FLEXI-STREAMS now passes tests on 1.3.0-dev (was Re: [armedbear-devel] Gray streams bugs)

[Theam Yong Chew]

On 1/17/14, Anton Vodonosov <avodonosov at yandex.ru> wrote:
> 16.01.2014, 21:22, "Mark Evenson" <evenson at panix.com>:
>> On Jan 16, 2014, at 13:44, Theam Yong Chew <senatorzergling at gmail.com>
>> wrote:
>>
>>>  Hi all,
>>>
>>>  I inadvertently discovered some bugs in the gray-streams
>>>  implementation due to a stale flexi-streams that did not get upgraded
>>>  properly.
>>
>> Applied in [r14602][1]:  FLEXI-STREAMS from Quicklisp passes all its
>> tests!  Fantastic!
>>
>> [1]: http://abcl.org/trac/changeset/14602
>
> Flexi-streams test suite passes even on ABCL 1.1.1
> http://common-lisp.net/project/cl-test-grid/library/flexi-streams.html

That's right. Just a quick note, I was going to say I didn't deserve
any credit. What really happened to me was the result of an upgrade
from a very old quicklisp.  trivial-gray-streams was upgraded, but
somehow my flexi-streams didn't recompile itself. The stale .abcls had
a different superclass hierarchy, mucking up the method dispatch. This
resulted in ABCL's own gray-streams methods being invoked.

Since in practice, most folks would implement both read-char &
read-sequence together (including flexi-streams), the ABCL gray
streams probably never got used thus far. That's why the bugs were
never uncovered. No bugs (not related to my topic anyway :-)) were
uncovered in flexi-streams or ABCL's support of flexi-streams this time.

Yong