Swazoo and multipart post

Hi,

Trying to port Aida to gst, I found a bug with multipart posts : it
makes gst freeze.

I also tried Seaside functional tests (the multipart form test), and gst
freezes too.

I tried to debug it without any success (I'm not a Sawzoo expert).

Is this a known bug?

Cheers!

Nico

_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk

nico wrote:
Probably; it's a pretty new feature in Swazoo 2.2.

Any change you can capture the HTTP request?

Thanks,

Paolo



_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk

Paolo Bonzini wrote:
> nico wrote:

>> Trying to port Aida to gst, I found a bug with multipart posts : it
>> makes gst freeze.
>>
>> I also tried Seaside functional tests (the multipart form test), and gst
>> freezes too.
>>
>> I tried to debug it without any success (I'm not a Sawzoo expert).
>>
>> Is this a known bug?

> Probably; it's a pretty new feature in Swazoo 2.2.
>
> Any change you can capture the HTTP request?

I chatted with Nico about that problem on IRC and it seems there is
infinite recursion somewhere in the Swazoo code. CPU is on 100% and
memory grows.

So I propose that a breakpoint is to be put somewhere at the start of
request parsing (let say HTTPRequest class #newFor:readFrom:) and step
through until the freeze. I think a thorough inspection of the methods
on the way will reveal some missing ^ or similar traditional error to
get in infinite recursion soon.

Best regards
Janko




--
Janko Mivšek
AIDA/Web
Smalltalk Web Application Server
http://www.aidaweb.si


_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk

> I chatted with Nico about that problem on IRC and it seems there is
> infinite recursion somewhere in the Swazoo code. CPU is on 100% and
> memory grows.
>
> So I propose that a breakpoint is to be put somewhere at the start of
> request parsing (let say HTTPRequest class #newFor:readFrom:) and step
> through until the freeze. I think a thorough inspection of the methods
> on the way will reveal some missing ^ or similar traditional error to
> get in infinite recursion soon.

Yeah, I wanted to debug it with a real HTTP request and I'm too lazy to
capture one myself. :-)

Paolo


_______________________________________________
help-smalltalk mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/help-smalltalk