_into_self ... ok
test cp::test_cp_hl ... ok
test cp::test_cp_special_bits ... ignored
test cp::test_cp_i_2 ... ok
test cp::test_cp_i ... ok
test cp::test_cp_thru_dangling ... ok
test cp::test_cp_special_f ... ok
test cp::test_cp_r_vs_symlink ... ok
test link::link_already_exists ... ok
test link::link_create ... ok
test cp::test_cp_trailing_slash ... ok
test cp::test_cp_part_symlink ... ok
test ls::test_ls_empty_directory ... ok
test ls::test_ls_infloop ... ok
test ls::test_ls_file_type ... ok
test ls::test_ls_no_arg ... ok
test ls::test_ls_dangle ... ok
test ls::test_ls_recursive ... ok
test ls::test_ls_rt_1 ... ok
test ls::test_ls_m_option ... ok
test mkdir::test_create_single_directory ... ok
test ls::test_ls_inode ... ok
test mkdir::test_directory_already_exists ... ok
test mkdir::test_invalid_mode ... ok
test ls::test_ls_size_align ... ok
test mkdir::test_set_directory_mode ... ok
test cp::test_cp_proc_short_read ... ok
test mv::test_mv_atomic2 ... ok
test mv::test_mv_atomic ... ok
test cp::test_cp_same_file ... ok
test mv::test_mv_dir2dir ... ok
test mv::test_mv_dir_file ... ok
test mv::test_mv_hard_4 ... ok
test mv::test_mv_hardlink_case ... ignored
test mv::test_mv_hard_2 ... ok
test mv::test_mv_childproof ... ok
test mv::test_mv_force ... ok
test mv::test_mv_i_1 ... ok
test mv::test_mv_i_link_no ... ignored
test mv::test_mv_dup_source ... ok
test mv::test_mv_i_5 ... ok
test mv::test_mv_hard_link_1 ... ok
test mv::test_mv_i_2 ... ok
test mv::test_mv_into_self ... ok
test mv::test_mv_i_4 ... ok
test mv::test_mv_into_self_3 ... ok
test mv::test_mv_into_self_4 ... ok
test mv::test_mv_part_fail ... ok
test mv::test_mv_into_self_2 ... ok
test mv::test_mv_perm_1 ... ok
test mv::test_mv_part_hardlink ... ok
test mv::test_mv_partition_perm ... ok
test mv::test_mv_sticky_to_xpart ... ignored
test mv::test_mv_symlink_onto_hardlink_to_self ... ok
test mv::test_mv_special ... ok
test mv::test_mv_symlink_onto_hardlink ... ok
test mv::test_mv_to_symlink ... ok
test mv::test_mv_part_symlink ... ok
test readlink::test_readlink_non_existent_file ... ok
test mv::test_mv_part_rename ... ok
test readlink::test_readlink_not_a_symlink ... ok
test mv::test_mv_trailing_slash ... ok
test readlink::test_readlink_valid_symlink ... ok
test readlink::test_readlink_valid_symlink_no_newline ... ok
test rm::test_rm_cycle ... ok
test rm::test_rm_dir_no_w ... ok
test rm::test_rm_dir_nonrecur ... ok
test rm::test_rm_dot_rel ... ok
test rm::test_rm_empty_inacc ... ok
test rm::test_rm_f_1 ... ok
test rm::test_rm_fail_2eperm ... ignored
test rm::test_rm_empty_name ... ok
test rm::test_rm_i_1 ... ok
test rm::test_rm_fail_eacces ... ok
test rm::test_rm_i_no_r ... ok
test rm::test_rm_ignorable ... ok
test rm::test_rm_ir_1 ... ok
test rm::test_rm_inaccessible ... ok
test rm::test_rm_no_give_up ... ignored
test rm::test_rm_dangling_symlink ... ok
test rm::test_rm_r_4 ... ok
test rm::test_rm_r_3 ... ok
test rm::test_rm_r_root ... ok
test rm::test_rm_rm1 ... ok
test rm::test_rm_rm2 ... ok
test rm::test_rm_rm3 ... ok
test rm::test_rm_rm4 ... ok
test rm::test_rm_rm5 ... ok
test rm::test_rm_sunos_1 ... ok
test rm::test_rm_unread2 ... ok
test rm::test_rm_readdir_bug ... ok
test rm::test_rm_unread3 ... ok
test rmdir::rmdir_remove_directory_with_parents ... ok
test rmdir::rmdir_remove_existing_directory ... ok
test rm::test_rm_unreadable ... ok
test rmdir::rmdir_remove_non_empty_directory ... ok
test rmdir::rmdir_remove_non_existent_directory ... ok
test unlink::unlink_remove_directory ... ok
test unlink::unlink_remove_existing_file ... ok
test unlink::unlink_remove_non_existing_file ... ok
test rm::test_rm_deep_2 ... ok
test mv::test_mv_leak_fd ... ok
test rm::test_rm_hash ... ok
test ls::test_ls_time ... ok
test cp::test_cp_preserve_slink_time ... ok
test rm::test_rm_isatty ... ok

failures:

---- cp::test_cp_fail_perm stdout ----
thread 'cp::test_cp_fail_perm' panicked at plib/src/testing.rs:68:5:
assertion `left == right` failed
  left: "cp: Permission denied\n"
 right: "cp: cannot open '/home/jgarzik/repo/posixutils-rs/target/tmp/test_cp_fail_perm/D/a' for reading: Permission denied\n"
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace


failures:
    cp::test_cp_fail_perm

test result: FAILED. 104 passed; 1 failed; 6 ignored; 0 measured; 0 filtered out; finished in 3.08s

error: test failed, to rerun pass `-p posixutils-tree --test tree-tests`

[jgarzik]

This fails on Linux in a new way:

---- cp::test_cp_fail_perm stdout ----
thread 'cp::test_cp_fail_perm' panicked at tree/tests/cp/mod.rs:348:34:
called `Result::unwrap()` on an `Err` value: Os { code: 13, kind: PermissionDenied, message: "Permission denied" }
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace


failures:
    cp::test_cp_fail_perm

test result: FAILED. 105 passed; 1 failed; 6 ignored; 0 measured; 0 filtered out; finished in 3.09s

Also tested on MacOS. All tests pass. No regressions.

[JRF63]

Forgot to reset the permissions for DD like in test_cp_issue199. Latest commit is the same as the previous plus the said modification:

@@ -342,7 +342,7 @@ fn test_cp_fail_perm() {
         1,
     );

-    for f in [test_dir, d, d_a] {
+    for f in [test_dir, d, d_a, dd] {
         fs::set_permissions(f, fs::Permissions::from_mode(0o777)).unwrap();
     }
     fs::remove_dir_all(test_dir).unwrap();

[jgarzik]

Testing successful, well done.

Fixed #199