most likely it can't find the writable cert db files or perhaps it just
can't find crypto.
do you think you could provide a stack trace? gecko has lots of lines
of code and predicting exactly where it will crash in an unspecified
version of gecko (hint: please specify) is hard.
nss3.dll!024820ce()
[Frames below may be incorrect and/or missing, no symbols loaded for
nss3.dll]
nss3.dll!02472d27()
ssl3.dll!02501a17()
ssl3.dll!0250acb1()
ssl3.dll!0250d94f()
ssl3.dll!0250d95f()
nspr4.dll!300180bc()
nspr4.dll!30015947()
ssl3.dll!0250f75b()
ssl3.dll!025105e5()
ssl3.dll!02510696()
ssl3.dll!0251343a()
pipnss.dll!023f1b47()
nspr4.dll!30015e7b()
ntdll.dll!7c926abe()
nspr4.dll!30017f2d()
> msvcr80.dll!_callthreadstartex() Line 348 + 0x6 bytes C
msvcr80.dll!_threadstartex(void * ptd=0x019ef618) Line 326 + 0x5 bytes C
kernel32.dll!7c80b683()
ntdll.dll!7c926abe()
"timeless" <time...@gmail.com> wrote in message
news:1164616312.9...@n67g2000cwd.googlegroups.com...
ntdll.dll!7c901230()
[Frames below may be incorrect and/or missing, no symbols loaded for
ntdll.dll]
nspr4.dll!PR_Assert(const char * s=0x023d7ad8, const char *
file=0x023d7aa0, int ln=0x00000734) Line 546 C
> nss3.dll!PK11_GetInternalSlot() Line 1844 + 0x1d bytes C
nss3.dll!PK11_TokenExists(unsigned long type=0x00000001) Line 1897 + 0x5
bytes C
ssl3.dll!ssl3_config_match_init(sslSocketStr * ss=0x03223410) Line 642 +
0x39 bytes C
ssl3.dll!ssl2_ConstructCipherSpecs(sslSocketStr * ss=0x03223410) Line 206
+ 0x9 bytes C
ssl3.dll!ssl2_BeginClientHandshake(sslSocketStr * ss=0x03223410) Line
3007 + 0x9 bytes C
ssl3.dll!ssl_Do1stHandshake(sslSocketStr * ss=0x03223410) Line 149 + 0xf
bytes C
ssl3.dll!ssl_SecureSend(sslSocketStr * ss=0x03223410, const unsigned char
* buf=0x030a6760, int len=0x00000196, int flags=0x00000000) Line 1090 + 0x9
bytes C
ssl3.dll!ssl_SecureWrite(sslSocketStr * ss=0x03223410, const unsigned char
* buf=0x030a6760, int len=0x00000196) Line 1128 + 0x13 bytes C
ssl3.dll!ssl_Write(PRFileDesc * fd=0x014075c8, const void *
buf=0x030a6760, int len=0x00000196) Line 1413 + 0x17 bytes C
pipnss.dll!nsSSLThread::Run() Line 919 C++
nspr4.dll!_PR_NativeRunThread(void * arg=0x30023b3d) Line 436 + 0x9 bytes
C
ntdll.dll!7c926a44()
msvcr80d.dll!_callthreadstartex() Line 348 + 0xf bytes C
msvcr80d.dll!_threadstartex(void * ptd=0x0300cc70) Line 331 C
kernel32.dll!7c80b683()
ntdll.dll!7c926a44()
PK11_GetInternalSlot() is returning null
For reference, here's my .mozconfig
mk_add_options MOZ_CO_PROJECT=browser
mk_add_options MOZ_OBJDIR=$topsrcdir/@CONFIG_GUESS@
ac_add_options --enable-application=browser
ac_add_options --disable-installer
ac_add_options --disable-static
ac_add_options --enable-shared
ac_add_options --enable-svg
ac_add_options --enable-crypto
ac_add_options --enable-debug
ac_add_options --disable-activex
ac_add_options --disable-activex-scripting
"timeless" <time...@gmail.com> wrote in message
news:1164616312.9...@n67g2000cwd.googlegroups.com...
I believe it's sufficient to do something like:
nsCOMPtr<nsISupports> psm = do_GetService("@mozilla.org/psm;1");
but it's been a few years. in short NSS needs to be initialized before
it's used, otherwise it kills itself and its consumer for violating its
api. in gecko, NSS is managed by PSM, which iirc is governed by that
singleton contract, however i'm not certain it's the right contract.
some breakpoints/tracing/testing might be warranted.
"timeless" <time...@gmail.com> wrote in message
news:1164749572.2...@n67g2000cwd.googlegroups.com...
Is your problem fixed ? I've tried to insert the proposed do_GetService
line but psm is NULL and the error code returned is NS_ERROR_FAILURE.
I've also take a look at the
https://bugzilla.mozilla.org/show_bug.cgi?id=309877 bug, but the two
proposed patches are obsolete...
I use the xulrunner-1.8.0.4 sources on a 64bits linux box compiled with
gcc-3.4.
Olivier