Drop -v from rm and cp commands in fixtures for portability
#1646
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This removes `-v` from an `rm -R` command in `make_basic_repo.sh`, which can cause tests that needed to run that fixture to fail on systems where `rm` does not support `-v`. The script generates a fixtured used in numerous tests, and it is always run even if `GIX_TEST_IGNORE_ARCHIVES` is not set, since its output is `.gitignore`d. Because the script is so heavily used, this change allows many more tests to pass on systems where `rm` does not recognize `-v` than before. `rm -v` is actually POSIX, but it was standardized very recently, in 1003.1-2024; it was not even required in 1003.1-2017. Some Unix systems' `rm` commands do not recognize `-v`. This includes most (or all?) illumos and Solaris systems (except when `/usr/gnu/bin` is present and placed early in the `$PATH`). It seems likely that there are a number of other such systems. Regarding the history of `rm -v` in POSIX, see: - https://pubs.opengroup.org/onlinepubs/9699919799/utilities/rm.html - https://pubs.opengroup.org/onlinepubs/9799919799/utilities/rm.html - https://www.austingroupbugs.net/view.php?id=1154 - https://www.austingroupbugs.net/view.php?id=1365 - https://www.austingroupbugs.net/view.php?id=1487 Based on a `git grep -En 'rm( -[^v ;&|]*)+v'` search, this occurrence in `make_basic_repo.sh` appears to be the only place where `-v` is passed to `rm` in any gitoxide fixture script. Two nearby similar, conceptually related commands omit `-v`, and when this command (rather, the command that turned into it) was introduced in the first version of the script in a4cec4a (GitoxideLabs#399), another such command without `-v` was also present. This suggests that `-v` may have been used for exploratory purposes during development, such that the approach here of omitting `-v` it without replacing it with anything might be sufficient. However, it is possible for information shown when the fixture script runs and fails to be less informative as a result of this omission. An alternative approach to the one taken here could be to have the script carry out a tiny experiment to check if `-v` is supported, then use it only if it was found to be accepted. (Another option could be to use `find` with both `-print` and `-delete`, though that might be stronger than desired given that the affected `rm` command does not pass `-f`.)
This change removes the `-v` option from the one `cp` invocation in `gix-dir/tests/fixtures/many.sh`, the one fixture script that uses it. This is roughly analogous to the preceding removal of `-v` from the one `mv` invocation that used that, and both fixture scripts are heavily used due to being used in multiple tests and due to being intentionally re-run even when `GIX_TEST_IGNORE_ARCHIVES` is not set. But while `rm -v` is standardized but only recently, `cp -v` is not standard. Even POSIX 1003.1-2024 does not require `cp` to recognize `-v`: - https://pubs.opengroup.org/onlinepubs/9799919799/utilities/cp.html See the preceding commit for general considerations, advantages and disadvantages, and alterntive approaches. As there, this currently just drops the `-v` without replacing it with anything, based on the idea that printing the list of files being operated on (here, copied) may not be important. - That assumption may be less well-founded here, since the affected command, introduced in 482d6f3 (GitoxideLabs#1285), does *not* have other nearby `cp` commands that are conceptually related but omit `-v`. - But, unlike there, it may be reasonable to consider the `-v` here to be a (minor) bug, since it is not standardized but it appears in test fixture code that seems intended to run all targets. Taken together, this and the preceding commit allow 106 formerly failing tests to pass, on the OmniOS system I used for testing.
Byron
approved these changes
Oct 24, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
In the fixture scripts, there is one invocation of
rmwith the-voption, which is only very recently standardized, and one invocation ofcpwith the-voption, which is nonstandard. The intent appears to be to make it easier to understand what a fixture has done if it fails, or if it is run manually to investigate it. This PR drops-vfrom those commands without replacing it with anything.The two affected fixture scripts are heavily used, and the archives they generate are
.gitignored. So this change causes 106 formerly failing tests to pass on my OmniOS test system (decreasing the number of failing tests from 108 to 2). I expect most illiumos and various other significant but less common Unix-like systems to have the same benefit.The tradeoff of not showing verbose output appears to be reasonable to me in context here, but it may not be the best approach. Alternative approaches, and other details about the change--including the conditions under which it makes a difference and its relationship to POSIX--are given in the commit messages.