Path: csiph.com!x330-a1.tempe.blueboxinc.net!usenet.pasdenom.info!gegeweb.org!de-l.enfer-du-nord.net!feeder2.enfer-du-nord.net!feeder.news-service.com!newsfeed.xs4all.nl!newsfeed6.news.xs4all.nl!xs4all!post.news.xs4all.nl!not-for-mail
Date: Thu, 01 Sep 2011 22:02:54 -0400
From: Sahil Tandon <sahil@FreeBSD.org>
User-Agent: Postbox Express 1.0.1 (Macintosh/20100705)
MIME-Version: 1.0
To: python-list@python.org
Subject: Re: idiomatic analogue of Perl's: while (<>) { ... }
References: <mailman.644.1314853527.27778.python-list@python.org> <4e5f2010$0$29987$c3e8da3$5496439d@news.astraweb.com>
In-Reply-To: <4e5f2010$0$29987$c3e8da3$5496439d@news.astraweb.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: list
Newsgroups: comp.lang.python
Message-ID: <mailman.678.1314928995.27778.python-list@python.org>
Lines: 57
NNTP-Posting-Host: 2001:888:2000:d::a6
Xref: x330-a1.tempe.blueboxinc.net comp.lang.python:12600

[Thanks to everyone who responded]

Steven D'Aprano wrote:
> On Thu, 1 Sep 2011 02:56 pm Sahil Tandon wrote:
>> %%
>> # unbuffer STDOUT
>> sys.stdout = os.fdopen(sys.stdout.fileno(), 'w', 0)
>
> I've never bothered with unbuffered stdout, but that looks fine to me.
>
> I'm not sure if it is necessary though, because print seems to automatically
> flush the buffer after each line in my testing. Unless you're printing
> repeatedly to the same line, I'm not sure unbuffered stdout is helpful.

I found it necessary because without reopening sys.stdout with buffering 
explicitly turned off, I would have to manually flush the buffer after 
each print.  This is because the program must reply (via writing to 
STDOUT) after parsing each line read via STDIN.  If I neither disable 
buffering nor manually flush after each print, the program just hangs 
instead of printing right away.

>> # process input, line-by-line, and print responses after parsing input
>> while 1:
>>    rval = parse(raw_input())
>>    if rval == None:
>>      print('foo')
>>    else:
>>      print('bar')
>> %%
>
> "while True" is considered slightly more idiomatic (readable), but
> otherwise, that seems fine.

Ah, thanks -- I've changed '1' to 'True'.

>> This works, but while reading the documentation, I thought of using 'for
>> line in fileinput.input()' in lieu of 'while 1:' construct.  This does
>> not work when debugging the program on the command line -- the script
>> appears to just hang no matter what is typed into STDIN.  I believe this
>> is because of some internal buffering when using fileinput.  Is there a
>> recommended way to disable such buffering?  Am I taking a totally wrong
>> approach?
>
> I'm not sure anything about fileinput is exactly *recommended*, it's kinda
> discouraged on account of being a bit slow. See help(fileinput) at the
> interactive prompt.
>
> For what it's worth, the default buffersize for fileinput.input is 0, so if
> that doesn't do what you want, I don't think fileinput is the right
> solution.

Got it.  Based on your and others' response, I will stick with my 
existing approach.

-- 
Sahil Tandon <sahil@FreeBSD.org>