You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Following a discussion with @NelleV I suggest to decrease the number of CI builds to help easing the backlog.
drop USE_PYTEST (now useless)
On Travis, e.g.:
py2.7 mock numpy=1.7 pandas docs
py3.4
py3.6 pandas docs pep8 delete-font-cache
osx (=2.7) mock
Note that we really should build the docs on 2.7 too (extra points if we build them on OSX instead of the Linux 2.7 job) -- right now they don't build on 2.7 at all (due to failure in examples).
On Appveyor the third job is currently a duplicate of the second(!) except that one has USE_PYTEST and not the other (which is the same). So we could be left with
py2.7 numpy=1.8 without extra dependencies
py3.5 numpy=1.10 without extra dependencies
py2.7 numpy=1.10 32-bit with extra dependencies
(the last two could be merged into e.g. py3.5 numpy=1.10 32-bit with extra dependencies, if we're still under backlog pressure)
More extra points if someone manages to build the docs on Windows...
Selfishly I am 👍 for CI build of the docs on Python 2.7 😈. FWIW, I have just successfully locally built the doc on Python 2.7 (and Linux), with #8165 in its current state (eventhough I still have to have a look to a comment @anntzer did on this PR).
@afvincent@NelleV There's a bunch of other entries in rcsetup.py that check for six.text_types. I have a strong suspicion that most of them should be replaced by six.string_types, otherwise they'll just break in the same way when a non-unicode string is passed in Py2.
I would also be inclined to agree with this conclusion: in #8165 I used six.text_type in first place because it seemed to be used quite a lot to do this kind of checking in rcsetup. And I was bitten rather bad for doing so instead of using six.text_type… However, based on my further experience in #8165, I would advise to be careful with six.string_types too, as it would also let pass strings like "Hello".encode("utf-16") (or with any other weird encoding that does not map smoothly with the ASCII one), which can have some weird/funny/unexpected side-effects too.
@QuLogic It appears that on a successful build, Appveyor will create a wheel to upload to conda (see appveyor.ymlafter_test). Perhaps we could save some time by always first building the wheel and then installing the wheel to run the tests? (especially if we don't have a C compiler cache)
The wheels are not uploaded, but just placed in the appveyor download "spot". The idea was to have a place where win users can download stuff to try fixes.
Activity
NelleV commentedon Mar 8, 2017
So… It seems that USE_PYTEST == false and true do not have the same behaviour.
I am investigating…
afvincent commentedon Mar 8, 2017
Selfishly I am 👍 for CI build of the docs on Python 2.7 😈. FWIW, I have just successfully locally built the doc on Python 2.7 (and Linux), with #8165 in its current state (eventhough I still have to have a look to a comment @anntzer did on this PR).
tacaswell commentedon Mar 8, 2017
I would rather a 3.5 test run than building the docs twice, the docs are just a smoke test (we should improve the test coverage of the validators).
anntzer commentedon Mar 8, 2017
Appveyor covers 3.5.
NelleV commentedon Mar 9, 2017
@afvincent indeed your patch will fix the documentation under 2.7.
anntzer commentedon Mar 9, 2017
@afvincent @NelleV There's a bunch of other entries in rcsetup.py that check for six.text_types. I have a strong suspicion that most of them should be replaced by six.string_types, otherwise they'll just break in the same way when a non-unicode string is passed in Py2.
NelleV commentedon Mar 9, 2017
Yep, that's the conclusion I came up with.
afvincent commentedon Mar 9, 2017
I would also be inclined to agree with this conclusion: in #8165 I used
six.text_typein first place because it seemed to be used quite a lot to do this kind of checking inrcsetup. And I was bitten rather bad for doing so instead of usingsix.text_type… However, based on my further experience in #8165, I would advise to be careful withsix.string_typestoo, as it would also let pass strings like"Hello".encode("utf-16")(or with any other weird encoding that does not map smoothly with the ASCII one), which can have some weird/funny/unexpected side-effects too.anntzer commentedon Mar 10, 2017
@QuLogic It appears that on a successful build, Appveyor will create a wheel to upload to conda (see
appveyor.ymlafter_test). Perhaps we could save some time by always first building the wheel and then installing the wheel to run the tests? (especially if we don't have a C compiler cache)QuLogic commentedon Mar 12, 2017
@anntzer That sounds like a good idea; I'm not sure why it was done this way or why we need to build the wheels every time either. Ping @JanSchulz.
jankatins commentedon Mar 12, 2017
The wheels are not uploaded, but just placed in the appveyor download "spot". The idea was to have a place where win users can download stuff to try fixes.
anntzer commentedon Mar 12, 2017
Sure, but they could still be built first and the test installation done using them.
phobson commentedon Apr 26, 2017
Quick, related questions: to me it appears that we always use python 2.7 to build the docs, no matter which version of python we've set travis to use: https://github.com/matplotlib/matplotlib/blob/master/ci/travis/test_script.sh#L38
Is that true or am I missing something?
anntzer commentedon Apr 26, 2017
That's just to run linkchecker.
phobson commentedon Apr 26, 2017
Ahh yes, i see it now.
anntzer commentedon Jul 21, 2017
(But in fact linkcheck is now builtin to sphinx so we don't even need that. See e.g. #5380.)
tacaswell commentedon Aug 10, 2017
Closing this as we are now down to:
We need to try turning the 2.7 build back on on appveyor and I might try adding a 3.6 build on OSX.