-
-
Notifications
You must be signed in to change notification settings - Fork 2.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
tools, ci: print info to use v symlink instead of v symlink -githubci
#21471
Conversation
I haven't made any fixes yet, the CI will fail at first. But if I may, I would make that internal update already (to always the same symlink) instead of awaiting a deprecation period and just warn about the flag being obsolete while keeping it functioning. |
This should stay for several months: if '-githubci' in os.args {
setup_symlink_github()
} Note, also that the CI passed fine in 279405a , which did use a single I've explained to you multiple times already, both here and on Discord, that I want to keep this explicit command: Changes that affect I do not want to affect CIs that do use I do not know how else to explain my position. I accept, that I will look as stubborn ass to you, but so be it :-| . |
It's fine for me to keep that as it is. Also it was understood and I didn't intended to remove it I didn't removed it here. Just seeing that it is failing with the flag where it isn't without it appeared like a selling point. I'd update the CI test then to test the flag to its current capabilities. Adding the warning is fine though? As said no functionalities changed. |
A selling point not to remove it, but to extend it's capabilities while still keeping it is what I meant. |
Adding a warning is fine, as long, as it does not affect the actual functionality. |
That can be done too in the future, but not now. This has known limitations, but also known advantages too. I do not want to complect what it does with creating a symlink for now. It is a very deliberate choice, that reduces the risk of breaking things, that currently work, and that provides a way to quickly go back to using |
Yes that's totally fine. I'm already onto supporting this decision by trying to increase coverage for the cases where it is used. Just to have it mentioned: |
I already stated my preference, which is that it should stay exactly as it is for now. |
💯 As said no issues with that, and supporting it. Still wanted to mention something that might not have been thought about. Thanks for the fix 👍 |
This one contains an interesting coincidence and revelation.
The last update used
./v symlink --githubci
(double-dashed) instead of./v symlink -githubci
to cover extended cases for the flag. However, the double-dashed flag results in the default symlink being used instead of the GITHUB_PATH-symlink that is based on the current directory. Now, using the symlink with the single dash flag-githubci
fails to pass the extended test cases.This shows that using the flag actually results in worse functionality. The solution would be to already use
setup_symlink
instead ofsetup_symlink_github
(even when the flag is passed), as it fixes an issue.Based on the recent heated(while still constructive) discussion, the coincidence might look like an evil plan, setting this revelation up, but it really was just a tiny mistake that happened in the last PR.