PythonWin IDE sucks!

[Alex Martelli]

"Dale Strickland-Clark" <dale at out-think.NOSPAMco.uk> wrote in message
news:7tuivs46osods22uumpvsvsap2jocrd4si at 4ax.com...
> I've sort of mentioned this before and all I got was general murmors
> of agreement but no solution.
>
> This is such a huge irritation and time waster for me that if it isn't
> sorted out I can't really consider Python for future large projects.

Given that your issues appear to be solely with the IDE, why don't
you propose to change development environments, but, rather,
_languages_?  Seems a non-sequitur if I ever saw one.

Previous grumblings of yours that I remember seeing about this
IDE appeared to be related to completely different issues -- hangs
in the IDE due to broken code, as I recall.  Those are architectural
issues -- similar to those which affect the VB IDE, by the way: as
it runs the code under development in the same process as the
IDE, if said code does something sufficiently disruptive (and bugs
will cause that, sometimes...), then the IDE itself hangs or
crashes.  (I consider this "proof" that an IDE must run the code
being developed under _another_, separate process -- but that
is quite another issue).


> To test a module, I import it into PythonWin using the Import/Reload
> button and then instantiate a class in the Interactive Window.
>
> After a failure, I fix the offending code and Import/Reload again.
> However, the same failure occurs showing the original line with it's
> original contents. The changed module hasn't been loaded.
>
> Why is this?

Can you reproduce this behavior "in the small"?  When this happens
to me, it tends to be a case of the module _not_ having been reloaded,
due to (e.g.) syntax errors, and me missing the feedback about it in
the statusbar.  If you have a sourcefile that reproducibly displays
this behavior (the smaller the better, of course), then it should not
be _too_ hard to diagnose the cause (and the fix is likely to be easy
once the cause is diagnosed).


> What can I do about it?

Get a reproducible bug.  Diagnosis and fix will follow -- I'll gladly
lend a hand myself, and I'm sure Hammond and others at
ActiveState are even more disposed to.  But dealing with bugs
that are not reproducible (and which _might_ be user errors...)
is just too hard.


> The only way to fix it is to quite and reload PythonWin but then you
> have to arrange windows, reestablish breakpoints and watches, etc.
> and it all takes far too long.

A consequential feature-request might be to give PythonWin a
"save-all-IDE-state"/"reload-all-IDE-state" functionality; that
would not only diminish this issue, but come in handy on many
other occasions.  Have you submitted such a feature request?


Alex