Testing and lock noise stipple.

[John Arbash Meinel]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Vincent Ladeuil wrote:
>>>>>> "robert" == Robert Collins <robert.collins at canonical.com> writes:
> 
>     robert> The test suite now looks for unreleased *physical*
>     robert> locks, and will output noise (via print)
> 
> To say the least...
> 
>     robert> when a test has unbalanced lock/unlock counts.
> 
>     robert> These can be turned into errors via the use of
>     robert> '-Dlock' when running the test suite.
> 
>     robert> Once we have the test suite passing cleanly, We can
>     robert> turn this into an error by default.
> 
>     robert> Please start using -Dlock when running new tests you
>     robert> have created; if you have time consider running with
>     robert> -Dlock at other times as well.
> 
> A bit of feedback:
> 
> 1 - the actual noise is *really* too much (may the message ought
>     to be displayed only when running with -Dlock ?). Granted,
>     the attached patch makes it worse :)
> 
> 2 - there is no *global* way to disable it (but making the output
>     optional may me enough),
> 
> 3 - there is no *local* way to disable it (on a test by test
>     basis),
> 
> Having said that, I'd love to find a better way to address the
> underlying problem so I tried to fix some tests in the attached
> patch.
> 
> I started with:
> 
>   ./bzr selftest -s bt.test_lockdir -Dlock
> 
> and fixed 7 of the 10 failures.
> 
> If you look at the patch *in context*
> test_42_confirm_broken_manually gives a strange feeling of
> artificial cleanup.
> 
> I couldn't find a good way to address the 3 remaining failures
> without violating the test spirit, feedback welcome.
> 
> test_break_lock: leaves the lock/unlock unbalanced on purpose
> 
> test_unlock_after_break_raises: requires that lock/unlock are unbalanced
> 
> test_43_break: makes it impossible to balance the lock/unlock by design
> 
> I didn't look at the 70 remaining failures yet so the 3 above may
> be pathological.
> 
>         Vincent
> 
> 

BTW, BB:approve on the patch, if it was meant for review.

John
=:->

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkoATTwACgkQJdeBCYSNAAPzpwCaAlESVDh0prkPoQMEYVxdsziE
CX4AoNahtF7bCbQ3aoKHTbof+xcHteL/
=iNEZ
-----END PGP SIGNATURE-----