From avodonosov at yandex.ru  Wed May 12 16:36:33 2010
From: avodonosov at yandex.ru (Anton Vodonosov)
Date: Wed, 12 May 2010 20:36:33 +0400
Subject: [cl-openid-devel] Fwd: Re: [cl-plus-ssl-devel] Two LISTEN bugs
Message-ID: <78521273682193@web120.yandex.ru>



12.05.10, 20:31, "Ron Garret" <ron at flownet.com>:


On May 12, 2010, at 9:29 AM, Anton Vodonosov wrote:

>>>> BTW, if it was not intended for CL+SSL to do this kind of buffering, why bother to allocate an input buffer?
>>> 
>>> Maybe to avoid consing, i.e. to not allocate new buffer every time when we do SSL_Read?
>> 
>> No, that's not what I mean.  CL+SSL allocates a 2000 byte input buffer (once).  But then it only ever uses the first byte.
>> 
> 
> There is also function stream-read-sequence, it uses more.

Ah.  Thanks.

rg




From avodonosov at yandex.ru  Tue May 25 21:25:59 2010
From: avodonosov at yandex.ru (Anton Vodonosov)
Date: Wed, 26 May 2010 01:25:59 +0400
Subject: [cl-openid-devel] [cl-plus-ssl-devel] Two LISTEN bugs
In-Reply-To: <78521273682193@web120.yandex.ru>
References: <78521273682193@web120.yandex.ru>
Message-ID: <113941274822759@web66.yandex.ru>

Hello, 

I added CCL version of the READ-CHAR-NO-HANG.

Ron Garret <ron at flownet.com>
> Just FYI, CL+SSL out of the box fails 4 of 8 tests on the CCL trunk. (My patch seems to have fixed one of them. I now fail only 3 of 8

The remaining failures are when the tests are run with Lisp-BIO, which is know to be broken (Lisp-BIO is a Lisp implementation of OpenSSL BIO: http://www.openssl.org/docs/crypto/bio.html).

With the new read-char-no-hang test we have 4 failures of 12 tests. The same on SBCL.

Best regards,
- Anton



From avodonosov at yandex.ru  Wed May 26 16:18:52 2010
From: avodonosov at yandex.ru (Anton Vodonosov)
Date: Wed, 26 May 2010 20:18:52 +0400
Subject: [cl-openid-devel] [cl-plus-ssl-devel] Two LISTEN bugs
Message-ID: <2001274890732@web65.yandex.ru>

Sorry, two previous letters get into the cl-openid list by mistake. They were intended for cl-plus-ssl.

- Anton