Saturday, September 29, 2012

Python 3.3 is my Favorite Python Release

Today, Python 3.3 was released. During the 4.5 years I've been a CPython core developer, 6 major Python releases (2.6, 2.7, 3.0, 3.1, 3.2, and 3.3) have past by me. In this post, I will explain why 3.3 is the most exciting Python release to me. I will be cherrypicking, consult "What's New in Python 3.3" and the Misc/NEWS file for complete details.

Unicode

PEP 393 completely changed the internal format of Python's Unicode implementation. It does away with the concept of wide and narrow unicode builds. The encoding of a string now depends on its maximum codepoint; there are 1-byte, 2-byte, or 4-byte strings internally. This means, for example, that strings with only ASCII characters can be represented in their most compact format. Partially as a consequence, Unicode standard compilance has improved. Indexing strings always gives code points not surrogates like on < 3.3 narrow builds. str.lower(), str.upper(), and str.title() have been fixed to use full Unicode case-mappings instead of the simple 1-1 ones. The str.casefold method implements the Unicode casefolding algorithm.
If the gods of PyCon talk selection smile on me, I will be giving a talk about this and the history of Unicode in Python.

Glorious Return of the "u" Prefix

Python 3.3 allows the u in front of strings again. Since the b prefix is supported from Python 2.6, code which wants to support 2.x and 3.3 shouldn't need to use unpleasant kludges like six's u() and b() functions. I don't think it would be unreasonable for libraries to only support 2.7 and 3.3+ now just to have the more natural string syntaxes.

Many Nice Things

One of the annoyances in previous Python 3 versions was it was impossible to turn off PEP 3134's implicit exception chaining. The raise exc from None syntax introduced in 3.3 prevents the __context__ of an exception from being printed.
There were improvements in exceptions themselves. PEP 3151 merged IOError, OSError, WindowsError, and various error types in the standard library. It also created a hierarchy of specialized exception subclasses. This means that most code dealing with IO errors won't have to dig into the errno module. For example, this standard pattern
try:
    fp = open("data", "rb")
except OSError as e:
    if e.errno != errno.ENOENT:
        raise
    # Create file
can become
try:
    fp = open("data", "rb")
except FileNotFoundError:
    # Create file
. (Of course, for this sort of thing you can also use the new "x" mode in open().) The errors from incorrect call signatures have improved:
Python 3.3.0+ (3.3:7e83c8ccb1ba, Sep 29 2012, 10:34:54) 
[GCC 4.5.4] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> def f(a, b, c=5, *, kw1, kw2): pass
... 
>>> f(1, kw2=42)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: f() missing 1 required positional argument: 'b'
>>> f(1, 2, kw2=42)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: f() missing 1 required keyword-only argument: 'kw1'
In the future, I think there should be a ArgumentsError subclass of TypeError which provides programmatic access to the signature mismatch, but this is a start. The new standard library modules, ipaddress, lzma, a dn unittest.mock are certainly worth a look.
The Windows installer has an option to set up PATH for you.







Saturday, May 12, 2012

The Architecture of Open Source Applications volume 2 has been published. It includes my chapter on PyPy. You can buy the dead tree version for $35 on Lulu where all the proceeds go to Amnesty International.

Wednesday, March 14, 2012

PyCon 2012 retrospective

I just returned from PyCon 2012, my first PyCon since 2009. It was great to meet everyone in person. I realized that everyone is much nicer in person than they appear on the mailing lists.

I met Dave Malcolm, the man behind the gcc Python plugin support. It looks like his project to automatically detect reference counting bugs in C extensions is going well. In the process of that, he has created Python bindings to gcc, which could be used for many interesting projects.

My talk on the PyPy JIT was on Saturday morning. It went fairly well; I wish I had practiced more. I also gave a lightning talk Saturday evening on what's new in Python 3.3.

On Sunday morning, Guido mentioned six in his keynote, so that was exciting.

I was able to stay for part of Monday sprints. I helped some people with fixing Python 3 tests on the PyPy py3k branch. I also advised some people about CPython work. It occurs to me now we forgot to have a key-signing party.

A lot of thanks needs to be extended to Jesse Noller and the entire PyCon team. That they rganized a convention with over 2,600 attendees on volunteer time is amazing.

Finally, Santa Clara has really nice weather!

Saturday, July 9, 2011

Behind the scenes of py.test's new assertion rewriting

py.test 2.1 was just released. py.test, which uses the Python assert statement to check test conditions, has long had support for displaying intermediate values in subexpressions of a failing assert statement. This feature is called assertion introspection. Historically, py.test performed assertion introspection by reinterpreting failed assertions in order to glean information about subexpressions. In assertion reinterpreting, py.test actually reruns the assertion noting intermediate values during interpretation. This works pretty well but is subject to several problems, most importantly that assert statements with side-effects can produce strange results because they are evaluated twice on failure. py.test 2.1's main new feature, which I wrote (with generous sponsorship from Merlinux GmbH), is a new assertion introspection technique called assertion rewriting. Assertion rewriting modifies the AST of test modules to produce subexpression information when assertions fail. This blog post will give a peek into how this is done and what the rewritten tests look like.

py.test tries to rewrite every module that it collects as a test module. Assertion rewriting uses a PEP 302 import hook to capture test modules for rewriting. I'm happy to report doing this was easier than I expected. Most of the code in the import hook I had to write was dealing with detecting test modules rather than supporting import's extremely complicated API. Rewriting has a non-zero cost during test collection, so py.test compiles rewritten modules to bytecode and caches them in the PEP 3147 PYC repository, __pycache__. One major thing I did have to account for was the possibility that multiple py.test processes would be writing PYC files. (This is a very real possibility when the xdist plugin is being used. Therefore, py.test uses only atomic operations on the rewritten PYC file. Windows, lacking atomic rename, was a pain here.

I'm now going to demonstrate what py.test's rewriting phase does to a test module. Let's dive in with a failing test for a (broken) function that is supposed to create empty files:

import os

def make_empty_file(name):
with open(name, "w") as fp:
fp.write("hello")

def test_make_empty_file():
name = "/tmp/empty_test"
make_empty_file(name)
with open(name, "r") as fp:
assert not fp.read()


This test nicely demonstrates the problem with py.test's old assertion method mentioned in the first paragraph. If we force the old assertion interpretation mode with --assert=reinterp, we see:


def test_make_empty_file():
name = "/tmp/empty_test"
make_empty_file(name)
with open(name, "r") as fp:
> assert not fp.read()
E AssertionError: (assertion failed, but when it was re-run for printing intermediate values, it did not fail. Suggestions: compute assert expression before the assert or use --no-assert)

test_empty_file.py:11: AssertionError


The problem is that assert statement has the side-effect of reading the file. When py.test reinterprets the assert statement, it uses the same file object, now at EOF, and read() returns an empty string. py.test's new rewriting mode fixes this by scanning the assert for introspection information before executing the test. Running py.test with assertion rewriting enabled gives the desired result:


def test_make_empty_file():
name = "/tmp/empty_test"
make_empty_file(name)
with open(name, "r") as fp:
> assert not fp.read()
E assert not 'hello'
E + where 'hello' = ()
E + where = .read

test_empty_file.py:11: AssertionError


So what magic has py.test worked to display such nice debugging information? This is what Python is actually executing:

def test_make_empty_file():
name = '/tmp/empty_test'
make_empty_file(name)
with open(name, 'r') as fp:
@py_assert1 = fp.read
@py_assert3 = @py_assert1()
@py_assert5 = (not @py_assert3)
if (not @py_assert5):
@py_format6 = ('assert not %(py4)s\n{%(py4)s = %(py2)s\n{%(py2)s = %(py0)s.read\n}()\n}' %
{'py0': (@pytest_ar._saferepr(fp) if ('fp' in @py_builtins.locals() is not @py_builtins.globals()) else 'fp'),
'py2': @pytest_ar._saferepr(@py_assert1),
'py4': @pytest_ar._saferepr(@py_assert3)})
raise AssertionError(@pytest_ar._format_explanation(@py_format6))
del @py_assert5, @py_assert1, @py_assert3


As you can see, it's not going to be winning any awards for beautiful Python! (Ideally, though, you'll never have to see or think about it.) Examining the rewritten code, we see a lot of internal variables starting with "@" have been created. The "@", invalid in Python identifiers, is to make sure internal names don't conflict with any user-defined names which might be in the scope. In the first four written lines under the with statement, the test of the assert statement has been expanded into its component subexpressions. This allows py.test to display the values of subexpressions should the assertion fail. If the assertion fails, the if statement in the fifth line of rewriting evaluates to True and a AssertionError will be raised. Under the if statement is the real mess. This is where the helpful error message is generated. The line starting with @py_format6 is simply does string formatting (with %) on a template generated from the structure of the assert statement. This template is filled in with the intermediate values of the expressions collected above. @py_builtins is the builtins module, used in case the test is shadowing builtins the rewriting code uses. The @pytest_ar variable is a special module of assertion formatting helpers. For example, @pytest_ar._saferepr is like builtin repr but gracefully handles long reprs and __repr__ methods that raise exceptions. A non-obvious trick in the format dict is the expression @pytest_ar._saferepr(fp) if ('fp' in @py_builtins.locals() is not @py_builtins.globals()) else 'fp'. This checks whether fp is a local variable or not and customizes the display accordingly. After the initial formatting, the helper function _format_explanation is called. This function produces the indentation and "+" you see in the error message. Finally, we note that if the assertion doesn't fail, py.test cleans up after itself by deleting temporary variables.

The example above is a fairly tame (and luckily also typical) assertion. Rewriting gets more "exciting" when boolean operations and comparisons enter because they require short circuit evaluation, which complicates both the expansion of expressions and formatting (think lots of nested ifs).

In conclusion, py.test's new assertion rewriting fixes some long standing issues with assertion introspection and continues py.test's long tradition of excellent debugging support. (There are now three(!) assertion introspection methods in py.test: two reinterpretation implementations as well as rewriting) I just hope I haven't scared you completely off py.test! :)

Thursday, May 12, 2011

Webpage layout without tables

The HTML5 standard states unequivocally that tables should not be used for layout. Yesterday, I was working on a old website of mine that used tables for that purpose and decided to remove them to conform with the standard. Not being a very experienced web-designer, I thought it would be a simple matter of replacing a lot of table tags with divs and finding the right CSS properties. Evidently, this is much easier said than done.

Originally, my webpage two table columns. The first had the navigation sidebar, and the second had the main content of the page. I could not figure out how to get proper columns with just CSS and divs. Making the sidebar float: left worked except that then the content of the page started wrapping around the bottom of it. Of course, I couldn't find any CSS to get a float to take up the whole height of its container. In the end, I resorted to setting a left margin on the content wider than width of sidebar, which made the two pieces stay out of each other's way.

I'm curious as to what real web-designers say about this problem. There must be some trick I'm missing. It seems many sites, like python.org, are forced to resort to position: absolute tricks to get nice sidebars. I also wonder if the supposed problem with table layouts, that screen readers have trouble with them, could be alleviated by adding a role="layout" attribute to tables. It's a case of "practicality beats purity" if I ever saw one.

Tuesday, April 5, 2011

The next few years of my life

I'm pleased to announce that next year I will be attending Yale college, joining the class of 2015. To say I'm excited, would be like saying PyPy might offer a minor performance advantage over CPython.

What will I study? Your guess is as good as mine. There will certainly be music math, and free software involved, but beyond that I'm leaving until later to figure out.

The only sad part is the large cost. It will doubtless be worth it, though.

Boola, Boola!

Tuesday, March 15, 2011

six 1.0.0 final finally released

I finally found time to release six 1.0.0. six is a Python 2 and 3 compatibility library. You can find the documentation and download it on PyPI.

There haven't been many changes since the beta: one bugfix and one new advanced feature. The bugfix is that unicode escapes are now properly decoded with the u() fake literal in both Python 2 and 3. The feature is that there is now an api for adding items to the "six.moves" interface. This was requested by ActiveState, which uses six in the ActivePython package manager.

Enjoy!