8000 MAINT: add 'nightly' (= 3.5ish) to travis build matrix by njsmith · Pull Request #6015 · numpy/numpy · GitHub
[go: up one dir, main page]

Skip to content

MAINT: add 'nightly' (= 3.5ish) to travis build matrix #6015

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jun 25, 2015

Conversation

njsmith
Copy link
Member
@njsmith njsmith commented Jun 25, 2015

Not sure if this works, but it's documented as being recently added so let's see what Travis makes of this PR.

@juliantaylor
Copy link
Contributor

fwiw, we already have a 3.5 test build in gh-5772, its going to fail on a zipfile change

@njsmith
Copy link
Member Author
njsmith commented Jun 25, 2015

Oh, interesting. My motivation is #6000, which is impossible to test at all without a real 3.5 build...

It does look like this PR is working, in that Travis does claim to be running a test against "Python 3.5.0a4+".

@juliantaylor
Copy link
Contributor

Currently debian is ahead of travis so called nightly. In the other PR more stuff is failing than here.

I'm not so sure about adding python nightlys to our regular tests, useful during cpython's release phase, before that it will probably just add noise. But it could be worth a try at least.

@njsmith
Copy link
Member Author
njsmith commented Jun 25, 2015

Yeah, I wouldn't want it in general, but between now and when 3.5-final gets released it seems pretty useful.

@njsmith
Copy link
Member Author
njsmith commented Jun 25, 2015

Have we filed a bug about the GzipFile thing upstream?

@juliantaylor
Copy link
Contributor

To my knowledge no one has yet.

@juliantaylor
Copy link
Contributor

mh the new failure (due to deprecation) is not in numpy but nose itself ... that will not be fun to get sorted.

@njsmith
Copy link
Member Author
njsmith commented Jun 25, 2015

wow, yeah, travis "nightly" says it's hg revision 1158f5d7a48e, which is from May 7. That's before they even branched 3.5 off from trunk.

@njsmith
Copy link
Member Author
njsmith commented Jun 25, 2015

We could also some version of 3.5 to travis as a "failure allowed" matrix entry, so at least we'll be able to easily peek in and check by hand to make sure we haven't broken new stuff.

@njsmith
Copy link
Member Author
njsmith commented Jun 25, 2015

Having taken a look at the GzipFile issue, I'm not going to bother filing a bug upstream: #6016

@njsmith
Copy link
Member Author
njsmith commented Jun 25, 2015

Also the nose issue is no big deal... it's just a deprecation warning, so worst case we just filter it out in the tests.

@charris
Copy link
Member
charris commented Jun 25, 2015

The imp module also raises a PendingDeprecation on import. I had a mostly complete PR for that...

@njsmith
Copy link
Member Author
njsmith commented Jun 25, 2015

I filed the nose issue upstream as nose-devs/nose#929.

charris added a commit that referenced this pull request Jun 25, 2015
MAINT: add 'nightly' (= 3.5ish) to travis build matrix
@charris charris merged commit c8759f1 into numpy:master Jun 25, 2015
@charris
Copy link
Member
charris commented Jun 25, 2015

Let's have this in for a while, we can always neglect the results ;) Thanks Nathaniel.

@njsmith
Copy link
Member Author
njsmith commented Jun 25, 2015

I'll just note for the record that afaict the way this works is that you
get whatever was the latest version of 3.5 (or maybe 3.6) at the moment
when Travis last rolled out a new machine image (which happens ~monthly).

A more predictable approach would be to have our test scripts pull down the
latest 3.5 prerelease and build it manually, but I can leave that until
someone feels inspired...
On Jun 25, 2015 4:13 PM, "Charles Harris" notifications@github.com wrote:

Let's have this in for a while, we can always neglect the results ;)
Thanks Nathaniel.


Reply to this email directly or view it on GitHub
#6015 (comment).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants
0