search

conda-forge / pillow-feedstock

A conda-smithy repository for pillow.
BSD 3-Clause "New" or "Revised" License
2 stars 30 forks source link

Bump to freetype 2.6, fix Tcl/Tk on Mac, better log output #8

Closed jakirkham closed 8 years ago

jakirkham commented 8 years ago
conda-forge-linter commented 8 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.

ocefpaf commented 8 years ago

@jakirkham see #6

jakirkham commented 8 years ago

Sorry, @ocefpaf, missed that one. Closing in favor of PR ( https://github.com/conda-forge/pillow-feedstock/pull/6 ).

ocefpaf commented 8 years ago

No problem @jakirkham. Just wanted to save your time. Although, I won't be able to focus on this right now. Feel free to re-open if you want to take over this issue.

jakirkham commented 8 years ago

Feel free to re-open if you want to take over this issue.

I might try the --debug flag, but I don't know that I will get any further than you did.

jakirkham commented 8 years ago

I have rebased on top of your PR ( https://github.com/conda-forge/pillow-feedstock/pull/6 ) and added the patch for the --debug flag and am using it for the build now. Let's wait and see what this tells us.

jakirkham commented 8 years ago

FYI, we have some other things we need to fix too. This is using the system Framework for Tk on Mac. That's a no go for us.

jakirkham commented 8 years ago

See this issue ( https://github.com/python-pillow/Pillow/issues/1168 ). Basically, they tried to address it, but it still appears to be happening on Mac. So, we can't disable Tcl/Tk outright. Not sure where these arguments are coming from.

I think I'm just going to use this patch from conda-recipes instead. This will give us Tcl/Tk support on Mac using the tk package instead of the system framework. We can discuss with upstream about a better way to do this going forward.

jakirkham commented 8 years ago

Had to tweak the patch because of improvements made for handling frameworks. It's sort of like past me is kicking future me. 😆 Does seem to work though. Will push soon.

jakirkham commented 8 years ago

@patricksnape, I'm borrowing your multiprocessing patch for Windows. Do we still need this? Would it help with our Python 3.4 32-bit dilemma? Do you have any other ideas?

jakirkham commented 8 years ago

So, I don't know guys. It seems like this Windows Python 3.4 32-bit thing is a tough nut to crack. 🌰 Though this pinning really isn't compliant and it is really causing me some pain. As none of us (myself included) have the time and/or know-how to figure this out, I'd like to propose a temporary solution.

Let's add a skip for Windows Python 3.4 32-bit and merge. We can open an issue mentioning this is broken Windows Python 3.4 32-bit and someone with the time and know-how can get it in.

Right now I'd just like to get this in a better state w.r.t. to our stack.

ocefpaf commented 8 years ago

Let's add a skip for Windows Python 3.4 32-bit and merge.

:+1: Don't even bother adding the skip. Let the failure show in the badge so we can remember to get back to this in case we miss the issue.

jakirkham commented 8 years ago

Awesome. Thanks, @ocefpaf. I debated proposing that, but thought people would be against it. SGTM.

I've updated the summary above to better explain what has happened here. Basically your pinning PR is absorbed and Tcl/Tk is correctly linked against on Mac.

jakirkham commented 8 years ago

Alright, I'm not waiting for the Windows CI as we already know how that is going to go.