| Message ID | 20240301034606.69673-2-shejialuo@gmail.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | t9117: prefer test_path_* helper functions | expand |
On Thu, Feb 29, 2024 at 10:46 PM shejialuo <shejialuo@gmail.com> wrote: > test -(e|f|d) does not provide a nice error message when we hit test > failures, so use test_path_exists, test_path_is_dir and > test_path_is_file instead. Thanks for rerolling. t9117 is indeed a better choice[1] than t3070 for the exercise of replacing `test -blah` with `test_path_foo`. [1]: https://lore.kernel.org/git/CAPig+cR2-6qONkosu7=qEQSJa_fvYuVQ0to47D5qx904zW08Eg@mail.gmail.com/ > Signed-off-by: shejialuo <shejialuo@gmail.com> > --- > diff --git a/t/t9117-git-svn-init-clone.sh b/t/t9117-git-svn-init-clone.sh > @@ -15,39 +15,39 @@ test_expect_success 'setup svnrepo' ' > test_expect_success 'basic clone' ' > - test ! -d trunk && > + ! test_path_is_dir trunk && Generally speaking, you don't want to use `!` to negate the result of a `path_is_foo` assertion function. To understand why, take a look at the definition of `test_path_is_dir`: test_path_is_dir () { if ! test -d "$1" then echo "Directory $1 doesn't exist" false fi } The test in question (t9117: "basic clone") is using `test ! -d` to assert that the directory `trunk` does not yet exist when the test begins; indeed, under normal circumstances, this directory should not yet be present. However, the call to test_path_is_dir() asserts that the directory _does_ exist, which is the opposite of `test ! -d`, and complains ("Directory trunk doesn't exist") when it doesn't exist. So, in the normal and typical case for all the tests in this script, `test_path_is_dir` is going to be complaining even though the non-existence of that directory is an expected condition. Although you make the test pass by using `!` to invert the result of `test_path_is_dir`, the complaint will nevertheless get lodged, and may very well be confusing for anyone scrutinizing the output of the tests when running the script with `-v` or `-x`. So, `test_path_is_dir` is not a good fit for this case which wants to assert that the path `trunk` does not yet exist. A better choice for this particular case would be `test_path_is_missing`. > git svn clone "$svnrepo"/project/trunk && > - test -d trunk/.git/svn && > - test -e trunk/foo && > + test_path_is_dir trunk/.git/svn && > + test_path_exists trunk/foo && These two changes make sense and the intent directly corresponds to the original code. > test_expect_success 'clone to target directory' ' > - test ! -d target && > + ! test_path_is_dir target && > git svn clone "$svnrepo"/project/trunk target && > - test -d target/.git/svn && > - test -e target/foo && > + test_path_is_dir target/.git/svn && > + test_path_exists target/foo && > rm -rf target > ' What follows is probably beyond the scope of your GSoC microproject, but there is a bit more of interest to note about these tests. Rather than asserting some initial condition at the start of the test, it is more common and more robust simply to _ensure_ that the desired initial condition holds. So, for instance, instead of asserting `test ! -d target`, modern practice is to ensure that `target` doesn't exist. Thus: test_expect_success 'clone to target directory' ' rm -rf target && git svn clone "$svnrepo"/project/trunk target && ... is a more robust implementation. This also addresses the problem that the `rm -rf target` at the very end of each test won't be executed if any command earlier in the test fails (due to the short-circuiting behavior of the &&-operator). As noted, this type of cleanup is probably overkill for your GSoC microproject so you need not tackle it. I mention it only for completeness. Also, if someone does tackle such a cleanup, it should be done as multiple patches, each making one distinct change (i.e. one patch dropping `test !-d` and moving `rm -rf` to the start of the test, and one which employs `test_path_foo` for the remaining `test -blah` invocations).
shejialuo <shejialuo@gmail.com> writes: > test_expect_success 'basic clone' ' > - test ! -d trunk && > + ! test_path_is_dir trunk && This is not quite right. Step back and think why we are trying to use the test_path_* helpers instead of "test [!] -d". What are the differences between them? The answer is that, unlike "test [!] -d dir" that is silent whether "dir" exists or missing, "test_path_is_dir dir" is *not* always silent. It gives useful messages as necessary. When does it do so? Here is the definition, from t/test-lib-functions.sh around line 930: test_path_is_dir () { test "$#" -ne 1 && BUG "1 param" if ! test -d "$1" then echo "Directory $1 doesn't exist" false fi } It succeeds silently when "test -d dir" is true, but it complains loudly when "test -d dir" does not hold. You will be told that the test is unhappy because "dir" does not exist. That would be easier to debug than one step among many in &&-chain silently fails. Now, let's look at the original you rewrote again: > - test ! -d trunk && It says "it is a failure if 'trunk' exists as a directory". If 'trunk' does not exist, it is a very happy state for us. So instead of silently failing when 'trunk' exists as a directory, you would want to improve it so that you will get a complaint in such a case, saying "trunk should *not* exist but it does". Did you succeed to do so with this rewrite? > + ! test_path_is_dir trunk && The helper "test_path_is_dir" is called with "trunk". As we saw, we will see complaint when "trunk" does *NOT* exist. When "trunk" does exist, it will be silent and "test_path_is_dir" will return a success, which will be inverted with "!" to make it a failure, causing &&-chain to fail. So the exit status is not wrong, but it issues a complaint under the wrong condition. That is not an improvement. Let's step back one more time. Is the original test happy when "trunk" existed as a regular file? "test ! -d trunk" says so, but should it really be? Think. I suspect that the test is not happy as long as 'trunk' exists, whether it is a directory or a regular file or a symbolic link. IOW, it says "I am unhappy if 'trunk' is a directory", but what it really meant to say was "I am unhappy if there is anything at the path 'trunk'". IOW, "test ! -e trunk" would be what it really meant, no? So the correct rewrite for it would rather be something like test_path_is_missing trunk && instead. This will fail if anything is at path 'trunk', with an error message saying there shouldn't be anything but there is. In a peculiar case, which I do not think this one is, a test may legitimately accept "path" to either (1) exist as long as it is not a directory, or (2) be missing, as success. In such a case, the original construct '! test -d path" (or "test ! -d path") would be appropriate. But I do not think we have a suitable wrapper to express such a case, i.e. we do not have a helper like this. test_path_is_not_dir () { if test -d "$1" then echo "$1 is a directory but it should not be" false fi } If such a use case were common, we might even do this: # "test_path_is_dir <dir>" expects <dir> to be a directory. # "test_path_is_dir ! <dir>" expects <dir> not to be a # directory. # In either case, complain only when the expectation is not met. test_path_is_dir () { if test "$1" = "!" then shift if test -d "$1" then echo "$1 is a directory but it should not be" return 1 fi else if test ! -d "$1" then echo "$1 is not a directory" return 1 fi fi true } but "we are happy even if path exists as long as it is not a directory" is a very uncommon thing we want to say in our tests, so that is why we do not have such a helper function. HTH.
Thanks for your comment. > Although you make the test pass by using `!` to invert the result of > `test_path_is_dir`, the complaint will nevertheless get lodged, and > may very well be confusing for anyone scrutinizing the output of the > tests when running the script with `-v` or `-x`. I have run the script with `-v`, I have got the following result: Directory trunk doesn't exist I come to realisize the fault with your dedicated comments. An assertion is an assertion. And I am impressed by the following idea: > Rather than asserting some initial condition at the start of the test, > it is more common and more robust simply to _ensure_ that the desired > initial condition holds. So, for instance, instead of asserting `test > ! -d target`, modern practice is to ensure that `target` doesn't > exist. Thus: > > test_expect_success 'clone to target directory' ' > rm -rf target && > git svn clone "$svnrepo"/project/trunk target && > ... > > is a more robust implementation. This also addresses the problem that > the `rm -rf target` at the very end of each test won't be executed if > any command earlier in the test fails (due to the short-circuiting > behavior of the &&-operator). The command `rm -rf target` ensures an exit status of 0 regardless of whether the `target` exists. Thus the code will elegant make sure the initial condition holds. I think I could add a patch to clean the code.
Thanks for your wonderful comments. I have known that semantics is important not only the functionality. I will send a new patch at now.
diff --git a/t/t9117-git-svn-init-clone.sh b/t/t9117-git-svn-init-clone.sh index 62de819a44..2f964f66aa 100755 --- a/t/t9117-git-svn-init-clone.sh +++ b/t/t9117-git-svn-init-clone.sh @@ -15,39 +15,39 @@ test_expect_success 'setup svnrepo' ' ' test_expect_success 'basic clone' ' - test ! -d trunk && + ! test_path_is_dir trunk && git svn clone "$svnrepo"/project/trunk && - test -d trunk/.git/svn && - test -e trunk/foo && + test_path_is_dir trunk/.git/svn && + test_path_exists trunk/foo && rm -rf trunk ' test_expect_success 'clone to target directory' ' - test ! -d target && + ! test_path_is_dir target && git svn clone "$svnrepo"/project/trunk target && - test -d target/.git/svn && - test -e target/foo && + test_path_is_dir target/.git/svn && + test_path_exists target/foo && rm -rf target ' test_expect_success 'clone with --stdlayout' ' - test ! -d project && + ! test_path_is_dir project && git svn clone -s "$svnrepo"/project && - test -d project/.git/svn && - test -e project/foo && + test_path_is_dir project/.git/svn && + test_path_exists project/foo && rm -rf project ' test_expect_success 'clone to target directory with --stdlayout' ' - test ! -d target && + ! test_path_is_dir target && git svn clone -s "$svnrepo"/project target && - test -d target/.git/svn && - test -e target/foo && + test_path_is_dir target/.git/svn && + test_path_exists target/foo && rm -rf target ' test_expect_success 'init without -s/-T/-b/-t does not warn' ' - test ! -d trunk && + ! test_path_is_dir trunk && git svn init "$svnrepo"/project/trunk trunk 2>warning && ! grep -q prefix warning && rm -rf trunk && @@ -55,7 +55,7 @@ test_expect_success 'init without -s/-T/-b/-t does not warn' ' ' test_expect_success 'clone without -s/-T/-b/-t does not warn' ' - test ! -d trunk && + ! test_path_is_dir trunk && git svn clone "$svnrepo"/project/trunk 2>warning && ! grep -q prefix warning && rm -rf trunk && @@ -69,7 +69,7 @@ project/trunk:refs/remotes/${prefix}trunk project/branches/*:refs/remotes/${prefix}* project/tags/*:refs/remotes/${prefix}tags/* EOF - test ! -f actual && + ! test_path_is_file actual && git --git-dir=project/.git config svn-remote.svn.fetch >>actual && git --git-dir=project/.git config svn-remote.svn.branches >>actual && git --git-dir=project/.git config svn-remote.svn.tags >>actual && @@ -78,7 +78,7 @@ EOF } test_expect_success 'init with -s/-T/-b/-t assumes --prefix=origin/' ' - test ! -d project && + ! test_path_is_dir project && git svn init -s "$svnrepo"/project project 2>warning && ! grep -q prefix warning && test_svn_configured_prefix "origin/" && @@ -87,7 +87,7 @@ test_expect_success 'init with -s/-T/-b/-t assumes --prefix=origin/' ' ' test_expect_success 'clone with -s/-T/-b/-t assumes --prefix=origin/' ' - test ! -d project && + ! test_path_is_dir project && git svn clone -s "$svnrepo"/project 2>warning && ! grep -q prefix warning && test_svn_configured_prefix "origin/" && @@ -96,7 +96,7 @@ test_expect_success 'clone with -s/-T/-b/-t assumes --prefix=origin/' ' ' test_expect_success 'init with -s/-T/-b/-t and --prefix "" still works' ' - test ! -d project && + ! test_path_is_dir project && git svn init -s "$svnrepo"/project project --prefix "" 2>warning && ! grep -q prefix warning && test_svn_configured_prefix "" && @@ -105,7 +105,7 @@ test_expect_success 'init with -s/-T/-b/-t and --prefix "" still works' ' ' test_expect_success 'clone with -s/-T/-b/-t and --prefix "" still works' ' - test ! -d project && + ! test_path_is_dir project && git svn clone -s "$svnrepo"/project --prefix "" 2>warning && ! grep -q prefix warning && test_svn_configured_prefix "" && @@ -114,7 +114,7 @@ test_expect_success 'clone with -s/-T/-b/-t and --prefix "" still works' ' ' test_expect_success 'init with -T as a full url works' ' - test ! -d project && + ! test_path_is_dir project && git svn init -T "$svnrepo"/project/trunk project && rm -rf project '
test -(e|f|d) does not provide a nice error message when we hit test failures, so use test_path_exists, test_path_is_dir and test_path_is_file instead. Signed-off-by: shejialuo <shejialuo@gmail.com> --- t/t9117-git-svn-init-clone.sh | 40 +++++++++++++++++------------------ 1 file changed, 20 insertions(+), 20 deletions(-)