[Openmcl-devel] Problem with read-sequence on binary streams

[Gary Byers]

On Wed, 11 Jun 2008, Wade Humeniuk wrote:

> Thanks for the mods, version  1.2-r9736M-RC1 is running fine with Drakma.

Thanks for finding and reporting the bug.

>
> The two versions of STREAM-READ-VECTOR would explain why my
> change to the made no difference on a rebuild.  Not
> something I would have thought to look for.

Whoever invented cut-and-paste editing clearly had no idea how
dangerous it could be in the wrong hands.


>
> Wade
>
>
>
> On Wed, Jun 11, 2008 at 12:23 PM, Gary Byers <gb at clozure.com> wrote:
>>
>>
>> On Wed, 11 Jun 2008, p2.edoc at googlemail.com wrote:
>>
>>> I wondered if it should be (read-byte stream nil :eof)
>>> but then there are plenty of other (read-byte stream) calls that seem
>>> to anticipate EOF conditions return :eof.
>>>
>>
>> I found a total of 3 such calls: the one that Wade ran into in the
>> STREAM-READ-VECTOR method on BINARY-INPUT-STREAM, another
>> in the STREAM-READ-LIST method on BINARY-INPUT-STREAM, and ... well,
>> there were apparently two identical copies of the STREAM-READ-VECTOR
>> method in l1-streams.lisp.  Those cases should be fixed in svn as
>> of <http://trac.clozure.com/openmcl/changeset/9730>,
>>
>> Did I miss others ?
>>
>
>