v 3.0.71 gives an error. downgrade to 3.0.67 fixes it

[JPSefe]

Describe the bug

LIB018]: "[email protected]" does not contain the following: js/bootstrap-image-gallery.js, css/bootstrap-image-gallery.css Valid files are package.json, .npmignore, README.md, css, img, index.html, js

To Reproduce

Steps to reproduce the behavior:

1. have the 3.0.71 as global tool installed
2. in libman.json:

{
  "provider": "unpkg",
  "library": "[email protected]",
  "files": [ "js/bootstrap-image-gallery.js", "css/bootstrap-image-gallery.css" ],
  "destination": "temp/blueimp-bootstrap"
}

1. failes:

LIB018]: "[email protected]" does not contain the following: js/bootstrap-image-gallery.js, css/bootstrap-image-gallery.css Valid files are package.json, .npmignore, README.md, css, img, index.html, js

Expected behavior

no issue

Workaround

dotnet tool update microsoft.web.librarymanager.cli --version 3.0.67 --global --allow-downgrade

[jimmylewis]

Another report workaround is to clear the cache

• libman cache clean, or
• delete %localappdata%\.librarymanager\cache on Windows, or the corresponding directory on other systems
  • libman cache list will show where the cache resides

[JPSefe]

Shouldn't it clean the cache then after an upgrade?

I can verify that the clean cache works, but it feels like a workaround:

• dotnet tool update microsoft.web.librarymanager.cli --global
• libman restore <-- FAIL
• libman cache clean
• libman restore <-- WORKS

Shouldn't it just clean the cache then after an upgrade to prevent these errors which seem unrelated?

[andyjhall]

+1 for cleaning the cache

[KeithDickson]

This seems related to a similar issue being tracked in Visual Studio.

https://developercommunity.microsoft.com/t/Unpkg-libraries-in-Libman-are-no-longer/10906411?q=Libman

I posted a comment in that bug report with the workaround described here, which worked for me. Not sure if these are directly related problems, but the symptoms are basically the same (failure to restore from unpkg). Clearing my cache manually allowed a previously-failing restore operation to work.

There doesn't appear to have been any activity on the issue here in several weeks, so I'm giving this a bump to hopefully get a better long-term fix.