[Partysip-dev] crash-patch and better network diagnostics

[Fredrik Thulin]

hi

attached two patches fixes some things i found when working with partysip on a 
train (meaning no network) :

        * fredrik-partysip-crash-unresolvable-host2:
          a bit hard to read diff - the important line is the added

        if (req->rsp_ctx != NULL && req->rsp_ctx->branches != NULL)

        with libosip 0.8.8 and partysip 0.4.5 (what i had on the train) it was 
really
        easy to make partysip dump core just by trying to call "sip:9999" from
        linphone. with CVS libosip and CVS partysip i had to have two
        linphones and make something like 10 calls from each of them to get a 
        sigsegv.

        * fredrik-partysip-better-network-diagnostics-patch2:
        says what goes wrong when sending to the network goes wrong. i'm
        unsure about strerror() and errno on WIN32 so they don't get more than
        the numeric error code.

        also prints nicer error message when bind() to port fails and adds
        setsockopt() SO_REUSEADDR since it helps when partysip dumps core ;)

(btw. the thing i was working on on the train is an ENUM plugin (rfc2916) for 
partysip. do you consider that interesting? here in sweden we will probably 
accomplish a delegation of 6.4.e164.arpa soon, and i want to use partysip in 
our (Stockholm university's) tests)

/Fredrik

fredrik-partysip-better-network-diagnostics-patch2
Description: Text Data

fredrik-partysip-crash-unresolvable-host-patch2
Description: Text Data