You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
I'm using cargo deny in a multi-crate workspace that is proprietary/private. In this case, the tool complains about empty license values for crates within the workspace, even though it seems I should be allowed to have this (that is, in my perception cargo deny should be about checking my dependencies more than about checking my own crates -- or it should allow me to have separate policies for these two categories).
Describe the solution you'd like
I think not warning about licenses for crates in the current workspace would be a decent default.
Describe alternatives you've considered
A more powerful alternative would be to have a whitelist for crates (I think there's a whitelist for licenses, but not for crates). For example, I also have the ring crate in my dependency graph, which does not have a license specified because a number of licenses are involved. It would be good if there was also a way to silence this output after I've audited the licenses involved. (Bonus points, of course, if I could silence this output only as long as the licenses don't change, probably involving some hash thing that I could put in the configuration.)
The text was updated successfully, but these errors were encountered:
This should be solved by #103, I just haven't published a new version yet with that change.
For the ring case, this is the intent of the [[licenses.clarify]] array, eg here is our config for one of our projects.
[[licenses.clarify]]
name = "ring"# SPDX considers OpenSSL to encompass both the OpenSSL and SSLeay licenses# https://spdx.org/licenses/OpenSSL.html# ISC - Both BoringSSL and ring use this for their new files# MIT - "Files in third_party/ have their own licenses, as described therein. The MIT# license, for third_party/fiat, which, unlike other third_party directories, is# compiled into non-test libraries, is included below."# OpenSSL - Obviouslyexpression = "ISC AND MIT AND OpenSSL"license-files = [
{ path = "LICENSE", hash = 0xbd0eed23 },
]
This configuration is kind of tedious because it's per project at the moment. Ideally it would be resolved in the upstream crate, but briansmith/ring#902 never got any attention. Another idea that I've had is basically having "overlay" databases in git that you can add to your config(s) that are queried if a crate eg doesn't have a readable license so that this kind of configuration doesn't have to repeated.
Ah yes, I just figured this out from the book. BTW, hiding a link to the book in a badge makes it really easy to overlook, I usually only closely look at the badges if I'm looking for something that's usually kept there (like links to API docs, CI, coverage perhaps).
Is your feature request related to a problem? Please describe.
I'm using cargo deny in a multi-crate workspace that is proprietary/private. In this case, the tool complains about empty
license
values for crates within the workspace, even though it seems I should be allowed to have this (that is, in my perception cargo deny should be about checking my dependencies more than about checking my own crates -- or it should allow me to have separate policies for these two categories).Describe the solution you'd like
I think not warning about licenses for crates in the current workspace would be a decent default.
Describe alternatives you've considered
A more powerful alternative would be to have a whitelist for crates (I think there's a whitelist for licenses, but not for crates). For example, I also have the
ring
crate in my dependency graph, which does not have alicense
specified because a number of licenses are involved. It would be good if there was also a way to silence this output after I've audited the licenses involved. (Bonus points, of course, if I could silence this output only as long as the licenses don't change, probably involving some hash thing that I could put in the configuration.)The text was updated successfully, but these errors were encountered: