[MERGE] Repository.iter_files_bytes()

[Vincent Ladeuil]

>>>>> "Andrew" == Andrew Bennetts <andrew.bennetts at canonical.com> writes:

    Andrew> Vincent Ladeuil wrote:
    Andrew> [...]
    >> Given that using '_' as an anonymous variable is a common idiom
    >> in python, this conflicts hard with gettext reserving '_' for its
    >> private usage by adding it to __global__ so that it can be
    >> accessed anywhere without importing gettext again.

    Andrew> I don't understand why that's such a big issue.

Because using '_' as an anonymous variable overrides the '_'
defined by gettext.

I don't remember the specifics right now, but if you restore bzr
and bzr-gtk before the mentioned patch you should be able to
reproduce the problem.

The thread that should contain the most relevant info is at
https://lists.canonical.com/archives/bzr-gtk/2008-January/000417.html.

    Andrew> How many functions both use '_' as an anonymous
    Andrew> variable *and* want to use gettext? 

The problem wasn't in functions that were using both, but indeed
the side-effect of one function using '_' as an anonymous and the
other using '_' as per gettext definition.

Which made it hard to debug since you had to execute the former
before the later or the bug didn't manifest itself.

    Andrew> The small minority that do can use your workaround of
    Andrew> importing gettext.gettext directly, or is the problem
    Andrew> that xgettext doesn't provide a way to cope with two
    Andrew> possible keywords?  Or does xgettext not cope with
    Andrew> “_” sometimes not being used as a gettext function?

I can't remember, sorry.

<snip/>

    Andrew> Huh?  It sounds like you're saying that code like:

    Andrew>     # foo.py
    Andrew>     x = 1
    Andrew>     def func():
    Andrew>         x = 2
    Andrew>     func()
    Andrew>     print x == 1

    Andrew> would print False, which is clearly not the case in
    Andrew> Python.  Assignment in a function doesn't modify
    Andrew> globals unless there is a “global” declaration in
    Andrew> that function, or unless I'm missing some very nasty
    Andrew> magic.

It seems it is nasty black magic indeed :)

Good hunt :)

     Vincent