Allow changing difflib._file_template character encoding.

[storchaka_gmail.com]

http://bugs.python.org/review/2052/diff/11867/Doc/library/difflib.rst
File Doc/library/difflib.rst (right):

http://bugs.python.org/review/2052/diff/11867/Doc/library/difflib.rst#newcode108
Doc/library/difflib.rst:108: numlines=5, charset='utf-8')
make_file has too many parameters. To specify positional 7th parameter you
should specify all 6 previous parameters. I believe that most (or even all)
users of make_file() will specify charset as keyword parameter, I suggest to
force this and make charset a keyword-only parameter. This will help when we
will rearrange or remove parameters after 6th parameter.

http://bugs.python.org/review/2052/diff/11867/Lib/difflib.py
File Lib/difflib.py (right):

http://bugs.python.org/review/2052/diff/11867/Lib/difflib.py#newcode1712
Lib/difflib.py:1712: table=self.make_table(fromlines,tolines,fromdesc,todesc,
May be make_file() result  should be encodable to the charset? Then table should
be transformed:

    table = table.encode(charset, 'xmlcharrefreplace').decode(charset)

[berkerpeksag]

http://bugs.python.org/review/2052/diff/11867/Doc/library/difflib.rst
File Doc/library/difflib.rst (right):

http://bugs.python.org/review/2052/diff/11867/Doc/library/difflib.rst#newcode108
Doc/library/difflib.rst:108: numlines=5, charset='utf-8')
On 2015/03/06 22:26:01, storchaka wrote:
> make_file has too many parameters. To specify positional 7th parameter you
> should specify all 6 previous parameters. I believe that most (or even all)
> users of make_file() will specify charset as keyword parameter, I suggest to
> force this and make charset a keyword-only parameter. This will help when we
> will rearrange or remove parameters after 6th parameter.

Done.

http://bugs.python.org/review/2052/diff/11867/Lib/difflib.py
File Lib/difflib.py (right):

http://bugs.python.org/review/2052/diff/11867/Lib/difflib.py#newcode1712
Lib/difflib.py:1712: table=self.make_table(fromlines,tolines,fromdesc,todesc,
On 2015/03/06 22:26:01, storchaka wrote:
> May be make_file() result  should be encodable to the charset? Then table
should
> be transformed:
> 
>     table = table.encode(charset, 'xmlcharrefreplace').decode(charset)

Done.

[storchaka_gmail.com]

http://bugs.python.org/review/2052/diff/14183/Lib/difflib.py
File Lib/difflib.py (right):

http://bugs.python.org/review/2052/diff/14183/Lib/difflib.py#newcode1688
Lib/difflib.py:1688: def make_file(self, fromlines, tolines, fromdesc='',
todesc='', context=False,
Too long line.

http://bugs.python.org/review/2052/diff/14183/Lib/difflib.py#newcode1689
Lib/difflib.py:1689: numlines=5, *, charset='utf-8'):
May be should be documented, that default charset was changed.

http://bugs.python.org/review/2052/diff/14183/Lib/difflib.py#newcode1712
Lib/difflib.py:1712: table=table.encode(charset,
'xmlcharrefreplace').decode(charset),
There is a possibility that self._styles or self._legend would be non-ascii. It
would be better to do encoding/decoding of all formatted result, not only a
table.

http://bugs.python.org/review/2052/diff/14183/Lib/test/test_difflib.py
File Lib/test/test_difflib.py (right):

http://bugs.python.org/review/2052/diff/14183/Lib/test/test_difflib.py#newcod...
Lib/test/test_difflib.py:232: def test_make_file_iso88591_charset(self):
Could you please add characters non encodable with iso-8859-1 to input data? Or
may be easier to make a test with ascii, if they already contain non-ascii data.

[berkerpeksag]

http://bugs.python.org/review/2052/diff/14183/Lib/difflib.py
File Lib/difflib.py (right):

http://bugs.python.org/review/2052/diff/14183/Lib/difflib.py#newcode1688
Lib/difflib.py:1688: def make_file(self, fromlines, tolines, fromdesc='',
todesc='', context=False,
On 2015/03/13 21:37:42, storchaka wrote:
> Too long line.

Done.

http://bugs.python.org/review/2052/diff/14183/Lib/difflib.py#newcode1689
Lib/difflib.py:1689: numlines=5, *, charset='utf-8'):
On 2015/03/13 21:37:42, storchaka wrote:
> May be should be documented, that default charset was changed.

Done.

http://bugs.python.org/review/2052/diff/14183/Lib/difflib.py#newcode1712
Lib/difflib.py:1712: table=table.encode(charset,
'xmlcharrefreplace').decode(charset),
On 2015/03/13 21:37:42, storchaka wrote:
> There is a possibility that self._styles or self._legend would be non-ascii.
It
> would be better to do encoding/decoding of all formatted result, not only a
> table.

You mean by subclassing HtmlDiff right? These attributes are private and
undocumented, but I guess it's doesn't hurt to do the encoding/decoding dance.

[berkerpeksag]

http://bugs.python.org/review/2052/diff/14183/Lib/test/test_difflib.py
File Lib/test/test_difflib.py (right):

http://bugs.python.org/review/2052/diff/14183/Lib/test/test_difflib.py#newcod...
Lib/test/test_difflib.py:232: def test_make_file_iso88591_charset(self):
On 2015/03/13 21:37:42, storchaka wrote:
> Could you please add characters non encodable with iso-8859-1 to input data?
Or
> may be easier to make a test with ascii, if they already contain non-ascii
data.

Done.

[storchaka_gmail.com]

https://bugs.python.org/review/2052/diff/14191/Doc/library/difflib.rst
File Doc/library/difflib.rst (right):

https://bugs.python.org/review/2052/diff/14191/Doc/library/difflib.rst#newcod...
Doc/library/difflib.rst:128: *charset* keyword-only argument was added. The
default charset of HTML document
PEP8: "Limit all lines to a maximum of 79 characters."

PEP8: "You should use two spaces after a sentence-ending period."

https://bugs.python.org/review/2052/diff/14191/Doc/whatsnew/3.5.rst
File Doc/whatsnew/3.5.rst (right):

https://bugs.python.org/review/2052/diff/14191/Doc/whatsnew/3.5.rst#newcode232
Doc/whatsnew/3.5.rst:232: can now be customized by using *charset* keyword-only
parameter. The default
The same about a period.

https://bugs.python.org/review/2052/diff/14191/Lib/difflib.py
File Lib/difflib.py (right):

https://bugs.python.org/review/2052/diff/14191/Lib/difflib.py#newcode1709
Lib/difflib.py:1709: styles = self._styles.encode(charset,
Oh, I meant to do the encoding/decoding dance on formatted result, not every
component.

return (self._file_template % dict(
            styles=styles,
            legend=legend,
            table=table,
            charset=charset
        )).encode(charset, 'xmlcharrefreplace').decode(charset)

[berkerpeksag]

https://bugs.python.org/review/2052/diff/14191/Doc/library/difflib.rst
File Doc/library/difflib.rst (right):

https://bugs.python.org/review/2052/diff/14191/Doc/library/difflib.rst#newcod...
Doc/library/difflib.rst:128: *charset* keyword-only argument was added. The
default charset of HTML document
On 2015/03/14 18:29:13, storchaka wrote:
> PEP8: "Limit all lines to a maximum of 79 characters."
> 
> PEP8: "You should use two spaces after a sentence-ending period."

Done.

https://bugs.python.org/review/2052/diff/14191/Doc/whatsnew/3.5.rst
File Doc/whatsnew/3.5.rst (right):

https://bugs.python.org/review/2052/diff/14191/Doc/whatsnew/3.5.rst#newcode232
Doc/whatsnew/3.5.rst:232: can now be customized by using *charset* keyword-only
parameter. The default
On 2015/03/14 18:29:13, storchaka wrote:
> The same about a period.

Done.

https://bugs.python.org/review/2052/diff/14191/Lib/difflib.py
File Lib/difflib.py (right):

https://bugs.python.org/review/2052/diff/14191/Lib/difflib.py#newcode1709
Lib/difflib.py:1709: styles = self._styles.encode(charset,
On 2015/03/14 18:29:13, storchaka wrote:
> Oh, I meant to do the encoding/decoding dance on formatted result, not every
> component.
> 
> return (self._file_template % dict(
>             styles=styles,
>             legend=legend,
>             table=table,
>             charset=charset
>         )).encode(charset, 'xmlcharrefreplace').decode(charset)

Done.