Use a set when calculating resolve names#18939
Conversation
|
Is it not unexpected to get duplicates here? Would be bad to hide a configuration error.. rather than giving an error (or at least a warning) Do you have an example for what scenario you have where duplicates are a valid thing to have? |
|
Ah, overlooked this was for the error message only. Nvm me then :) |
kaos
left a comment
There was a problem hiding this comment.
I still wonder why you'd get duplicates though, and can you export them properly still for those that have dupes?
|
So I did |
|
Feels like there might be something registered twice then... perhaps 🤷🏽 👀 |
|
as such, that feels like a bug we want to squash rather than bury. I'll dig a little. |
|
OK- now I see why it gets duplicated, and it's all right by design. As we list all user resolves along with all tool resolves, and in case you have a custom lockfile for a tool, that gets duplicate entries in this list. So all good. 👍🏽 |
|
It was reassuring to see there was no duplicate entries registered, which helped guide me to look in the proper spot (which ironically was in the PR diff context already) |
|
Thanks for doing the homework <3 |
I saw errors with duplicates, so using a `set` (which is OK since this gets `sorted`) to stop that.
I saw errors with duplicates, so using a
set(which is OK since this getssorted) to stop that.