Re: (A) native Windows port

On Wed, 2002-07-10 at 19:56, Lamar Owen wrote:
> On Wednesday 10 July 2002 11:48 am, Hannu Krosing wrote:
> > On Wed, 2002-07-10 at 16:20, Lamar Owen wrote:
> > > On Wednesday 10 July 2002 09:11 am, Hannu Krosing wrote:
> > > > And I have written custom postgres table dumpers in python without too
> > > > much effort (except reverse-engineering the page structure ;) for both
> > > > 6.x and 7.x database tables, so we could actually use python here too.
>
> > > I'm willing to look into this.  However, the dump still has to be pulled
> > > with a standalone backend -- no networking availability can be assumed.
>
> > Actually it works on raw table file ;)
>
> > the script is meant for quick and dirty resque operations, and requires
> > that one writes their own data-field extractor code. I have used it
> > mainly to resurrect accidentally deleted data.
>
> > it is for 7.x style pagefile layout
>
> Hmmm.  This is interesting stuff.  I'll have to take a look at it once I'm
> finished re-learning Fortran 77 for a project I'm doing (34MB of DEC Fortran
> source that g77 doesn't like very well) for work.  I have a hard time
> switching language gears.  Particularly the Fortran 77 -> Python gear... :-)
> Although at least the fixed-form paradigm stays there in the transition. :-)
> It's been a very long time since I've done Fortran of this complexity.
> Actually, I've never done Fortran of _this_ complexity -- this is serious
> number-crunching stuff that uses all manners of higher math (tensors, even).
> There is no direct C equivalent to some of the stuff this code is doing --
> which is part of the reason g77 is having problems.  But I digress.

Once you understand what te code is doing you can port it to python
using Numerical Python (http://www.pfdubois.com/numpy/) and/or
Scientific Python (http://starship.python.net/~hinsen/ScientificPython/)
to get a head-start in total conversion to python ;)

You may even try using F2PY ­ Fortran to Python Interface Generator
(http://cens.ioc.ee/projects/f2py2e/).
> Getting the %pre scriptlet to use a non-sh interpreter is undocumented, but
> not hard.  :-) (actually, I stumbled upon it by accident one time -- that
> time it was a bug....)  Now to see if it can be done consistently in both the
> anaconda chroot as well as a standard rpm command line invocation.

Actually, if the python dumper can be made to work somewhat reiably it
can be run after install/upgrade without too much trouble.

--------------
Hannu