Bug: Replace Selection changes double backslashes to single backslashes

[Khris777]

Description of your problem

When doing a Replace in a selection that has a string with double backslashes, the Replace command will change the double backslashes to single backslashes even though something completely different is being replaced.

What steps will reproduce the problem?

1. Example code:

a = 1
b = 2
c = a+b
s = "C:\\Test\\Test2\\output.txt"

2. Mark the example code and replace selection a with x
3. The output will be this:

x = 1
b = 2
c = x+b
s = "E:\Test\Test2\output.txt"

What is the expected output?

x = 1
b = 2
c = x+b
s = "E:\\Test\\Test2\\output.txt"

Please provide any additional information below

It only happens on Replace selection, not when doing Replace all.

Versions and main components

• Spyder Version: 3.2.6
• Python Version: 3.6.4
• Qt Version: 5.6.2
• PyQt Version: 5.6
• Operating system: Windows 10

Dependencies

IPython >=4.0 : 6.2.1 (OK)
cython >=0.21 : 0.27.3 (OK)
jedi >=0.9.0 : 0.11.1 (OK)
nbconvert >=4.0 : 5.3.1 (OK)
numpy >=1.7 : 1.13.3 (OK)
pandas >=0.13.1 : 0.22.0 (OK)
pycodestyle >=2.3: 2.3.1 (OK)
pyflakes >=0.6.0 : 1.6.0 (OK)
pygments >=2.0 : 2.2.0 (OK)
pylint >=0.25 : 1.8.1 (OK)
qtconsole >=4.2.0: 4.3.1 (OK)
rope >=0.9.4 : 0.10.7 (OK)
sphinx >=0.6.6 : 1.6.6 (OK)
sympy >=0.7.3 : 1.1.1 (OK)

[ccordoba12]

Thanks for reporting. We'll take a look at this in a future release.

[CAM-Gerlach]

@ccordoba12 Likely related to #5494 and the associated #5555 ; that apparently fixes other escaped cases but not escaping the escape char itself.

[oscargus]

This becomes even more of a problem considering raw strings. While adding \ to the current reg-exp would solve this particular issue, it will still break on lines like:

b = r'\a\v\b'

Now, one may think that since all the escaped characters in the reg-exp are non-letters one can simply add all letters (and numbers). However, in e.g. SymPy, there is LaTeX printing, so a line in the editor may very well look like

r'\{7 \& 8\}'

If it wasn't for the word option, one could simply use subs instead for queries that are not regular expressions. Right now, I do not really have any idea how to deal with it...