Bug#637653: pu: gdebi/0.6.4+squeeze1

["Adam D. Barratt" <adam@adam-barratt.org.uk>]

On Sat, 2011-08-13 at 15:15 +0200, Luca Falavigna wrote:
> I'd like to get your permission to upload to stable a fix for a bug that
> happens only on certain locales (#629403). This solution should be more
> robust than relying on correctness of the translations.

Thanks for working on this.

> I attach a debdiff of the proposed changes. It's quite heavy because of
> the offsets introduced in .po files (regenerated at build time) due to
> line changes in the patch, real changes are limited to the files below:

The regular expression being passed to findall() looks slightly
confused, although I realise that in the context of the data being
matched the differences never actually apply currently.

(Specifically, brackets and parentheses don't need escaping inside a
character class or special treatment other than brackets appearing at
the limits of the class, nor is | used as a separator.)  By way of a
slightly contrived example:

>>> msg = "a|b|c[y/n]hello"
>>> print findall("[\[|\(](\S+)/\S+[\]|\)]", msg)
['b|c[y']

As I understand the intent, the following appears to match more closely:

>>> msg = "a|b|c[y/n]hello"
>>> print findall("[[(](\S+)/\S+[])]", msg)
['y']

Are there actually any translations using parentheses rather than
brackets as delimiters?

Regards,

Adam