reviewing lockfiles for 2.0 #11135
reviewing lockfiles for 2.0 #11135
Conversation
|
what's the summary of the review? discussion is pretty long, it's hard to extract the final resolution |
It is a draft, it will be posted on the Tribe repo when ready, and most likely submitted to the mailing list |
conans/cli/commands/export.py
Outdated
| add_lockfile_args(parser) | ||
| parser.add_argument("-l", "--lockfile", action=OnceArgument, | ||
| help="Path to a lockfile.") | ||
| parser.add_argument("--lockfile-no-strict", action="store_true", |
There was a problem hiding this comment.
It could be a complete lockfile, but you changed one requires = xxx in your recipe and use against the lockfile, want to "lock whatever you can, but if there is something you cannot lock, then evaluate it as normal". I am not sure -lockfile-partial would be closer to that than no-strict
There was a problem hiding this comment.
"lock whatever you can, but if there is something you cannot lock, then evaluate it as normal"
So it applies the lockfile partially. Honestly, the strict might mean anything.
There was a problem hiding this comment.
Good for me. @czoido @franramirez688 ? --lockfile-partial better naming than --lockfile-no-strict?
There was a problem hiding this comment.
I have to say I don't love --lockfile-partial either... maybe we need more ideas.
There was a problem hiding this comment.
I have used --lockfile-partial I agree it is slightly better than --lockfile-strict in the meaning.
If anyone else got other ideas, they are welcome.
Review from Tribe proposal conan-io/tribe#34:
--lockfile=conan.lock, automatically by commands that implement this argumentconan lock createwill also default to output "conan.lock", even without--lockfile-outargument--lockfile-outwill allow to save a lockfile as a result, even if a lockfile was not provided as input--lockfile-no-strict(arg name not great, we can try to think a better one) will allow non-lockedrequiresto not fail--cleanargument that was existing only forconan lock createhas been renamed to--lockfile-clean, because it now can be used by the other commands that save a--lockfile-out.Second round of review:
conan.lockis assumed by default besides theconanfile, not in thecwd--lockfile-no-strict=>--lockfile-partial, it seems a bit more aligned with the actual behavior