[cl-plus-ssl-devel] Possible bug in test-https-server (2008-11-04)

[Anton Vodonosov]

Hello John.

I am not the original author, just happen to have write access to the
repository, therefore can't be 100% sure that it is not intentional
(esp. taking into account ccl:make-socket call in the example.lisp,
that allows to suppose that cl+ssl was developed on CCL).

But it seems to be an error, closing ssl wrapper before wrapped socket
is reasonable, moreover you reported that it works only that
way on CCL.

Therefore I committed this change (for both test-https-client
and test-https-server).

Thanks for the report.

Best regards,
-Anton

on Saturday, January 17, 2009, 4:49:41 PM John wrote:

> All,

> I've recently tried to use CL+SSL on CCL (Clozure Common Lisp Version
> 1.2-r9226-RC1).

> When I loaded cl+ssl (the latest tarball, 2008-11004) and its various
> dependencies (whatever was latest on the web), I tried to load and run
> the samples in example.lisp

> (ssl-test::test-https-client)

> works fine, and produces useful output. However,

> (ssl-test::test-https-server)

> almost works, but fails to return output to the connecting browser:


> ? (ssl-test::test-https-server)
> SSL server listening on port 8080
HTTPS>> GET / HTTP/1.1
HTTPS>> User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10_5_6;
> en-us) AppleWebKit/525.27.1 (KHTML, like Gecko) Version/3.2.1 Safari/
> 525.27.1
HTTPS>> Accept: text/xml,application/xml,application/xhtml+xml,text/
> html;q=0.9,text/plain;q=0.8,image/png,*/*;q=0.5
HTTPS>> Accept-Language: en-us
HTTPS>> Accept-Encoding: gzip, deflate
HTTPS>> Connection: keep-alive
HTTPS>> Host: debian-lisp.local:8080
>> Error: An I/O error occurred: undocumented reason. (return code:
> 5)SSL error queue:
>>
>> While executing: CL+SSL::SSL-SIGNAL-ERROR, in process listener(1).
>> Type :POP to abort, :R for a list of available restarts.
>> Type :? for other options.
1 >> :b
>   (2AAAAACBB848) : 0 (SSL-SIGNAL-ERROR #<A Foreign Pointer #x684B70>
> #<Compiled-function CL+SSL::SSL-WRITE #x300040FEF01F> 5 -1) 397
>   (2AAAAACBB880) : 1 (FUNCALL #'#<#<STANDARD-METHOD STREAM-FORCE-
> OUTPUT (CL+SSL::SSL-STREAM)>> #<SSL-SERVER-STREAM for 6>) 1037
>   (2AAAAACBB908) : 2 (FORCE-OUTPUT [...]) 69
>   (2AAAAACBB920) : 3 (FUNCALL #'#<#<STANDARD-METHOD CLOSE (CL+SSL::SSL-
STREAM)>>> #<SSL-SERVER-STREAM for 6> [...]) 157
>   (2AAAAACBB950) : 4 (%%CHECK-KEYWORDS #(1 #(:ABORT) #<METHOD-FUNCTION
CLOSE #>>) 5864062285622) 245
>   (2AAAAACBB9D0) : 6 (%%CHECK-KEYWORDS #(1 #(:ABORT) #<METHOD-FUNCTION
CLOSE #>>) 5864062285637) 245
>   (2AAAAACBBA38) : 8 (TEST-HTTPS-SERVER [...]) 557

> (other items from stack trimmed for brevity)

> Investigation by myself and Gary Byers on the openmcl-dev mailing list
> suggests that the underlying cause is that the test server is trying  
> to write its output via a handle it has already closed. No output is  
> attempted until the close triggers some sort of flush. Looking at the
> source:

>       (let* ((socket (trivial-sockets:accept-connection
>                       server
>                       :element-type '(unsigned-byte 8)))
>              (client (cl+ssl:make-ssl-server-stream
>                       (cl+ssl:stream-fd socket)
>                       :external-format :iso-8859-1
>                       :certificate cert
>                       :key key)))

> <snip>

>           (close socket)  ; possibly closing the shared fd
>           (close client)  ; possibly forcing output to the closed fd

> It looks like it closes socket before client, yet client depends on  
> socket for the server to return data to the browser.

> If I swap the order of the close calls:

>           (close client)
>           (close socket)

> All is well, and the test server returns a page to the browser.

> I had a look at the CVS, to see if this was some sort of recent  
> change, but the web interface befuddled me, and I couldn't see history
> for the example.lisp file.

> Is this a bug? Or is there something amiss with my lisp config/ 
> dependencies, etc?

> J



> _______________________________________________
> cl-plus-ssl-devel mailing list
> cl-plus-ssl-devel at common-lisp.net
> http://common-lisp.net/cgi-bin/mailman/listinfo/cl-plus-ssl-devel