Noissue - Fix edge cases with raw strings in flake8-abp

Noissue - Fix edge cases with raw strings in flake8-abp

[Sebastian Noack, 2016-05-30 09:57:43]

https://codereview.adblockplus.org/29345337/diff/29345338/flake8-abp/flake8_a...
File flake8-abp/flake8_abp.py (left):

https://codereview.adblockplus.org/29345337/diff/29345338/flake8-abp/flake8_a...
flake8-abp/flake8_abp.py:402: literal = re.sub(r'\\(?!{})'.format(literal[0]),
The problem with old logic was that there are cases in which a string if using
raw syntax requires double quotes while repr() would use single quotes.

Example: r"'\"" -> '\'\\"'

Since backslashes in raw strings are both interpreted as literal backslashes and
escape quotes at the same time. We cannot write the above raw string like r'\'"'
since this is interpreted as '\\\'"'.

https://codereview.adblockplus.org/29345337/diff/29345338/flake8-abp/flake8_a...
File flake8-abp/flake8_abp.py (right):

https://codereview.adblockplus.org/29345337/diff/29345338/flake8-abp/flake8_a...
flake8-abp/flake8_abp.py:387: match = re.search(r'^([rub]*)([\'"]{1,3})(.*)\2$',
On Python 3, you can write br'' as well as rb''. The old regexp, however, relied
on 'b' always occurring before 'r'.

[Vasily Kuznetsov, 2016-05-30 10:08:19]

https://codereview.adblockplus.org/29345337/diff/29345338/flake8-abp/flake8_a...
File flake8-abp/flake8_abp.py (left):

https://codereview.adblockplus.org/29345337/diff/29345338/flake8-abp/flake8_a...
flake8-abp/flake8_abp.py:402: literal = re.sub(r'\\(?!{})'.format(literal[0]),
On 2016/05/30 09:57:42, Sebastian Noack wrote:
> The problem with old logic was that there are cases in which a string if using
> raw syntax requires double quotes while repr() would use single quotes.
> 
> Example: r"'\"" -> '\'\\"'
> 
> Since backslashes in raw strings are both interpreted as literal backslashes
and
> escape quotes at the same time. We cannot write the above raw string like
r'\'"'
> since this is interpreted as '\\\'"'.

I think it would be good to add a test for this case.

[Sebastian Noack, 2016-05-30 10:18:04]

https://codereview.adblockplus.org/29345337/diff/29345338/flake8-abp/flake8_a...
File flake8-abp/flake8_abp.py (left):

https://codereview.adblockplus.org/29345337/diff/29345338/flake8-abp/flake8_a...
flake8-abp/flake8_abp.py:402: literal = re.sub(r'\\(?!{})'.format(literal[0]),
On 2016/05/30 10:08:19, Vasily Kuznetsov wrote:
> On 2016/05/30 09:57:42, Sebastian Noack wrote:
> > The problem with old logic was that there are cases in which a string if
using
> > raw syntax requires double quotes while repr() would use single quotes.
> > 
> > Example: r"'\"" -> '\'\\"'
> > 
> > Since backslashes in raw strings are both interpreted as literal backslashes
> and
> > escape quotes at the same time. We cannot write the above raw string like
> r'\'"'
> > since this is interpreted as '\\\'"'.
> 
> I think it would be good to add a test for this case.

Done.

[Vasily Kuznetsov, 2016-05-30 11:43:45]

LGTM

[Sebastian Noack, 2016-05-30 12:33:42]

https://codereview.adblockplus.org/29345337/diff/29345340/flake8-abp/flake8_a...
File flake8-abp/flake8_abp.py (left):

https://codereview.adblockplus.org/29345337/diff/29345340/flake8-abp/flake8_a...
flake8-abp/flake8_abp.py:400: elif start[0] == end[0]:
I just realized, after pushing this change, that we could have avoided a level
of indentation/complexity here:

if first_token and re.search(...):
    ...
elif start[0] != end[0]:
    pass
elif 'r' in prefixes:
    ...
else:
    ...

I leave it up to Jon whether to change it while touching this code again while
working on #4044.