This commit fixes a regression that sneaked into Python 3.3 where importlib
was not respecting -E when checking for the PYTHONCASEOK environment variable.
"The fix" is to tell Mercurial that the test files are binary.
Windows developers: to get the correct line endings in your checkout,
delete Lib\test\xmltestdata, and then "hg revert" that directory.
Why the Windows buildbots didn't fail test_sax remains a mystery :-(
Patch import and initialization in SearchEngine to make testing easier.
Improve docstrings, especially to clarify the double role of 'ok' parameters.
Original patch by Phil Webster.
runner were run in an environment where the process inherited an ignore
setting for SIGINT. Restore the SIGINT handler to the desired
KeyboardInterrupt raising one during that test.
Using ``method`` markup because the method is on a callback object, not an
explicitly documented method. :meth: markup creates links within the current
class which is incorrect.
In addition, indent the paragraph correctly.
Previously it generated them if they were given a value, but completely
ignored them if they were present in the string passed in to be parsed. Now
if the flag appears on a cookie, the corresponding Morsel key will reference a
True value. Other pre-existing behavior is retained in this maintenance
patch: if the source contains something like 'secure=foo', morsel['secure']
will return 'foo'. Since such a value doesn't round trip and never did (and
would be a surprising occurrence) a subsequent non-bug-fix patch may change
this behavior.
Inspired by a patch from Julien Phalip, who reviewed this one.
With importlib getting used more and more, changing __import__ will
not work as well as people used to hope as it will potentially bypass
importers, etc. It also will not work with importlib.import_module()
as it uses "importlib.__import__" (i.e. importlib's implementation of
import) directly and not builtins.__import__.
This also backs out the previous fixes for for #14360, #1717, and #16564.
Those bugs were actually caused by the fact that set_payload didn't decode to
str, thus rendering the model inconsistent. This fix does mean the data
processed by the encoder functions goes through an extra encode/decode cycle,
but it means the model is always consistent. Future API updates will provide
a better way to encode payloads, which will bypass this minor de-optimization.
Tests by Vajrasky Kok.
Christian Heimes
Victor Stinner
Antoine Pitrou
Meador Inge
Tim Peters
Eli Bendersky
Terry Jan Reedy
Charles-François Natali
Gregory P. Smith
Serhiy Storchaka
Ezio Melotti
R David Murray
Brett Cannon