Closed fjetter closed 3 years ago
No, I don't think we have such a thing. We have tried to use github projects and whatever they called their meta-issues on there.
The simple approach does work: an issue listing all the connected issues and their status.
I am personally pro the proliferation of project-specific labels. For instance, I'd like to have a parquet
label on dask/dask. I got the sense from #50 that we didn't want to go down that path, so I've held off. What do you think @jacobtomlinson, should we standardize or just let people add labels as convenient?
As you can see in #50 I started walking the path of standardization and automating that process. However I think there is a good case to be made for repo specific labels. I would be lost in dask-cloudprovider
without the vendor labels for sure.
That added enough complexity to automating the standardization that I got distracted by other things.
Ideally I would like to see broad labels like type/bug
standardized across all our repos. But leave project specific labels in place.
Yeah that makes sense. It wasn't clear to me that those were meant to be a subset of total labels. I might go add a parquet
label to dask/dask then :) @jrbourbeau is that cool with you?
Ideally I would like to see broad labels like type/bug standardized across all our repos. But leave project specific labels in place
Yeah, that sounds like a good place to work towards
@fjetter you should feel free to add project specific labels to distributed
directly or raise an issue in the distributed
repo if you'd like to engage other folks beforehand.
@jsignell a parquet
label seems reasonable. For example, some folks might want to look at parquet
issues specifically instead of all I/O-related issues (which we currently use the io
label for).
Do we have any convention regarding problem specific labels? I have the particular use case that there is a large scale problem bothering us (https://github.com/dask/distributed/issues/4413) for a while now and there are multiple issues connected to this (Example https://github.com/dask/distributed/issues/4698 but there are many more). I do not think the individual issues can be sensibly fixed on their own but I also do not take it for granted that they will be fixed automagically. The root cause issue requires a refactoring of the code base and I'd like to get back to the individual tickets once the refactoring is done. One possibility to keep track of this would be to use labels, either very specific ones
worker-state-machine
or more broader ones likestability
Another way to handle this would be to maintain a list with refs to the individual tickets (w/ or w/out closing them) and I was wondering if there are any best practices or suggestions about smth like this.Note there is an issue open about standardising issues across projects https://github.com/dask/community/issues/50