[DRAFT] Proposed release schedule for 1.9.x

[Derek Atkins]

Quoting Josh Sled <jsled at asynchronous.org>:

> On Mon, 2006-01-16 at 16:21 -0500, Derek Atkins wrote:
>> I gave up that fight a couple years ago.  Getting people to make sure
>> "make check" worked was an uphill battle.  Worse, there are a number
>> of tests that use random input and, as a result, randomly fail.  Some
>> people wanted to fix this by making the test non-random..  But I think
>> that only masks the actual bug that certain inputs triggered.
>
> Then we should remove broken tests, so we know if something was broken
> un-intentionally.

Sometimes it's hard to differentiate between a broken test and a test
that detects broken code.  The latter should not be removed, the
former should (or at least should be fixed so the test isn't broken).

I think historically devs are too willing to call the test broken or
to call "random tests" bad...  And therefore disable the test instead
of tracking down and fixing (potentially) esoteric bugs.

-derek

-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord at MIT.EDU                        PGP key available