599177 - Python FutureWarning about behavior of has_key method on Python 2.7

Status: RESOLVED FIXED

(Add-on SDK Graveyard :: General, defect)

Description

[Myk Melez [:myk] [@mykmelez]]

over in bug 596528, comment 7, Matej reported that every time he runs cfx he sees the following warnings:

/home/matej/projekty/jetpack-sdk/python-lib/mozrunner/__init__.py:221:
FutureWarning: The behavior of this method will change in future versions.  Use
specific 'len(elem)' or 'elem is not None' test instead.
  if desc and desc.attrib.has_key('{http://www.mozilla.org/2004/em-rdf#}id'):
/home/matej/projekty/jetpack-sdk/python-lib/mozrunner/__init__.py:223:
FutureWarning: The behavior of this method will change in future versions.  Use
specific 'len(elem)' or 'elem is not None' test instead.
  elif desc and desc.find('.//{http://www.mozilla.org/2004/em-rdf#}id') is not
None:

Matej is using Python 2.7 on Fedora14Beta.  I don't see this on my Linux installation running Python 2.6.4, but presumably it's something new in Python 2.7.  To avoid showing users these warnings (and continue working on future versions of Python), the SDK should implement the recommended changes and push the changes upstream.

Brian: you are our resident Python expert; might this be something you could tackle?

Comment 1

[Brian Warner [:warner :bwarner]]

sure, I'll look at it.

Comment 2

[Brian Warner [:warner :bwarner]]

Attachment: mozrunner patch to hush FutureWarning on py2.7

Some digging revealed that the actual issue is that the new version of
ElementTree (included in python2.7 for the first time) has deprecated the use
of boolean(element), in preparation for changing the way it works in some
future version. So in the line:

  if desc and desc.attrib.has_key('{http://www.mozilla.org/2004/em-rdf#}id'):

the actual problem was the "if desc" part (which is equivalent to "if
bool(desc)"), not the call to has_key().

http://effbot.org/zone/elementtree-13-intro.htm#truth-testing has the release
notes about the change.

My best guess is that this code is testing whether the "Description" XML
element contains non-"targetApplication" children, so the implicit bool(desc)
test is checking whether there are any child elements. If that's correct,
then replacing "desc" with "len(desc)" is the right fix.

This patch also changes a "len(about) is 0" into "len(about) == 0", since
numbers can't be compared by identity. This happens to work because Python
optimizes some math by interning a bunch of small integers, but e.g. "1000 is
1001-1" evaluates as False. Comparing with == is more robust.

Comment 3

[Brian Warner [:warner :bwarner]]

copying Clint to see how he wants to handle this upstream

Comment 4

[Atul Varma [:atul]]

Comment on attachment 478121 [details] [diff] [review]
mozrunner patch to hush FutureWarning on py2.7

Looks good to me!

Comment 5

[Brian Warner [:warner :bwarner]]

landed, in http://hg.mozilla.org/labs/jetpack-sdk/rev/55bcbfcd04c6

Comment 6

[cmtalbert]

Comment on attachment 478121 [details] [diff] [review]
mozrunner patch to hush FutureWarning on py2.7

Let's take this for 1.5.1

Comment 7

[cmtalbert]

Heather, can you shepherd this change into 1.5.1 hotfix, please?

Comment 8

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

(In reply to comment #7)
> Heather, can you shepherd this change into 1.5.1 hotfix, please?

Was this pushed? (Just wondering...)

Comment 9

[Henrik Skupin [:whimboo][⌚️UTC+1]]

(In reply to comment #7)
> Heather, can you shepherd this change into 1.5.1 hotfix, please?

Looks like that has been forgotten. I had to update the patch for the hotfix-1.5.1 and master branch. Both have been landed now:

http://github.com/mozautomation/mozmill/commit/a8ca453b8a90cb3bf545fc9d25bae44e283f744c (master)
http://github.com/mozautomation/mozmill/commit/1aa3a377642ae02913a8a02c1e989c268ecfab73 (hotfix-1.5.1)

Comment 10

[Myk Melez [:myk] [@mykmelez]]

The Add-on SDK is no longer a Mozilla Labs experiment and has become a big enough project to warrant its own Bugzilla product, so the "Add-on SDK" product has been created for it, and I am moving its bugs to that product.

To filter bugmail related to this change, filter on the word "looptid".