[Python] Style question...

Sorry, but yes, I find that ugly. If you're just starting out with
Python, please just give it a shot without that approach for a while
and see whether you become much more comfortable *without* the token
than you ever were with it.

-Peter

Well, when in Rome I guess.

Having the closing token at the same indent level as the compound statement
is ugly anyway. :)

I do this myself at times, to help emacs. But I'd suggest you not
overdo it. It is more "pythonic" to simply leave the line blank (ie.
vertical whitespace), than have "pass" everywhere.

Pressing backspace once (to undo the auto-indent that emacs gives inside
loops) is less typing than 'pass'. And if you need to reindent large
chunks of code, you are better off using emacs block indent/dedent
feature than relying on "pass" as defacto block delimiter (this has been
my experience)

Also, look into the pindent.py script (included in Python's "Tools"
directory with the distribution), for another way of producing block
closing tokens.

--

Chad Netzer




From n/a Thu Aug 28 01:26:06 2003
From: n/a (Mike Thompson)
Date: Thu, 28 Aug 2003 09:26:06 +1000
Subject: My future Python IDE article
References: <mailman.1061920192.21278.python-list@python.org><3f4ca68c$0$4190$afc38c87@news.optusnet.com.au> <mailman.1061989090.23114.python-list@python.org>
Message-ID: <3f4d3e12$0$4190$afc38c87@news.optusnet.com.au>

"Gerhard H?ring" <gh at ghaering.de> wrote in message
news:mailman.1061989090.23114.python-list at python.org...
finding

That's not my experience. I've found Boa both stable, functional and well
priced.

I tried Wing and generally liked it, except that I'm used to an editor with
tabs for each open file and I found Wing's "one file at a time arrangement"
didn't match my way of working.

I tried Komodo, which has an enormous feature set that goes well beyond python,
which I liked until I came to run my first wxPython based program under the
debugger and it froze. I never did get to the bottom of why.

In between all this I attempted to use Eric but had difficulty getting it
setup. From memory, after some googling around, I found a few similar reports
and abandoned the effort.

I then tried Boa and haven't moved since.

My search has occurred over the last six months. About twelve months ago I also
tried the Secret Labs IDE (can't remember the name) which I found a bit of work
initially, but ended up quite liking. However this product now seems to have
been withdrawn from the market.

If you have to restrict it to 3, my suggestion would be:
Wing
Boa
IDLE

That would give you a mix of commerical and free. All are cross-platform.

If not IDLE, then Komodo which has a very inexpensive licence for personal
use..

"Anthony Roberts" <anthonyr-at-hotmail-dot-com at nospam.com> wrote in message
news:Uva3b.49430$la.826046 at news1.calgary.shaw.ca...

Hi. I tried to do something like that too, early on. (I recall being
frustrated to find I couldn't just alias "end = pass", and use 'end' as a
block delimiter... heh). I didn't actually want to use 'end' (I like
significant whitespace), I just saw a lot of posts with complaints about
"no-block-delimiter", and I thought, I wonder if you can make one. Well, you
can (of course), but not like that.

Anyway. I think the suggested idiom for people who must have a visible block
delimiter is "# end <keyword>". If you use this, I think emacs can "figure
it out" (I don't know, I don't use emacs, but I think pymode can handle
this), and you can also use Tools/scripts/pindent.py, if you like.

"from pindent.py"
# This file contains a class and a main program that perform three
# related (though complimentary) formatting operations on Python
# programs. When called as "pindent -c", it takes a valid Python
# program as input and outputs a version augmented with block-closing
# comments. When called as "pindent -d", it assumes its input is a
# Python program with block-closing comments and outputs a commentless
# version. When called as "pindent -r" it assumes its input is a
# Python program with block-closing comments but with its indentation
# messed up, and outputs a properly indented version.

[snip]

# Secret feature:
# - On input, a block may also be closed with an "end statement" --
# this is a block-closing comment without the '#' sign.

Hmm. Looks like you can also use "end", or perhaps 'end' . I don't know if
the <keyword> is optional or not, having never used this module.
Anyway, if you use this format, then, if someone doesn't like seeing all
that noise when they read your code, they can strip it out with
"pindent -d". Handy. And, when you read other peoples code, you can clutter
it up nicely with "pindent -c". Best of both worlds, really.

HTH
Sean

My emacs can't handle it... I think I'm going to subject myself to a bit
more of the Python way to see if I change my mind.

"Anthony Roberts" <anthonyr-at-hotmail-dot-com at nospam.com> wrote in message
news:Uva3b.49430$la.826046 at news1.calgary.shaw.ca...
Isn't there some kind of emacs plugin, macro or mode that
handles Python? I'm not an emacs person, but I've heard that
there is, and I would expect that it would handle things like
auto-indent for you.

John Roth

There is, but it can't tell if I want some stuff indented or not.

Sure it can. Hit tab or backspace.

I think the problem is that people learn to reindent chunks of code by
the following technique:

<arrow down>
<TAB>
<arrow down>
<TAB>
.
.
.


This has the unfortunate consequence of breaking your code when you hit
multiply indented loops, nested functions, function or loop boundaries,
etc.

However, by sticking 'pass' at the end of loops, emacs's python-mode
will assume a dedent and preserve the indenting at boundaries.

But, this technique will STILL fail on multiply nested loops, etc.
(emacs does it's best, but it cannot mind read) So it is a terrible
habit to get into, because it WILL cause you problems if relied upon
blindly.

So, it is better to NOT rely on the 'pass' crutch, and instead rely on
the python mode operations for reindenting blocks of code (which Eric
provided; surely this is becoming a FAQ)

So, to Anthony, if you are using the TAB key to indent blocks of code,
and relying on 'pass' to make it easier, just be aware that this
technique is NOT robust.

Agreed. If you're going to write code in a particular language, you
should fully embrace the language and its style, or move on to something
that you feel more comfortable with. These kind of crutches will only
serve to bite you later -- they aren't panaceas, and you may forget to
use them or be hobbled when interacting with someone else's code (where
you're not going to convince others to use your crutches). There's also
an indirect effect; if I'm reading someone else's Python code, and, say,
see someone using semicolons as statement delimiters on every line, I'm
not exactly going to be struck with a great deal of confidence.

It sits better with me if I have a good idea why I'm doing it.

Yup. In emacs, it is robust for C and C decendants. Indent region can do it
to entire files.

It's good for structural modifications when you want to change some outer
loop that ends up changing the indent/nesting level of some inner loop.

In Java, which is where I really started to lean of that feature heavily, it
comes with so *many* data structure classes that even though they mostly
share interfaces, you still come up with a cleaner way later on and go back
to change it. I'm finding Python is more a matter of "for x in
<whatever>:"...

But I still change my mind sometimes. :)

What's the keystroke for that?

C-c < and C-c >.

Thanks. :)

BTW: I've had your website bookmarked since 2001 because a number of the
things there are useful and cool.

:-).

It's worth doing "C-h m" and "C-c ?" and reading the results, too.

Cheers,
mwh

LOL!

I'm sorry. :)

Re: Anthony Roberts wrote:
Re: >
Re: > If I end indentation levels with "pass" statements, will I piss off people
Re: > that have to read my code? eg:
Re: >
Re: > for i in xrange(0,5):
Re: > if i:
Re: > print i
Re: > pass
Re: > print i * -1
Re: > pass
Re: >
Re: > I ask for two reasons... a) it helps emacs figure it out, and b) I'm more
Re: > comfortable ending compound statements with a token.
Re:
Re: Sorry, but yes, I find that ugly. If you're just starting out with
Re: Python, please just give it a shot without that approach for a while
Re: and see whether you become much more comfortable *without* the token
Re: than you ever were with it.
Re:
Re: -Peter
Re: --


Sorry, but no ;-) Ok it's "one line more", and
is not very useful. But...
For the moment I work on (a small part of) a project
that has hundreds of thousands of lines, almost
eveything in Python (and some in C and Fortran),
and when you have a loop several pages long, or
nested blocks to 8 levels or more, I may say
it's *very* convenient to see where the end
of a block exactly is. It's convenient to
have some long variable names too.

Sorry, but yes * 2. ;-) Please just don't tolerate loops several pages long
and/or nested blocks that deep - such things should occur very, very rarely
(essentially never). Putting an end-block delimiter after such beasts is like
putting a band-aid on top of a compound fracture.

-Dave

I use Scite, which provides a graphical means of representation
the indentation, with vertical lines that extend down at every
indent level. (It would have to be later in the day for me to
be coherent enough to describe it better.) Suffice to say that
it provides more than adequate (for me) visual indication of the
end of a block that stretches too far.

A better point I'd make however is that if you have a loop several
pages long, or a nested block 8 levels deep, the code needs
refactoring! Saying this occurs in a large project does nothing
to excuse it: the more code you have, the greater the need for
structuring it well.

-Peter

The only times I've seen the lack of a block delimiter as a problem is
when somebody plops some sample code in a usenet message and uses tabs
for indenting. OE strips out the leading tabs for some reason, so
everything winds up flush to the left margin. Once in awhile it is not
possible to recreate the proper indention, and for that reason I try
to include comment delimiters in code I send. When I don't forget, at
least. A better solution might be to physically restrain the posters
of tab-infested code from doing that, but there are only so many hours
in the day. And if OE wouldn't do what it does ... but then, where
would be the fund in that?

Well, actually, a better solution would be to skip OE for something
sensible, no? :)

OE is far from the only news user-agent that doesn't display tabs the
way you might like -- KDE's KNode has similar issues, for example.


Alex

On my own hook, I'm doing that, but here in the office I don't have
the choice. I do have the choice to spell "fun" correctly, though I
see I opted not to.

--
rzed




From tina_li23AThotmailDOTcom Fri Aug 29 17:19:41 2003
From: tina_li23AThotmailDOTcom (Tina Li)
Date: Fri, 29 Aug 2003 11:19:41 -0400
Subject: pmw menu item's 'command' attribute
References: <3f4e5687_4@corp.newsgroups.com> <bilrdk$6j5$04$1@news.t-online.com>
Message-ID: <3f4f6f5f$1_4@corp.newsgroups.com>

Thanks so much! Both the explanation and the code worked. Basically when a
lambda function is created, any references to global variables are not
evaluated until when it's executed. So if the global var changes value
before that the lambda function follows suit and takes the last value it has
settled at.

Thanks again!

Tina


"Peter Otten" <__peter__ at web.de> wrote in message
news:bilrdk$6j5$04$1 at news.t-online.com...


-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----