On Thu, Mar 3, 2011 at 11:49 PM, Steve Borho <st...@borho.org> wrote: > On Thu, Mar 3, 2011 at 2:08 AM, TK Soh <teekay...@gmail.com> wrote: >> On Wed, Oct 21, 2009 at 6:31 AM, Martin Geisler <m...@lazybytes.net> wrote: >>> Steve Borho <st...@borho.org> writes: >>> >>>> On Tue, Oct 20, 2009 at 3:55 PM, Greg Ward <g...@gerg.ca> wrote: >>>>> On Mon, Oct 19, 2009 at 8:46 PM, Steve Borho <st...@borho.org> wrote: >>>>>>> BTW, it looks like someone else has figured out a workaround for this >>>>>>> bug: >>>>>>> >>>>>>> http://hg.clusterlabs.org/pacemaker/pygui/rev/460feb8039c1 >>>>>>> >>>>>>> but that sure looks painful and awkward. >>>>>> >>>>>> Oh, I'm certain we have worse workarounds in gtklib.py. But I'm glad >>>>>> you've found a cleaner solution. >>>>> >>>>> Shortly after I sent that message, it occurred to me that a good >>>>> compromise might be >>>>> >>>>> if pygobject version <= 2.12.1: >>>>> raise ???("incompatible version of pygobject") >>>>> >>>>> At least that way, the next poor slob who tries to run TortoiseHg on >>>>> RHEL/CentOS 5 will get a clear error message and not have to spend 2 >>>>> hours doing failed bisects and google'ing the error message. They'll >>>>> still have to upgrade pygobject, though. ;-) >>>> >>>> Good idea. I added a check next to a comment that references this mail >>>> thread. >>> >>> Uhm, I'm using Red Hat Enterprise Linux Client release 5.4 (Tikanga) at >>> my university so this would effect me and the guys I've convinced to use >>> Mercurial there :-( >>> >>> If >>> >>> http://hg.clusterlabs.org/pacemaker/pygui/rev/460feb8039c1 >>> >>> really fixes it, then I would like to see it used. Note that the >>> pygobject_newer function seems much too complex. It should really be: >>> >>> def pygobject_newer(args*): >>> return gobject.pygobject_version > args >>> >>> I'm sorry I haven't fixed this myself -- I only saw the error today. >> >> Sorry to bring back an old thread, but is this still an issue? I try >> to install thg 1.1.8 on our recently acquired REL5 box and got hit by >> it. > > We added a check in gtklib.py for gobject version some time ago, but > that's the last I heard of it.
I wonder if workaround suggested by Martin would fix it, or has it been tried and failed. It's very difficult, if at all possible, for me to upgrade the pygtk/pygobject version on those REL5 boxes, since they are part of production machines. ------------------------------------------------------------------------------ What You Don't Know About Data Connectivity CAN Hurt You This paper provides an overview of data connectivity, details its effect on application quality, and explores various alternative solutions. http://p.sf.net/sfu/progress-d2d _______________________________________________ Tortoisehg-develop mailing list Tortoisehg-develop@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tortoisehg-develop
