In particular, document the new comprehensions and remove mentions of long integers.
Fix a bunch of related things in the lib ref.

1. coding_spec() only looks at first two lines of bytes to avoid
   a UnicodeDecodeError if rest of file is e.g. latin-1
2. coding_spec() handles \n or \r
3. Clarify that locale_encoding is used by calling it that.  However, it's
   still called IOBinding.encoding in other parts of IDLE and that usage
   needs to be checked to verify that's still what is desired.
4. Return None from _decode() if decoding fails.
5. Name the vars representing bytes and strings or chars appropriately.

and deterministic.

function.

2. If the load fails, close the half-built edit window.
3. Don't reopen the file to check the shebang.
4. Clarify that we're setting tabs in Tk.

M    EditorWindow.py
M    FileList.py

collections with some brief notes.

The formatting of errors from pdb's own print command is different
from the formatting of errors when pdb exec's an arbitrary command;
the introduction of print as a pdb command caused this test to use
the former instead of the latter, causing the test to fail.

Corrected code for invalid conversion specifier.
Added tests to verify.

Modified string.Formatter to correctly expand format_spec's,
and added a limit to recursion depth.  Added _vformat()
method to support both of these.

This should fix the Lib\test\test_uuid.py test on Windows.

interactive mode. Fixes #1100.

Fix the string formatting docs for now; if the feature is changed again, the docs can be changed too.

the acronym.

rather than just complaining about stderr not being what we expected masking
the true problem.