-
Notifications
You must be signed in to change notification settings - Fork 357
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
Remove superfluous data structure for secondary events #2502
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@jougs Thanks for tidying this up. I am still a bit in doubt why it is okay to remove the create secondary machinery as you do. Maybe easier to discuss that one-to-one.
@heplesser, @med-ayssar: Thanks for your valuable input. In a video call with @heplesser and @suku248 on Friday, we concluded that these changes probably don't go far enough and we should revisit the handling of secondary events more thoroughly. I will mark this as draft for now with the option of either adding more commits later on or closing it in favor of another PR. |
Co-authored-by: Hans Ekkehard Plesser <hans.ekkehard.plesser@nmbu.no>
Co-authored-by: Hans Ekkehard Plesser <hans.ekkehard.plesser@nmbu.no>
I have added a bunch of new commits to this. It should be much more complete now. The description was updated to explain what was added in detail. This should now also fix #2492. |
@med-ayssar: I don't have any failures locally anymore, let's see what the CI thinks. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@jougs Nice work! I added a few more comments. Unfortunately, I haven't spotted yet where things go wrong :(.
Co-authored-by: Hans Ekkehard Plesser <hans.ekkehard.plesser@nmbu.no>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@heplesser: I've addressed your concerns in my latest commits and comments
@heplesser: What do you mean by "things go wrong?" I don't see anything going wrong anymore. |
When I reviewed, the most recent run of the testsuite on Github still showed many failing tests, but clearly that is fixed now :). |
@jougs Could you remove the |
…tor into refactor_module_system
Removed in a63a3bb. However, I'm a bit shocked that this was even in (in master). If we have a bug, we ave a bug and I don't think that such a workaround should ever be in master just to allow other PRs to go in. How would I even have known that this needs to be removed if you did not tell me? |
I agree with you that this is borderline and we need to discuss how to handle situations like this in the future. In this case, #2379 fixed a concrete bug (problems copying all kinds of models), but also brought and improved test that detected the error described in #2492. This left the choice of either delaying #2379 until #2492 was fixed or weakening the test. I opted for the latter and put into #2492 so that we would not forget to remove the restriction from the test. That info would probably have been more visible in the issue description, maybe even in form of a to-do list. |
This PR does the following:
nestkernel/secondary_event.h
ModelManager
in favor of the events that are stored in the connection modelsOn the way, I cleaned up the includes in the ModelsModule and some other minor things.
Update on October 18:
SecondaryEvent
spristine_supported_syn_ids
fromSecondaryEvent
sConnectionManager::connect_()
DynamicLoaderModule
from being used undefinedFixes #2492.