Re: [Visualpython-users] WinXP and numpy 1.3

[Gary P. <gar...@gm...>]

We should probably confirm that 1.3 does break visual before proceeding.
I'm surprised that a change from 1.3 -> 1.4 would cause breakage... the API
is supposed to be the same.

But wait ... I now recall that there is a discussion going on about numpy
1.4 ... evidently it actually breaks the ABI, the application binary
interface.  There have been over 200 posts on the subject on the numpy
discussion list. The break comes in date/time support, so I wouldn't think
it would affect visual.  I glanced at some of the posts, and they talk as if
1.4 has been released, but I can't find it.  Maybe they've withdrawn it.

On the one hand, this will bite only people who  install the current numpy
after installing the current visual *and* are on Windows *and* are novice
enough not to figure out how to fix their problem.  Most people who install
numpy independently from visual will probably be able to figure things out
as I did.

On the other hand, you eliminate even that small set of people by
recompiling.  That set might have zero people in it. But to be entirely neat
and clean, one would not want visual to break if one installs the current
numpy.  By the way, AFAICT, the current version of numpy is 1.3.  ... if
anyone sees something I've missed, and the current version is indeed 1.4,
please let me know.

Overall, I'd say it's not critical, but some day when you have time it would
be better to recompile.  But someone in addition to me should try to install
the latest numpy 1.3 on top of the latest visual on WinXP and see if it
breaks.  Just in case the problem is idiosyncratic to me.

-g

On Mon, Feb 15, 2010 at 12:04 PM, Bruce Sherwood <Bru...@nc...>wrote:

>  I would guess (without actually trying it) that yes, Visual would surely
> work with numpy 1.3 if (but only if) Visual had been compiled with numpy 1.3
> present. You make me realize that I've been too casual in what numpy I
> install before building on Windows. I've always just used the most up to
> date version of numpy, without checking to see what version of numpy is
> being used by scipy. Of course there may be yet another major application
> that wants even a different numpy, but surely scipy is the most likely
> accompaniment to Visual.
>
> Should I rebuild the Windows installer using numpy 1.3?
>
> Bruce Sherwood
>
> Gary Pajer wrote:
>
> I installed VPython on a new WinXP machine.  First I installed python, then
> visual, and all was good.
>
> Then I went about and installed the other stuff I use ... numpy, scipy,
> matplotlib ...
>
> And VPython crashed python at "import visual"
>
> Then I discovered that the VPython installer installs numpy 1.4, but the
> latest version on numpy.scipy.org is numpy 1.3.  So I had unknowingly
> downgraded numpy.  So I uninstalled numpy and visual (and python for good
> measure) and reinstalled, and it works again.
>
> Is it possible (or already known) that visual does not work with numpy 1.3,
> the current release?   I'm hesitant to try it myself.  Everything is working
> now, but I was surprised, and I thought I might document this potential
> gotcha.
>
> -gary
>
> ------------------------------
>
> ------------------------------------------------------------------------------
> SOLARIS 10 is the OS for Data Centers - provides features such as DTrace,
> Predictive Self Healing and Award Winning ZFS. Get Solaris 10 NOWhttp://p.sf.net/sfu/solaris-dev2dev
>
> ------------------------------
>
> _______________________________________________
> Visualpython-users mailing lis...@li...://lists.sourceforge.net/lists/listinfo/visualpython-users
>
>
>
> ------------------------------------------------------------------------------
> SOLARIS 10 is the OS for Data Centers - provides features such as DTrace,
> Predictive Self Healing and Award Winning ZFS. Get Solaris 10 NOW
> http://p.sf.net/sfu/solaris-dev2dev
> _______________________________________________
> Visualpython-users mailing list
> Vis...@li...
> https://lists.sourceforge.net/lists/listinfo/visualpython-users
>
>