John H Palmieri wrote:
On Feb 3, 11:33 pm, Dr David Kirkby <drkir...@gmail.com> wrote:
Than you for finding this error John. It's strange, I never se this on
Solaris myself, though William says he has seen it before on Linux.
IHopefully, this bug can be squashed, though it is not as serious as
the one which is stopping Sage building on Solaris. I could image I
could have wasted a lot of time on this, if it was intermittant, but
under the false impression is was introduced under one of the 5 or so
tickets merged, though as William says, this is perhaps a new
invocation of it. Anyway,once my build has finished, I'll see whether
I get the error, then try your patch.
Here's the situation, as far as I understand it: in versions from
around August 2009 up to and including 4.3.0, the first time you would
execute a graph theory command, say graphs.CompleteGraph(2), it would
call kpsewhich. If you didn't have a tex installation, it would
probably print the error message. (I always have tex installed, and I
rarely use graph theory, so I wasn't going to see this unless I made a
real effort.) Version 4.3.1 (and also 4.3.0.1.alpha3) included a patch
which ended up calling graphs.CompleteGraph(2) when Sage started up
(as the default value of an optional argument of a function imported
in sage.all), and this is what was going on with the t2 build.
Somewhere in the 4.3.2 development, a patch got put in that deferred
the call to graphs.... until needed, so the message went away until
you ran a graph theory function. That's why I didn't see it when I
tried reproducing it with 4.3.2.alpha2 on my Mac.
--
John
Thank you John. Your explanation makes perfect sence. Although William has seen
this occasoinally on Linux, like Minh, I'm seeing it every time Sage starts on
Solaris. The fact it is now executed at startup would explain this. The fix
should solve it, though I have not tried the patch. It is a trivial problem in
comparison to the more pressing issue of what broken the Solaris build on 4.3.1.
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
