Hello,

I am now running 1.6R-i386 from CVS sources dated to about 48 hours ago,
and have all threads apps (mostly X stuff) using native threads.  I noticed
that xmms starts up way faster and seems to work moderately well (it wedged
on me once while the system was under high load).

I had previously been running mozilla native from CVS sources also about 48
hours old, with no problems.  Once I upgraded to the latest NetBSD code,
mozilla broke.  It appeared to be due to conflicts between GNU pth and
native pthreads.  So, I rebuilt all the dependency libraries from pkgsrc
(which is how I got to all native pthreads), removed pth, and
reconfigured/rebuilt mozilla - I did not update the mozilla sources.  Now
mozilla core dumps on startup before ever drawing anything to the display.
Here is the tail of the ktrace:

   860 mozilla-bin NAMI  "/usr/home/bryan/etc/src/mozilla/dist/bin/libxpcom.so"
   860 mozilla-bin RET   open 7
   860 mozilla-bin CALL  __fstat13(0x7,0xbfbff18c)
   860 mozilla-bin RET   __fstat13 0
   860 mozilla-bin CALL  mmap(0,0x1000,0x1,0x1,0x7,0,0)
   860 mozilla-bin RET   mmap 1213018112/0x484d3000
   860 mozilla-bin CALL  munmap(0x484d3000,0x1000)
   860 mozilla-bin RET   munmap 0
   860 mozilla-bin CALL  mmap(0,0x105000,0x5,0x2,0x7,0,0)
   860 mozilla-bin RET   mmap 1213018112/0x484d3000
   860 mozilla-bin CALL  mmap(0x485c8000,0x10000,0x3,0x12,0x7,0,0xf5000)
   860 mozilla-bin RET   mmap 1214021632/0x485c8000
   860 mozilla-bin CALL  mmap(0x485d8000,0,0x3,0x1012,0xffffffff,0,0)
   860 mozilla-bin RET   mmap 1214087168/0x485d8000
   860 mozilla-bin CALL  close(0x7)
   860 mozilla-bin RET   close 0
   860 mozilla-bin CALL  __sysctl(0x4808c008,0x2,0xbfbff3fc,0xbfbff3f8,0,0)
   860 mozilla-bin RET   __sysctl 0
   860 mozilla-bin PSIG  SIGSEGV SIG_DFL
   860 mozilla-bin NAMI  "mozilla-bin.core"

The backtrace looks like this:

   #5231 0x8075394 in NS_GetMemoryManager ()
   #5232 0x8074935 in GlueStartupMemory ()
   #5233 0x807518c in XPCOMGlueStartup ()
   #5234 0x80754d0 in GRE_Startup ()
   #5235 0x8058a06 in main ()
   #5236 0x8054254 in ___start ()

I'll have to reconfigure and rebuild with debugging enabled.  In the
mean-time, if there is anyone out there who knows if these bleeding-edge
sources can be stitched together before they heal a bit more, I'd love to
hear. :-)

-bp
--
# Software Engineer