allow inplace addition to update currval

[Midnighter]

This is an easy implementation for the feature discussed in #6.

[wolph]

Thank you for the addition @Midnighter, I can't merge the pull request yet because it breaks the build ( https://travis-ci.org/WoLpH/python-progressbar/builds/24969677 ). This repository has a 100% test coverage requirement so all new code needs to have a test as well.

You can add tests if you want, otherwise I'll do it as soon as I'll have a bit more time (probably in about a week).

[Midnighter]

Can't say that I see the value of having 100% coverage when your tests are really just examples but I guess it's a good idea to add examples for the new behavior anyway.

[wolph]

They serve as both examples and as a simple test that the basics work. If something would be broken, the code wouldn't be callable anymore and the tests would break.

The tests aren't the highest quality indeed, but at least we know that every line of code is being run and therefore working for at least 1 use-case.

[Midnighter]

So would you like me to add more examples at the end and keep the numbering or intersperse existing examples with alternative versions that use the new __iadd__?

[wolph]

Just an additional example that shows the usage of the new operator should do the trick :)

Basically the code that you showed in #6 should work I think.

[Midnighter]

Strangely, Travis now complains about lines 304-307 which I didn't even touch. They are within the update method, though.

if self.redirect_stdout and sys.stdout.tell():
    self.fd.write('\r' + ' ' * self.term_width + '\r')
    self._stdout.write(sys.stdout.getvalue())
    self._stdout.flush()
    sys.stdout = StringIO()

[wolph]

It appears you might have removed one of the tests?
https://github.com/Midnighter/python-progressbar/commit/6679eca388e049495bf5a3b96b39d5e3a714a2ea#diff-c64b3378e7819bcbb93349363c67699cL265

[Midnighter]

Yeah, pyflakes was showing it as an exact duplicate. Both by function name and the content looked the same, too. Maybe I was just tired, though. I'll try with putting it back.

[Midnighter]

Guess I was tired, didn't notice that one case was treating stdout and the other stderr. I have renamed them to versions a and b in order to avoid a conflict.

[wolph]

Awesome, thank you very much for the pull request :)

I'll push out a new version as well