-
Notifications
You must be signed in to change notification settings - Fork 8
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
Release jupyter_kernel_mgmt 0.5.0 #33
Comments
@takluyver @kevin-bates We will start tomorrow during the jupyterlab dev meeting to pepare the release of jupyter server. This depends on a released jupyter_kernel_mgmt. What are we still missing to get a 0.5.0 release? |
I think we're ready to cut 0.5.0. @takluyver - if you agree, is this something you have bandwidth to do, or do you want me to push things through? |
If you have the bandwidth, go for it. I may be able to get to it if you can't, but I'm not going to be upset if you get there first. 😉 I've just pushed some docs changes I did earlier but failed to push at the time. |
Hmm - @takluyver I'm getting errors building the docs. The errors are relative to the high-level api rework and generate this page: https://jupyter-kernel-mgmt.readthedocs.io/en/latest/api/hl.html This is an 0.5 blocker. Here are the errors produced from
|
@kevin-bates I think the issue does not come from @takluyver last commit but from the changes I have brought to the conda env. To generate the API doc, sphinx needs to get the code it inspects installed. There were some rest of jupyter_client so I added it in the conda file but did not test, and of course, jupyter_client is not in conda repos. so the conda env was not created. Let me open a new PR to fix this. Side question, I think the published docs does not contain the API details. Even with a fix for this, I believe the API will not be created. Let's see... |
Should the readthedoc.io be configured so it first |
I've just pushed commit 86bb83c which should ensure the development version of JKM is installed on RTD before building. For building locally, you need to ensure the development version is importable, either by installing it in the same environment as Sphinx runs, or adding it to |
No, still not quite working. Something is importing |
Doh! Thanks to both of you. (I forgot about installing the dev version prior to build - sorry about that.) The high-level API page looks good. Hmm - @takluyver - just saw your update. The docs build, but I see refs to jc in two test files and docs/conf.py...
|
The build succeeds, but the autodoc part is not working at the moment. |
yep. I searched for jupyter_client and it was in the python comments, not in the source if I remember well. This is why I added it (badly) to the env. Right now none of the API pages are populated. They are created but not populated. At the single warning, I believe the autodoc are not run. To get it working on my local env, I had to install jupyter_client. |
Installing jupyter_client should fix it. But it shouldn't require that, so I'm trying to understand why it does. |
I think we have to update the conda env with
Then it wiill work |
True. I spent time also to understand why it needed jupyter_client. I think sphynx is clever and see we are referring to it (even without importing it). |
Sphinx is clever in many ways, but I don't think that's one of them 😉 |
:) jupyter_client should be removed from https://github.com/takluyver/jupyter_kernel_mgmt/blob/master/docs/api/client.rst |
The link there is an intersphinx thing, it doesn't require importing the Python module. The imports are a leftover dependency in the released version of
|
BTW to release JKM, we should first release JKP |
OK, the autodoc sections now work correctly: https://jupyter-kernel-mgmt.readthedocs.io/en/latest/api/hl.html Once we have released jupyter_protocol again, we should be able to remove jupyter_client from the |
Thanks guys. My docs built fine sans jc in env.yml, but I definitely had the dev version of jupyter_protocol in my env. Can we get JP released? I will then release 0.5 (including the removal of jc in environment.yml). |
Makes sense to first release JP if possible. Then update setup.py and conda env. jc is like https://i.pinimg.com/originals/aa/3d/fb/aa3dfb7a15b817572b1f5a00d7621975.jpg |
I've released jupyter_protocol 0.2, and removed https://readthedocs.org/projects/jupyter-kernel-mgmt/builds/10048031/ |
Awesome Thomas! Thx a lot. Really exciting to get this released. |
Yes - thank you very much! I'll take a crack at 0.5.0 right now. |
0.5.0 has been released! |
Thx Kevin. Congrats to you and Thomas for all the great work. |
@takluyver - Sigh - of course this is discovered after the release! This is from the jupyter_server build and only occurs on python 3.5. If we're going to support 3.5, I think we should add it to the build matrix. IIRC, you mentioned 3.5 would not be supported as of this month - correct? I'm kinda inclined to move jupyter_server > 3.5 actually.
|
3.5 is theoretically still alive until 2020-09-13 https://devguide.python.org/#status-of-python-branches But in practice it seems like all latest standard platforms don't use it anymore python-trio/trio#75 (comment) I am also in favor of python > 3.5 |
Maybe we were a bit overeager in removing 3.5 from CI on this repo. 😉 I'm puzzled why it fails on that line, though: jupyter_kernel_mgmt/jupyter_kernel_mgmt/hl.py Lines 49 to 51 in b109f0b
Keyword-only arguments are valid since Python 3.0, I thought. 😕 If you prefer to figure out the fix to make it work with 3.5, I'm happy enough. Otherwise, NEP 29 recommends that 3.5 can already be dropped, so I wouldn't spend too much effort on it. |
I have opened jupyter-server/jupyter_server#142 in jupyter_server to enforce 3.6. If this gets merged, we can leave it as it is. |
The syntax error on 3.5 is due to the trailing comma on the function declaration: https://github.com/takluyver/jupyter_kernel_mgmt/blob/master/jupyter_kernel_mgmt/hl.py#L49-L51 I'm fine dropping 3.5 as well and I don't think it's removal warrants another release at this time. |
Huh, I didn't know that was new. It seems like we're all in agreement to drop 3.5 support anyway, but it's good to know what the issue is in case we do need to revisit that. |
This issue is created to track the release jupyter_kernel_mgmt 0.5.0.
The text was updated successfully, but these errors were encountered: