Re: [Fastcgipp-users] SIGPIPE causes SEGFAULT

[Eddie Carle]

On Tue, 2010-11-09 at 11:57 +0100, Volker Schreiner wrote:
> I checked the databaseExample and experienced the same behaviour. If i
> use apache benchmark and cause nginx to generate a SIGPIPE signal the
> example database application also causes an segmentation fault:
> 
> #0  0x00541087 in ?? () from /usr/lib/libfastcgipp.so.2
> #1  0x00528fce in Fastcgipp::Transceiver::Buffer::freeRead
> (this=0xb3c117f8, size=16) at transceiver.cpp:177
> #2  0x0052954f in Fastcgipp::Transceiver::transmit (this=0xbfbadac4)
> at
> transceiver.cpp:43
> #3  0x00524d7d in Fastcgipp::Request<wchar_t>::complete() () from
> /usr/lib/libfastcgipp.so.2
> #4  0x0052566f in Fastcgipp::Request<wchar_t>::handler() () from
> /usr/lib/libfastcgipp.so.2
> #5  0x0813aa49 in Fastcgipp::Manager<Database>::handler
> (this=0xbfbadac4) at /usr/local/include/fastcgi++/manager.hpp:345
> #6  0x08137d61 in main () at ../src/main.cpp:281
> 
> So it seems that this is a problem of the blocked SIGPIPE signal and
> not a problem of my thread per request design.

Interestingly enough, the database example uses two threads. One for the
requests, and one for communicating with the database. Can you do a
similar test with say echo, helloworld or session? These are single
thread examples.
-- 
        Eddie Carle