Re: [PATCH 1/3] t0000: set TEST_OUTPUT_DIRECTORY for sub-tests

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi,

Jeff King wrote:

> Once upon a time, the test-lib library would create trash
> directories in the current working directory, unless we were
> explicitly told to put it elsewhere via --root. As a result,
> t0000 created the sub-test trash directories inside its own
> trash directory.
>
> However, we noticed that this did not cover all cases, since
> we would need to respect $TEST_OUTPUT_DIRECTORY even if
> --root is not given (or is relative). Commit 38b074d fixed
> this to consistently use the full path.

So the idea if I am reading correctly is "Instead of relying on the
implicit output directory chosen with chdir, which doesn't even work
any more, set TEST_OUTPUT_DIRECTORY to decide where output for the
sub-tests used by t0000's sanity checks for the test harness go".

I'm not sure I completely understand the regression caused by 38b074d.
Is the idea that before that commit, TEST_OUTPUT_DIRECTORY was only
used for the test-results/ directory so the only harm done was some
mixing of test results?

What is the symptom this patch alleviates?

> As a result, t0000's sub-tests are now created in git's
> original test output directory rather than in our trash
> directory.

This might be the source of my confusion.  Is "sub-tests" an
abbreviation for "sub-test trash directories" here?

>            Furthermore, since some of the sub-tests simulate
> failures, the trash directories do not get cleaned up, and
> the cruft is left in the t/ directory.
>
> We could fix this by passing a new "--root=$TRASH_DIRECTORY"
> option to the sub-test. However, we do not want the sub-tests
> to write anything at all to git's directory (e.g., they
> should not be writing to t/test-results, either, although
> this is already handled by separate code).

Ah, HARNESS_ACTIVE prevents output of test-results.

Does the git test harness write something else to
TEST_OUTPUT_DIRECTORY?  Is the idea that using --root would be
functionally equivalent but (1) more confusing and (2) less
futureproof?

>                                            So the best
> solution is to simply reset $TEST_OUTPUT_DIRECTORY entirely
> in the sub-test, which covers this case, as well as any
> future ones.

So, to sum up: if I understand correctly

 - git used to only use TEST_OUTPUT_DIRECTORY to decide where test
   results go.  You'd have to use --root to set a custom location for
   trash directories.

 - in that old setup, t0000 leaves around extra trash directories with
   --root, since the sub-tests inherit the parent test's $root and put
   trash directories there.

 - after 38b074d, that old problem still exists and furthermore
   t0000 leaves around extra trash directories even when --root is not
   in use, since the sub-tests inherit the value of
   TEST_OUTPUT_DIRECTORY from the parent test.

 - this patch fixes the TEST_OUTPUT_DIRECTORY problem (but not the $root
   problem) by setting TEST_OUTPUT_DIRECTORY explicitly

Does that sound right?  If so, should sub-tests unset $root, too?

Thanks and hope that helps,
Jonathan
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html




[Index of Archives]     [Linux Kernel Development]     [Gcc Help]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [V4L]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]     [Fedora Users]