Hi Daniel/Igor,

I have a problem where the process hangs while trying to post data using
libcurl.
The stack trace of the thread shows the folllwoing.
-----------------------------------
 ff065994 lwp_park (0, 0, 0)
 ff06166c mutex_lock_queue (ff078b44, 0, 3c4628, ff078000, 0, 0) + 104
 ff06206c slow_lock (3c4628, fef00200, 34, 3c4400, 1000, 39a400) + 58
 00308540 malloc (d, fed7b05c, 0, 7efefeff, 39a498, ff0000) + 48
 fefd4224 strdup (fed7b04c, fed7b04c, 100, fed7b23b, 430e5172, 8) + c
 0020da00 Curl_store_ip_addr (419800, 4322c0, 419898, fed7b1c4,
fed7b23b, 0) + 48
 00218100 ConnectPlease (419800, 4322c0, fed7b23b, 1a9e0, 21b2a0, 1000) + e0
 0021b43c SetupConnection (419800, 4322c0, fed7b322, fed7b323, 3aec00,
1000) + 15c
 0021b5f8 Curl_connect (d24000, fed7b3a0, fed7b323, fed7b322, ff0ea8d4,
25e) + 90
 0022986c Curl_connect_host (d24000, fed7b3a0, 0, 3e, 4, 0) + 54
 00229b80 Curl_perform (d24000, 271a, fed7b4b4, fed7b416, 3c7fc0,
ff266498) + 58
 0020cbf8 curl_easy_perform (d24000, 271a, fed7b4b4, 67650063, 6765,
587499) + 100
 000ed2f8
__1cJproxycommNURLConnectionLperformSend6kMrknDstdMbasic_string4Ccn0CLchar_traits4Cc__n0CJallocator4Cc____pn0CNbasic_ist
ream4Ccn0D___kIkpGpcIIpv_IkpG9AII9B_Ipn0CNbasic_ostream4Ccn0D____v_
(83bba8, fed7b7a8, fed7b878, 234, eb310, eb4b8) + 1d8
 000ec9a0
__1cJproxycommNURLConnectionEsend6kMrknDstdMbasic_string4Ccn0CLchar_traits4Cc__n0CJallocator4Cc____rn0CSbasic_strings
----------------------------------------

One of the threads hang at hostip.c in libcurl, where it was trying to
get the hostname.
Other threads were waiting for it. Does this patch/fix for
gethostbyname_thread:Access violation fix above problem.
problem occurs on libcurl 7.14.0.

Thanks & Regards,
Prasanna

Daniel Stenberg wrote:

> On Sun, 28 Aug 2005, Igor Polyakov wrote:
>
>> I think I've fixed this bug. Could anybody test my fix ?
>
>
> I've now applied this patch (with only minor indenting changes) so now
> hopefully it'll be easier for people to test it. Tomorrow's daily
> snapshot will contain this fix.
>
> Please people, test this and tell us how it works. I don't want to
> release 7.14.1 without at least a few positive tests results on this.
>
Received on 2005-08-30