-
Notifications
You must be signed in to change notification settings - Fork 6
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
Prettier Stata error handling #22
Comments
I now think this is a flaw in the pystata code, so I am going to reach out to their technical support today to inquire about it. |
Here's what I heard back:
|
The underlying pystata flaw was fixed in today's Stata update, so it should be possible to handle errors in Stata code more gracefully now. |
Unfortunately, the Stata/pystata update seems to have made it so that My updated version of this kernel (with prettier error handling that works on the latest Stata update) is here: https://github.com/hugetim/nbstata |
My apology for being absent for so long. Sickness and work have kept me from coming online. Let me see if I understand the situation correctly:
|
I'm glad that you're better and back at it. 🙂
|
I'm going to be rolling out pystata-kernel (or some variant) for this upcoming Spring semester for a class I'm teaching beginning in three weeks. I've been following this issue as it seems super important for students who are learning stata to have error messages. I'm wondering what the status is here versus the fork that seems to fix it by @hugetim. Are these changes going to be merged back into to pystata-kernel or is nbstata the "new" working version? |
I think of nbstata as the new working version, myself, in short. I think the only potentially "breaking" change to be aware of is that I changed the default to echo 'None'. Let me know if any further changes would benefit your use case. |
My department's computing cluster has not yet been updated to the newest version of Stata---we only update between semesters---so it will take me some time to incorporate the changes back from @hugetim's fork. We are upgrading this week, then I will get to test it out. I agree that if this issue is important right now, use nbstata. I have added a note on this on the front page. |
Thanks. Is it documented which stata update level . update query
(contacting http://www.stata.com)
Update status
Last check for updates: 03 Jan 2023
New update available: 15 Nov 2022 (what's new)
Current update level: 28 Jul 2022 (what's new) and on another machine I get this: update query
(contacting http://www.stata.com)
Update status
Last check for updates: 03 Jan 2023
New update available: 15 Nov 2022
Current update level: 23 Aug 2022 Note that on both I can only go to the 15 November version, I can't incrementally upgrade the first one to 23 August. Again, I could be wrong about this, but I haven't been able to find a way to do it. In the future are you planning on keeping @hugetim I suppose that |
I'd say that's up for discussion. Ideally it would be compatible with all builds--or else maintain separate versions that those stuck on a non-compatible Stata build could pin to. (My company doesn't always purchase the latest license, so there's a chance I won't be able to test it with Stata 18, myself, whenever that comes out. But then maybe Stata 18 will come packaged with its own kernel.) At this point, I've only tested |
Sorry. Didn't mean to be presumptuous. I am in a similar situation to @ticoneva: my cluster/jupyterhub manager decides when to install/upgrade stata and I am trying to map out the options to make things as easy as possible for them as they don't like to deal with stata even during the best of times (no technical glitches). I should be able to test either pystata-kernel or nbstata against the latest version as our campus license keeps us current. |
Oh, nothing presumptuous came across to me. And that all makes sense. I appreciate you raising this potential issue to my attention. |
Though it is out of date now, I thought to re-up this comment posted in another place, realizing that no one may have seen it, since it was a comment on a closed pull request: #21 (comment) Oops, and for myself, I only just now noticed the recent commit which refers people to |
Stata's inability to target specify update level is quite unfortunate. I do have access to snapshots of Stata 17 on our computing cluster, so it is possible to test packages against various Stata 17 versions if there is a need. |
Nothing in the Jan 10 2023 Stata update appears relevant, but I did rerun my |
I have ported the new error handling mechanism from |
Currently, running
list, nobs
results in this error message:Preferably, it would instead display the more Stata-like message:
Unfortunately, it doesn't appear possible to catch the
SystemError
raised by pystata in the way I expected. It is instead caught by the "threading.py" code referred to in the traceback. So I'm stumped on how to proceed, myself.The text was updated successfully, but these errors were encountered: