Duplicate Library (10922)

Oops! I might have been a little sloppy in my last pio publish and sent the wrong information…? I ended up creating a second U8glib-HAL library under my personal GitHub account name instead of updating marlinfirmware/U8glib-HAL. So library number 10922 should be deleted, expunged, and de-rezzed.


Was publication less than 72 hours ago? If yes, use pio publish unpublish

1 Like

I’ve just deleted that package. Maybe, we should add confirmation information before publishing.

1 Like

Yes, that would be helpful! It wasn’t clear to me that the published library would end up being from my fork instead of updating the existing library according to the library.json. Did it get that information from the git remote of my working copy, or…? I did change the email of the author —assuming it would just apply to the new version onward— but that might be a separate issue from why it chose my fork for the identifier.

I can fix up the situation for my case by carefully using the U8glib-HAL checked out directly from the main repo.

Would I have caught and fixed it if I had been given a warning and instructions? Maybe! Before publishing I’d like to have seen a list of all the existing libraries having the same library.json information, with a mark on the one that I’m about to update, or a warning if I’m about to create a new one. As a maintainer, if I do want to change “identifying information” on a library, I’d like to be able to specify the library’s ID number, (or otherwise specify the “root” library, if there is such a thing in the db).

I can see why the library db should include alternative forks, and I imagine it’s a fun challenge to keep it working as well as it does. If it doesn’t have the concept of a “root version” at this time, that might be something to consider adding later. When there’s more than one with the same version identifier, but a different source URL or fork, pio prints the “library ambiguous” message. Would it be better or worse if pio assumed you meant the “root version” of the library whenever there’s more than one library with otherwise-matching information?

@thinkyhead thanks again for all reports and feature requests. Everything is implemented! :rocket:

Could you try pio upgrade --dev and re-test with pio package publish?

1 Like

Oops, I did a publish apparently with the non-development version of PIO and it did the same thing. Sorry about that! It’s been over 72 hours, so can you delete it ? (12378) and I will give the devel version a try now to see that it does the right thing.

Right, so now it helpfully warns me to add an --owner flag. Much better!

A package with the name `U8glib-HAL` is already
published by the next owners: marlinfirmware

I noticed that the --owner MarlinFirmware version of the library also got updated, so I assume that was picked up automatically by the platformio library scraper.

@ivankravets — I don’t know how library 12378 got published, since I didn’t run any platformio commands to publish my personal fork that I can recall. I simply updated the version over at the MarlinFirmware/U8glib-HAL repository and left it to the automation. To avoid any more confusion I will simply delete my personal fork completely. In the meantime, please delete 12378.

The library 12378 has been deleted!

Is there a way to quiet/ignore the warning/confirmation when publishing from a github CI workflow? I’m looking at automating publishing multiple libs to PIO from github CI workflows. It is tedious to have to release each libs 2x (1 for github, 1 for PIO).

You need this command https://docs.platformio.org/core/userguide/package/cmd_publish.html

Oh I didn’t see the --non-interactive option. That’s exactly what I was looking for.

1 Like