- add enum entries for other types of errors
- omit stack frames for outer error in dagit log view
Details
- Reviewers
sandyryza dish prha - Commits
- R1:b57b91ea7666: [dagit] error source
before
after
dialogs
Diff Detail
- Repository
- R1 dagster
- Lint
Lint Not Applicable - Unit
Tests Not Applicable
Event Timeline
js_modules/dagit/src/app/PythonErrorInfo.tsx | ||
---|---|---|
28 | I should probably do something with the error source in the pop up here as well |
js_modules/dagit/src/app/PythonErrorInfo.tsx | ||
---|---|---|
61 | It doesn't seem that helpful to actually see the error source (e.g. USER_CODE_ERROR) in the message. Seems more useful to follow-up with steps for recourse. e.g. Please check your repository definition for X. What class of problems are framework errors, as opposed to unexpected errors? |
requesting review to get opinions on what to do with dialog context line. My current ordered preference is
- just drop the enum name
- only show on unexpected error to request for issue filing
what about yall?
js_modules/dagit/src/app/PythonErrorInfo.tsx | ||
---|---|---|
58 | Ya i think we can iterate to something better here. My thoughts for adding this or something like it:
| |
61 | ya I was thinking thats useful for us if they copy / screenshot and send it to us but i do agree its not ideal and we can just match on the description text if we keep it |
I think just dropping the enum name is fine. I'm curious how users would separate out the steps to resolve framework vs unexpected errors.
I guess because framework errors might be instance configuration errors?
js_modules/dagit/src/app/PythonErrorInfo.tsx | ||
---|---|---|
61 | I think it's not a blocker, but I have a natural aversion to having anything in the product talk about "users" and calling anything "user-provided" code. If there's a way to be specific about the error boundary (e.g. this solid name or this schedule definition), we could avoid this distinction. |
js_modules/dagit/src/app/PythonErrorInfo.tsx | ||
---|---|---|
61 | Yeah - I think I share the same aversion. |