-
Notifications
You must be signed in to change notification settings - Fork 75
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
fix tests on julia v0.6-dev #158
Conversation
Thanks for this. I haven't tried it out but it looks good, nothing controversial afaict. For mine I don't think performance issues on 0.4 are a major issue, but thanks for pointing it out. Pardon my ignorance, but what's the difference with the anonymous closures vs the previous? Fyi I threw together some very rough notes a couple of months ago about the IJulia and Interact display code, hopefully IJulia hasn't changed too much since then: https://github.com/JuliaGizmos/Interact.jl/blob/master/doc/dev%20guide.md |
re: anonymous functions, I don't really grok the new world-age system yet, but the current implementation resulted in errors caused by the fact that the Thanks for the dev guide; it's exactly what I've been looking for! |
I restarted the mac builds a few times, 0.4 went through 0.5 seems to hang. I'm merging this anyway. |
I just tried this out in IJulia on yesterday's 0.6 with cells:
Output: some deprecation warnings
No errors
world age related errors I get the same errors regarding world age with or without this change - just the names in the error messages are clearer with the named functions. I'm curious, what situation(s) did the anonymous functions fix? I didn't get the impression from the docs that they should behave any differently to named functions/methods. I had a look into it and it seems that we run into this situation on IJulia. Changing this line in IJulia to Apparently this is going to be merged which (I think) will mean we'll use |
Also, it's far from impossible that I'm completely wrong about this world age stuff, I don't fully grok it yet either. @stevengj is it true that this line in IJulia will need to change to: |
@JobJob that's extremely weird. I'm not sure why it resolved the world age errors in my case, but I agree that your solution (using eval() and then eventually invokelatest()) makes more sense. Sorry for the incorrect patch. |
Absolutely no problem at all, the contribution is appreciated. |
This brings in enough updates to ensure that the unit tests pass on Julia v0.6-dev. I'm still having issues with actually using Interact in v0.6, but I think this is at least a step in the right direction.
With this update, I can execute an
@manipulate
cell in a notebook without errors, but the output of the cell does not update as I move the slider. I don't understand enough about the way IJulia actually works to be able to fix that yet.One other note: this converts the nested functions in
IJulia/setup.jl
to anonymous closures to avoid world-age problems. That will hurt performance on Julia v0.4, so if you do go with this method, it might be worth bumping the julia requirement to v0.5 for these changes.