Programming

Changes that caught my eye in Python 3.2

Samat Jain — Wed, 23 Feb 2011 01:30:03 +0000

Python 3.2 was released a few days ago. Reading the What’s New document, some of the stuff that caught my eye…

argparse is a new command-line option parsing module that replaces optparse. I found optparse, now deprecated, a pain to use—argparse looks a lot better. argparse can be used in older programs today with a third-party module (E.g. argparse Debian package). I actively avoided adding command-line parsing to my applications because optparse was such a pain—something now much less of a problem.

The new concurrent.futures module seems to be a port of Java’s java.util.concurrent package. I haven’t looked into it yet, but with the GIL, how many people really care about Python threads? The simple examples given can easily be done with Python’s multiprocessing module, for both processes and threads, though without the Java paradigm. The threading module also has a new threading.Barrier class.

WSGI for Python 3 appears to be finalized. I’m looking forward to finally porting my WSGI web applications to Python 3. The email module also received the same makeover that WSGI needed for Python 3.

ElementTree has been updated to v1.3. Not that I care—lxml’s namespace handling is much better (which is to say, it’s just terrible as opposed to near unusable, as is with pure ElementTree).

Quite possibly the most badass addition to Python 3.2: the functools.lru_cache decorator. I use this pattern all the time, but have never gotten around to making it a generic class or decorator I could reuse across applications. With Python 3.2, I no longer need to. The idea: decorate a slow-performing function (where it makes sense) with functools.lru_cache, and Python will instantly and easily memoize the return values of that function making subsequent calls to the function faster.

html.escape() is a simple function that escapes HTML markup for you with the appropriate HTML/XML entities (I forget how I used to do this usually; something in one of the CGI utils, I think? Or maybe stuff in my template engine…).

This is just from a quick reading of the release notes—it’s very likely I missed something. What’d I miss? And what do you like that’s new in Python 3.2?

Topic:

Programming

Python

Python-like tuple unpacking for PHP

Samat Jain — Wed, 29 Oct 2008 21:55:47 +0000

Python provides a neat way for functions to return multiple arguments via “tuple unpacking”. For example:

def blah:
  return ('one', 'two')

rval_1, rval_2 = blah()

The same can be done in PHP relatively easily via the list construct:

function blah()
{
  return array('one', 'two'); }

list($rval_1, $rval_2) = blah();

Topic:

Programming

Python

PHP

jQuery: the new defacto Javascript web framework

Samat Jain — Thu, 02 Oct 2008 16:38:47 +0000

News from a couple days ago: both Microsoft and Nokia are now including the jQuery Javascript framework as part of their development kits. That is: jQuery will be part of Microsoft’s ASP.NET AJAX framework and be available for use in applications written for ASP.NET; and jQuery will also be distributed on millions of Nokia phones.

Defacto standards, I believe, are a good way to inform the development of real standards. Standards developed the other way around, at least in the tech industry, have had a habit of taking a very long time to reach end consumers… for example, how many decades has it taken for your average web user to gain access to a fully CSS2-compliant web browser? How many more decades will it take for OASIS’s OpenDocument format to supplant Microsoft Word and its *.doc files?

Hopefully, this is the beginning of a path that will lead to jQuery’s inclusion into the Javascript language, as well as initiatives that will improve jQuery’s performance.

I like the fact that Microsoft and Nokia are not trying to reinvent the wheel, and roll their own Javascript frameworks. Sun did this with Java Server Faces. A frequent lament with JSF is that it’s nearly impossible to customize any of the widgets. There is too much complex, custom Javascript, and the adoption of the frameworks used makes figuring out how to work with them difficult.

Also, as others have noted, this is the first time Microsoft itself is distributing an open-source project with one of their products. A sign of things to come?

Topic:

Python 2.5 alpha released

Samat Jain — Thu, 06 Apr 2006 00:00:00 +0000

Python 2.5’s first alpha has been released. From a quick look at the What’s New in Python 2.5 document, what I thought was interesting…

Topic:

Programming

Python

Computer Organization and Design For More Practice Problem 2.28

Samat Jain — Thu, 16 Feb 2006 08:37:10 +0000

I’m using Computer Organization and Design, 3rd ed (David Patterson and John Hennessy) for an electrical engineering course (that I have to take to graduate) in computer systems architecture. Assigned as homework, from the For More Practice section, is problem 2.28:

Topic:

Programming

Academia

PHP and passing non-variables by reference

Samat Jain — Wed, 14 Sep 2005 06:03:00 +0000

In PHP 5.0.5, they’ve now made it a requirement that things that are passed by reference must now explicitly be a variable. You would think this kind of behavior is obvious, but apparently it’s been allowed for all versions of PHP previous. Appararently without even warnings. You’ll get an error:

Fatal error: Only variables can be passed by reference ...

So you may think, where is this useful? Consider something short and concise like:

$only_element_we_care_about = array_pop(explode($seperator, $string));

You cannot do this now. The return value of a function is not a “variable” and cannot be passed by reference; a temporary variable needs to be used instead:

$tempory_variable = explode($seperator, $string); $only_element_we_care_about = array_pop($temporary_variable);

Yes, there are other ways to do the above, but that isn’t the point. The fix is not difficult, but it is a total complete pain to go back to legacy code and fix things like this.

Remind me to find another web programming language.

Topic:

Programming

PHP

Looking at Python web development frameworks

Samat Jain — Mon, 25 Jul 2005 20:05:00 +0000

I’ve a lot of interactive websites I’ve been wanting to do lately. After learning Python earlier this year, I’m wanting to write all these projects in Python versus something conventional like PHP. I’ve come to the conclusion that web development with Python sucks; it’s no wonder it’s not being used very much. There are too many frameworks, many doing similar things in different ways, or doing totally dissimilar things in totally different ways. I’ve been spending the last week reviewing several different frameworks and deciding which ones I thought were best.

Topic:

Programming

Python

Today's g++/C++ funny: ofstream's open constructor

Samat Jain — Thu, 07 Jul 2005 22:43:00 +0000

So, I’m working on a project, and have some code like this:

string blah="somefile"; ofstream fp;

fp.open(blah);

And it doesn’t work. Apparently, what you need to do when using ofstream’s open method:

fp.open(blah.c_str());

Yes, it looks like what it is. Apparently g++’s iostreams library (gcc 3.2) does not support things that have been part of the C++ standard for years–like C++ STL strings… You have to convert back to a C string.

Last I checked this was 2005, not 1999. You’d have thought someone would have overloaded this by now… I’ve only checked this in g++; will check with Intel’s C++ compiler and Microsoft Visual Studio later, though I’m not sure whether to expect better. What didn’t help me to find this out, of course, is g++’s totally wonderful (sarcasm) compiler error messages…

Topic:

Programming

C++

Splitting a Sequence into Subsequences with Python

Samat Jain — Sun, 05 Jun 2005 22:50:00 +0000

I’m writing a load balancing algorithm for this cluster I’m working on, and have a bunch of things in a Python list sequence. So came the issue, how to divide the list into sublists for each node in the cluster to compute? I came up with the algorithm:

def split_seq_stride(l, n):
  """Splits a sequence l into a list of subsequences containing at least n
  elements each, not preserving order. The first few subsequences may contain
  n+1 elements, containing the last few elements. (Algorithm is good for load
  balancing)"""
  r=[]
  k=len(l)/n
  [r.append(l[i::k]) for i in range(k)]
  return r

which behaves like:




  
    
      l = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
      print split_seq_stride(l, 3)
      [[1, 4, 7, 10], [2, 5, 8], [3, 6, 9]]

Assuming each task is an equal amount of work, this algorithm is a good load balancing algorithm. it prevents any node from having to do any less significant work than the others.

Probably useful for someone: what about splitting up a sequence in-order, with the last subsequence containing fewer elements? This code segment does just that:


def split_seq(l, n):
  """Splits a sequence l into a list of subsequences containing at most n
  elements each, preserving order. The last subsequence may contain less
  than n elements."""
  r=[]
  [r.append(l[s:s+n]) for s in range(0, len(l), n)]
  return r

which behaves like:




  
    
      l = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
      print split_seq(l, 3)
      [[1, 2, 3], [4, 5, 6], [7, 8, 9], [10]]

Thanks to yason and others on EFnet’s #python for pointers.

Topic:

Programming

Python

High Performance Computing

The semantics of a fork, followed by setsid, and another fork

Samat Jain — Thu, 02 Jun 2005 22:34:00 +0000

So I came across a piece of code that looked like this:

if (fork() != 0)
  exit(0);
setsid();
if (fork() != 0)
  exit(0);

What each individual system call can easily be looked up and understood (well, maybe not setsid), but what the entire block of code is doing is not clear at all

…and I really can’t say I understand the mechanics of it all to explain myself. But this python code segment’s comments does a good job of what the entire block of code is trying to do—daemonize a process completely.

Topic:

Programming