[PATCH][Bug 86451]Normalise path separators in ignore patterns

[Kent Gibson]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

This bundle is as per the previous, but includes some of the test
cases for bug 86451 that I had neglected to carry over from my initial
patch.

Kent Gibson wrote:
>
>
> John Arbash Meinel wrote:
>
>> Thanks for updating your patch. I'd really like to get this in
>> 0.15. There are a few small things, though.
>
>> ...
>
>> v- Your spacing here is a little odd, I think it is better as:
>
>> if (len(m) % 2) != 0:
>
> Yeah - that was odd.  I hadn't even noticed.  Must be fat fingers.
>> Also, while not strictly required, it would be good to have a
>> simple docstring. Maybe "check for trailing backslashes in the
>> pattern"
>
> I thought it was clear enough from the function name and warning.
> But if it will help....
>
>> + +_normalize_re = re.compile(ur'\\(?!x\d\d|\d\d\d|u\d\d\d\d)') +
>> + +def normalize_pattern(pattern): +    """Converts backslashes
>> in path patterns to forward slashes. + +    Avoids converting
>> escape backslashes. +    """ +    if not
>> pattern.startswith('RE:'): + pattern = _normalize_re.sub('/',
>> pattern) +    return pattern.rstrip('/')
>
>> ^- If I understand correctly, you are allowing backslashes for
>> character codes (hex) \x00 or (octal) \000 (unicode decimal)
>> \u0000.  There are a few problems, though.
>
>> 1) I don't think you need to do this. I'm guessing you
>> misunderstood the tests. But the data in .bzrignore is meant to
>> be utf-8 data. So non-ascii characters are expected to be
>> present, but not escapes.
> I understood that at least one of the tests failed without it.
> Generally when I make a change and tests fail I assume it is
> because I broke the code. But based on your comments, I think this
> is because some of the tests use raw strings when they should not.
> I've fixed those and replaced the regex with a simple replace.
>
> That's a relief.  That there could be escapes in ignore patterns
> (other than REs) had me worried.
>
> Updated bundle attached.
>
> Can someone can tell me why the bundle claims I changed a large
> chunk of test_regex, when I only dropped the r in the RE lines? bzr
> gvimdiff shows I only changed those lines, not the whole block.
> What's da diff?
>
> Cheers, Kent.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFF59sFgoxTFTi1P8QRAozoAJwIb5TMEf8nJDL8D5Od/bRBuMO5SQCfbwEm
1fi7BDRvk/+OuT3hBdIK1rU=
=VBV9
-----END PGP SIGNATURE-----

-------------- next part --------------
A non-text attachment was scrubbed...
Name: lp86451_c.patch
Type: text/x-patch
Size: 21658 bytes
Desc: not available
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20070302/6c95ebe4/attachment-0001.bin