Is Python any good with MySQL?

Alex Martelli aleaxit at yahoo.com
Tue Nov 7 09:04:41 EST 2000


"Keith MacDonald" <keith at nojunk.textpad.com> wrote in message
news:YoEN5.6830$Fi.23379 at NewsReader...
> Skip,
>
> Thanks for the suggestions.
>
> The problem is that the link command includes "/EXPORT:init_mysqlmodule",
> but I can't figure out where that code is emitted.  It does not seem to
have
> anything to do with the files that are part of the MySQLdb distribution.

That looks like something that the Distutils might emit, if (perhaps
wrongly) informed that '_mysqlmodule' is the modulename of the C they're
compiling-and-linking.


> This is really the essence of the problem.  Python has an enthusiastic
> following, who presumably know it inside out, but it just keeps throwing
up
> hurdles at newcomers.

Very, very few of us, Python's, "enthusiastic followers" `know it inside
out`, particularly if "it" is taken to include (as you seem to repeatedly
imply) _every_ possible module that somebody out there may have released
for Python.  For example, I've never had any use on Earth for MySql, so
I have absolutely no idea of what's up with the various ways to interface
Python with MySql, nor do I intend to look it up in the near future (nor,
possibly, ever).  Ditto for Sybase, Oracle, and a zillion other systems
that somebody has interfaced to Python.

I do have some (incomplete and imperfect) knowledge of possible interfaces
to databases which have been of my specific interest in my short Python
career so far (Gadfly, Microsoft Jet, Microsoft SQL Server/MSDE), and
that stands me in good stead, so far -- basically, I've seen that I can
use ADO for my specific DB-interfacing needs (unless I want to use
Gadfly, which uses a DB-API compliant interface, more or less), and
that is sufficient to me for now.


> There are various ways of judging success, but in
> this context, Python will only be really successful if new users can just
> pick it up and run with it.

In some ways, surely.  For example, all potential 'new users' who
are basically interested in the Windows platform can do that -- ADO
and other COM-based technologies are what they'd be using in Visual
Basic, Python lets them keep doing that, they're happy & running.

If MySql-on-Windows is as crucially important, as hugely widespread,
as to be the typical requirement of a "new user" (AND said "new
user" can't/won't just download and use the ready-made binary
package which I've already repeatedly pointed to -- or does _that_
have problems, too?  In the latter case, FAQTS would no doubt
like to learn about it!), no doubt ActiveState will 'adopt' some
solution for their pre-cooked distro (as you say they've solved your
problems by doing for Perl), and there you are.

What has "Python" gotta do with ActiveState's commercial decisions...?

>From my vantage point, very myopic no doubt, "being able to recompile
the interface to MySql on Windows" seems an issue that defines a rather
small target-audience -- although, no doubt (though I've never seen
an explanation of why you can't just use the binary?), it must be
a very important issue to you.  But surely you can see it's utterly
ridiculous to keep claiming, as you do, that this very specific
interfacing-module-recompilation-issue means you're "wasting time
making [Python] work at all'.  Puh-leeze.  Python IS working; your
time, if being wasted, it's being wasted in trying to recompile a
given, very specific, C-written interfacing-module (apparently, not
well packaged by its author/distributor/maintainers/...?).


It will no doubt make everybody's life easier if a vast array of
pre-built and/or ready-to-rebuild modules is put together, sure.  A
'CPAN-equivalent'... it's ofteen been wished-for, nor will I say
I wouldn't enjoy it myself just as anybody else!  And/or if
ActiveState will pick up more modules in (or, ideally, _on the
side of_:-) its distribution[s] of Python, as they do for Perl.

Just like having beautiful books (we now do), nice-looking icons
and logos (ditto), free or cheap webspace suppliers supporting
the use of Python for users' CGIs (not quite as mature a situation,
sigh), killer applications (currently Zope:-), CP4E, celebrity
testimonials, unit-test setups, standardized distribution utilities,
refactoring browsers, point-and-grunt GUI designers ... all kinds
of nice-to-have, up to can't-live-without-it, marginalia...

...but none of these is "the" one and only crux!-)


Alex






More information about the Python-list mailing list