-
Notifications
You must be signed in to change notification settings - Fork 972
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
[bug] Automated config update exception needed for conan config install commands to prevent catch-22 #7297
Comments
We added to Conan 1.27 two commands:
To be able to remove configurations you don't want anymore. Would they solve the issue? What we didn't want to add is to quiet or silently pass over missing configuration origins. That should be an error and that should raise, then Conan provide the tool to manage the What is not clear is why when you do |
Yes, there was an auto scheduled config due. I can't make any more conan calls without it trying to check for updates from the old, non-existent location. |
Ok, I see what is happening. You should at least be able to However, for cases where it fails (for example, no internet connection), it is likely that users don't want to remove the origins. So most likely, the scheduled task should always succeed in the sense that it update the timestamp, and wait for next scheduled call. What do you think @uilianries? I'd say to improve this for next release. |
In this case, the command I was trying to run was to sent a new origin, which I was unable to do.
It seems like there should be an exception to the update check for the I updated to 1.27, and it still suffers from this catch-22:
It's very easy to repro -- just fork your config repo, set |
Thanks for the detailed report. You are right, this is a completely a bug, needs to be fixed. Please @uilianries have a look. |
No worries. It's a nice feature. You want to make an omelette, you need to break a few eggs. |
@memsharded Not only this problem, but if there is no config listed in the index we will have a problem. I think we should put a warning, and ignore the error. |
Signed-off-by: Uilian Ries <uilianries@gmail.com>
Signed-off-by: Uilian Ries <uilianries@gmail.com>
Signed-off-by: Uilian Ries <uilianries@gmail.com>
Signed-off-by: Uilian Ries <uilianries@gmail.com>
* #7297 fix config install when scheduled Signed-off-by: Uilian Ries <uilianries@gmail.com> * #7297 Do not run sched config and empty file Signed-off-by: Uilian Ries <uilianries@gmail.com> * #7297 Do not sched config with empty configs file Signed-off-by: Uilian Ries <uilianries@gmail.com> * #7297 config is not only subcommands Signed-off-by: Uilian Ries <uilianries@gmail.com>
Merged the fix, will be release in 1.28 |
I upgraded to v 1.28.1 and can confirm this is fixed for me now. Thanks! |
Great, thanks for checking it and reporting @michaelmaguire! |
Note: Don't forget after testing this to use |
Relates to the feature discussed in #7282
I think I just ran into a bug with this.
I had made a personal fork of our teams config to test out making this change, and used it to set up my own client with:
I tested that, it worked fine, I got approval and landed the PR on our team's repo, and deleted my fork.
Today I went to perform a conan operation, which failed because the fork no longer exists.
So I thought, "no problem" I'll just re-set up my config to be from my team's repo:
It now won't let me fix the problem, because before the new config install location can be set, it tries to check the old, non-existent location and errors out.
Of course I worked around this, but can we please add an exception to the automated scheduler checks for
conan config install
commands themselves?The text was updated successfully, but these errors were encountered: