On Mon, 26 Mar 2007, Troy Dawson wrote:

Hi Jon,
I don't know if this matters or not, but these are the changes we make to the SL4 thunderbird source rpm in order for it to compile and work on SL3.

-#sw start
-sed -i 's/enable-pango/disable-pango/' .mozconfig
-#sw end

So, you see that we only disable pango when it's being compiled. But that doesn't look like the problem you are having.

Indeed it doesn't seem very likely does it!

Are you sure that the SL3 machine has it's timezone and glibc up to date?

I tested it on another SL3 machine and got the same results (crashing in the same place), on that I have:

$ rpm -q tzdata glibc seamonkey seamonkey-nss thunderbird
tzdata-2006m-3.el3
glibc-2.3.2-95.44
seamonkey-1.0.8-0.2.SL3
seamonkey-nss-1.0.8-0.2.SL3
thunderbird-1.5.0.10-0.1.SL3

which I believe are current for SL3.0.8 though I suppose I might have missed something. On the SL4 system we have:

$ rpm -q tzdata glibc seamonkey seamonkey-nss thunderbird
tzdata-2006m-3.el4
glibc-2.3.4-2.25
seamonkey-1.0.8-0.2.el4
seamonkey-nss-1.0.8-0.2.el4
thunderbird-1.5.0.10-0.1.el4

btw both the SL3 and SL4 machines are i386 arch since I forgot to say before.

Also, does he have the new seamonkey that just came out last week. The one before it broke evolution, gaim, and potentionally other things, maybe by some freak chance, it broke lightning as well.

Yup we have the new seamonkey{-nss,} etc, setting LD_ASSUME_KERNEL also makes no difference so it seems unlikely to be an issue with the nptl libraries.

Troy

Jon Peatfield wrote:
 One of our users tried installing lightning 0.3.1 in thunderbird 1.5.0.10
 on our SL3 systems and it apperently installs ok but then won't let
 thunderbird start up properly.  On SL4 systems (using the same profile) it
 works ok.

  SL4 system has:  thunderbird-1.5.0.10-0.1.el4
  SL3 system has:  thunderbird-1.5.0.10-0.1.SL3

 The lightning installed was the one from
 
https://addons.mozilla.org/en-US/thunderbird/downloads/file/8814/lightning-0.3.1-tb-linux.xpi

 Starting with thunderbird -g (to run under a debugger), gives me:

 Program received signal SIGSEGV, Segmentation fault.
 [Switching to Thread 1073855744 (LWP 31316)]
 0x003ccc74 in nsACString_internal::~nsACString_internal$base () from
 /usr/lib/thunderbird-1.5.0.10/libxpcom_core.so

 entering 'where', shows that it is crashing while trying to do something
 with timezones...

 (gdb) where
 #0  0x003ccc74 in nsACString_internal::~nsACString_internal$base () from
 /usr/lib/thunderbird-1.5.0.10/libxpcom_core.so
 #1  0x05586b1e in calICSService::GetTimezoneIds (this=0x9092400,
 aTzids=0xbfff8ac4) at
 ../../../dist/include/string/nsTDependentSubstring.h:62
 ...

 so it seems to be unhappy with the timezone info in a way which cases a
 segv, but this doesn't happen on SL4 systems.

 Does anyone else see this or is it something odd about our setup?
 sunbird/lightning has always had weirdness with timezones (I believe that
 it tries to do all the work itself which is just stupid on platforms with
 zoneinfo),

 Apparently a binary of thunderbird downloaded from mozilla doesn't show
 this effect -- the user tells me, but he didn't say which version that was
 so it might not be directly comparable.

 Of course TUV doesn't support thunderbird on EL3 so it might be something
 SL3 specifc that isn't as expected or just as easily it might be some
 problem in the way that *our* systems are set up (e.g. brokenness in
 timezones etc)...

 Sadly the user didn't tell me if this just started happening or not!

   -- Jon

--
Jon Peatfield,  Computer Officer,  DAMTP,  University of Cambridge
Mail:  [EMAIL PROTECTED]     Web:  http://www.damtp.cam.ac.uk/

Reply via email to