On 10/5/2014 2:36 PM, Ken Moffat wrote:
On Sun, Oct 05, 2014 at 01:44:55PM -0400, Michael Shell wrote:
On Sun, 05 Oct 2014 07:59:46 -0400
Alan Feuerbacher <[email protected]> wrote:
"Fatal Python error: Invalid thread state for this thread . . ."
Google found this:
http://lists.freedesktop.org/archives/libreoffice/2013-April/050229.html
. . .
It seems the debug option --enable-dbgutil can cause it by detecting
An interesting link, there is a similar thread (initially for build
failures on windoze, I think, but it apparently affected OSX "make
check" related to their use of clang) at
http://lists.freedesktop.org/archives/libreoffice/2013-April/050068.html
Alan, are you using clang ?
Well, it was installed along with LLVM-3.5.0 .
The problem, of course (which Michael might not be aware of) is
that building LO takes for ever.
Well, the better part of two hours for me. :-)
Perhaps using strace to run LO might eventually give some pointers.
I suspect it will take some time to find the correct parms for strace
(there are several programs invoked : lowriter or whatever is just a
script to invoke 'soffice' which is itself a script. /me looks
at the soffice script:
Perhaps invoking lowriter from the commandline as
$ lowriter --writer --strace
might provide strace.log in $PWD
As with _any_ graphical program, it will probably be an enormous
log, perhaps best read by starting at the end.
I may do the above, given that I haven't really figured out why changing
the switch --enable-python=system to --enable-python=fully-internal
seems to have solved the problem.
Alan
--
http://lists.linuxfromscratch.org/listinfo/blfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
