search

conda-forge / cyclus-feedstock

A conda-smithy repository for cyclus.
BSD 3-Clause "New" or "Revised" License
1 stars 14 forks source link

bump Cyclus version #42

Closed bam241 closed 5 years ago

bam241 commented 5 years ago

Checklist

conda-forge-linter commented 5 years ago

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

bam241 commented 5 years ago

@conda-forge-admin please rerender

bam241 commented 5 years ago

@conda-forge-admin please rerender

bam241 commented 5 years ago

@conda-forge-admin please rerender

bam241 commented 5 years ago

@conda-forge-admin please rerender

conda-forge-linter commented 5 years ago

Hi! This is the friendly automated conda-forge-webservice.

I tried to re-render for you, but it looks like there was nothing to do.

bam241 commented 5 years ago

@conda-forge-admin please rerender

bam241 commented 5 years ago

@conda-forge-admin please rerender

conda-forge-linter commented 5 years ago

Hi! This is the friendly automated conda-forge-webservice.

I tried to re-render for you, but it looks like there was nothing to do.

bam241 commented 5 years ago

@conda-forge-admin please rerender

conda-forge-linter commented 5 years ago

Hi! This is the friendly automated conda-forge-webservice.

I tried to re-render for you, but it looks like there was nothing to do.

xylar commented 5 years ago

@bam241, I can see that this has been quite a struggle.

Looking at https://github.com/cyclus/cyclus/blob/master/setup.py, I don't wee a clear reason that cyclus shouldn't support python 3.7. I wasn't able to look at the most recent built that tried py37 but if I had to guess, there must be some requirement or other that hasn't yet been built on conda-forge with py37 and that's the reason for the failure. Is that right?

If so, I suggest we work together on getting all of the requirements built with python 3.7, rather than opting to not support it here.

I'm going to try a local python 3.7 build on my machine and see if I can figure out what's going wrong.

bam241 commented 5 years ago

@xylar the setup.py is only part of the install, (we actually use install.py that call later the setup.py the problem with python37 are:

xylar commented 5 years ago

Okay, I see. Thanks for clarifying. If you are a developer of cyclus, these are probably things that need to be addressed but obviously not for this build.

Your current solution of pinning python seems to work. I'm not sure why your skip command with py37 wasn't working. It looks fine to me. But what you have is also okay.

bam241 commented 5 years ago

@conda-forge-admin please rerender

bam241 commented 5 years ago

@xylar do you think I can manually tweak the different file to get ride of python3.7 ?

xylar commented 5 years ago

No, that's not really a good solution. It'll just get overwritten the next time you try to rerender.

So I guess the problem is just that the bot isn't picking up on the fact that py37 CI shouldn't be running and isn't disabling it? That's not really all that big a deal. You can merge and the py36 versions will get built and uploaded even if the py37 build continues to fail. It's a bit irritating but it won't break anything.

bam241 commented 5 years ago

ok thx!

xylar commented 5 years ago

Maybe hold off on merging a little longer. I'm going to see if I can figure out why the skip isn't working. It's pretty odd...

xylar commented 5 years ago

Would you be willing to try the following?

  skip: True  # [win or py27 or py>36]

See if it rerenders without py37. Even if not, maybe the py37 builds will succeed trivially (producing no package). For whatever reason, py>3.6 works on my linux laptop (the build does nothing, as expected) whereas py37 leads to an attempted build that fails on the cython incompatibility you mentioned.

bam241 commented 5 years ago

@conda-forge-admin please rerender

conda-forge-linter commented 5 years ago

Hi! This is the friendly automated conda-forge-webservice.

I tried to re-render for you, but it looks like there was nothing to do.

xylar commented 5 years ago

Maybe there was a misunderstanding. I'm asking if you would first edit the meta.yaml with the change above and then rerender. You can edit on GitHub.

bam241 commented 5 years ago

sorry I did not pay attention... my push did not went though

bam241 commented 5 years ago

@conda-forge-admin please rerender

bam241 commented 5 years ago

now the test are gone :(

xylar commented 5 years ago

The bot is clearly confused. It deleted the python 3.6 build as well. I'm afraid this is beyond my expertise. If the bot were behaving, I think the build would work. But I can't be any help with the bot. I'd ask for help again on Gitter. Describe in more detail which skip lines you've tried and what the unexpected effects were.

xylar commented 5 years ago

Or I can post something...

xylar commented 5 years ago

One last idea, could you try the following:

 skip: True  # [win or py27 or py>=37]

This variant also worked on my laptop and maybe it'll work with the conda-smithy bot.

xylar commented 5 years ago

Okay, so the suggestion is to just limit python <=3.6 in the host section (basically as you were before). Then python 3.7 CI will still run but it will finish trivially.

Hmm, that suggestion on Gitter was immediately deleted so maybe the commenter thought better of it. I still think it's a reasonable option.

bam241 commented 5 years ago

@conda-forge-admin please rerender

conda-forge-linter commented 5 years ago

Hi! This is the friendly automated conda-forge-webservice.

I tried to re-render for you but ran into some issues, please ping conda-forge/core for further assistance.

bam241 commented 5 years ago

@conda-forge-admin please rerender

conda-forge-linter commented 5 years ago

Hi! This is the friendly automated conda-forge-webservice.

I tried to re-render for you but ran into some issues, please ping conda-forge/core for further assistance.

xylar commented 5 years ago

I think the bot is really confused. Given that another attempt at rerendering didn't work, I'd suggest that you make a copy of the meta.yaml, do a git reset --hard to the current master and copy back the meta.yaml, commit it, and do a git push --force. That should clean things up and give you a fresh start on this PR. Hopefully a rerender after that will work.

bam241 commented 5 years ago

@conda-forge-admin please rerender

conda-forge-linter commented 5 years ago

Hi! This is the friendly automated conda-forge-webservice.

I tried to re-render for you but ran into some issues, please ping conda-forge/core for further assistance.

bam241 commented 5 years ago

@conda-forge-admin please rerender

conda-forge-linter commented 5 years ago

Hi! This is the friendly automated conda-forge-webservice.

I tried to re-render for you but ran into some issues, please ping conda-forge/core for further assistance.

bam241 commented 5 years ago

@conda-forge-admin please rerender

xylar commented 5 years ago

Really odd. For reasons I don't understand, it tried to build with python 3.7.

xylar commented 5 years ago

@bam241, I seem to have coaxed the bot to do our bidding but it took an odd sequence of steps. If you want, you can try them here and see if they work for you.

bam241 commented 5 years ago

@conda-forge-admin, please rerender

xylar commented 5 years ago

One of the tests failed under my build. Hopefully that's something you can figure out.

xylar commented 5 years ago

Just a tip. You seem to like to merge, but a git reset --hard to the remote branch will get you to the same state without an extra merge commit.

bam241 commented 5 years ago

@conda-forge-admin, please rerender

bam241 commented 5 years ago

Just a tip. You seem to like to merge, but a git reset --hard to the remote branch will get you to the same state without an extra merge commit.

thx for the tips I did not know :)

bam241 commented 5 years ago

@conda-forge-admin, please rerender

bam241 commented 5 years ago

thank you @xylar

xylar commented 5 years ago

Okay, it looks like there are still some build issues to short out but it looks like you're on track. Sorry this has been such a frustrating process. Once you get the feedstock back in shape, it shouldn't be so bad the next time you update versions (fingers crossed). I'll leave you to it and close my test branch.