Drop duplicated tests by default #1609

[ioggstream]

When using pytest --skip-duplicates:

• ignores duplicate tests on the basis of the purebasename;
• take into account only the first one;

[coveralls]

Coverage decreased (-0.01%) to 93.026% when pulling 83fabcc on ioggstream:1609-features into ae07985 on pytest-dev:features.

[nicoddemus]

Hey @ioggstream,

I'm wondering, in which situations would be useful to drop different tests just because they have the same name?

At first I thought this was implementing #1609.

[nicoddemus]

(At the moment I'm a strong 👎 on this, unless I misunderstood the implementation... it makes it very easy for someone to rename a test in some separate directory to the same name used elsewhere by accident and it will simply skip the second test).

[coveralls]

Coverage decreased (-0.01%) to 93.026% when pulling a3c716b on ioggstream:1609-features into ae07985 on pytest-dev:features.

[ioggstream]

@nicoddemus fixed to avoid duplicate paths, not names.

[nicoddemus]

Oh OK, that makes more sense, thanks. 😅

I think dropping duplicated test paths on the command line should be the default behavior as discussed in #1609 though. Would you mind implementing it as such and adding a --keep-duplicates instead?

Also, I think we need a small doc update with the new option.

[ioggstream]

I implemented it this way to not change the default behavior.

Should I turn it to keep-duplicate?

Peace.
R
Il 23/lug/2016 17:46, "Bruno Oliveira" notifications@github.com ha
scritto:

Oh OK, that makes more sense, thanks. 😅

I think dropping duplicated test paths on the command line should be the
default behavior as discussed in #1609
#1609 though. Would you mind
implementing it as such and adding a --keep-duplicates instead?

Also, I think we need a small doc update with the new option.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#1756 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/ABFobIZuGnHbz5ChBabdOUwhGVYevb95ks5qYjdugaJpZM4JTX5M
.

[nicoddemus]

Should I turn it to keep-duplicate?

Yes please. We were planning to do this backward incompatible change eventually, and 3.0 seems like an excellent opportunity to do so.

[tomviner]

fpath is unused, which is currently failing the lint test.

[coveralls]

Coverage decreased (-0.01%) to 93.047% when pulling ab6aef1 on ioggstream:1609-features into a24146d on pytest-dev:features.

[ioggstream]

@nicoddemus @tomviner I had to change this test because the patch changed the default behavior.

[nicoddemus]

Thanks @ioggstream! 👍