You're right, that doesn't work correctly. Although the fix is actually still a bit "wrong" as it keeps the counter at 1 once it's been incremented. Shouldn't really be a problem, but it's not as neat :P

I guess when resetting the stream it should just set the counter to 0 in all cases.

@wolph That was my fault... fixed. [It didn't always reset, since class is instantiated at import time.]

Not sure what's triggering coverage that's different than master. Perhaps builds are broken there too?

When are you likely to be able to ship a new release with this patch?

My weekend is a bit busy but I might have time on sunday, monday at the latest :)

As for the tests, I'm guessing the tests are simply broken. They worked for the old (wrong) code so with the new code they need to be changed as well: https://github.com/WoLpH/python-progressbar/blob/6ca8abf16f8b7f2ca9d958ea19a7fe8887d9d403/tests/test_stream.py
I can take a look when I merge if you'd prefer, shouldn't be too hard to fix them

@wolph It's actually not the tests that are failing... but coverage. But since the patch only changes code within the methods, I'm not sure what's going on there.

Monday sounds great. Thanks for being so responsive!

I understand, but due to the changed behaviour the else of that if statement is never reached which is why coverage is saying it isn't covered.

I just realised as much (hitherto not used coverage). Will fix that too.

Testing locally is much easier than using travis :)

pip install -r tests/requirements.txt
py.test

After that you should have a htmlcov directory containing the coverage. Simply open the index.html in your browser and you'll see which part is covered and which isn't.

Yep... figured it out :). Just adding a test now. Thanks!

I added a fairly esoteric test to test_stream.py, which tests that progress bars work on non-standard file-type outputs (i.e. StringIO vs sys.stdout/sys.stderr), which rounds coverage back out to 100%. Feel free to modify the test if needed.

@matthewwardrop the new release is online and it seems to do the trick :)

Thanks for the help!

Thanks @wolph !