Skip to content

Windows: Fix compile after SB interface#277

Merged
mavenugo merged 1 commit intomoby:masterfrom
microsoft:10662-compile3
Jun 12, 2015
Merged

Windows: Fix compile after SB interface#277
mavenugo merged 1 commit intomoby:masterfrom
microsoft:10662-compile3

Conversation

@lowenna
Copy link
Member

@lowenna lowenna commented Jun 11, 2015

Signed-off-by: John Howard jhoward@microsoft.com

@swernli. After Sandbox was refactored to an Interface, Windows compilation was broken again. This fixes it up.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I know it can be frustrating to see windows builds broken all the time. But this cannot be a non-windows version. I don't think we should duplicate the interface itself

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@mrjana - Why not? I don't understand. Based on previous statements, at this time we do not expect Windows to be using Sandbox, which is why _unsupported is set to build on Windows (via +build !linux). This change simply addresses the compile break from introducing the interface. As I've stated, we still fully expect docker (via libnetwork) to be passthrough on Windows down to the Hyper-V virtual switch via HCS and not using sandbox. What sandbox does on Linux has no parallel on Windows - the networking stacks are so vastly different.

(BTW, at the time of submission, I did have Windows compilation working. However, revisiting, it appears that is not the case, not sure why. Will update this PR or submit a new one when that is solved - working on it pri1. However, the sentiment of this PR still remains)

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@jhowardmsft I agree with @mrjana. The main reason for having a generic Sandbox interface and platform specific implementation is to provide just that support. Its perfectly understandable that the implementation of that API interface in windows could be different (or even be absent for the fore-seeable future). But that detail must be kept on the implementation and keep the API free of platform specific details.

Yes, without a proper CI, Linux specific implementation commits are inadvertently breaking other platforms. As Maintainers we will be extra cautious. And infact push harder to have a good CI in place for cross-platform testing.
I think that will solve the nagging issue that you are facing.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@mavenugo Sure, I can rework this if you feel necessary (I tried and failed on 2 attempts - it's really not that easy). But, I really do have to ask why. You're asking to keep an interface in a platform which isn't being used, requiring far more code to be pulled in for compilation and execution path than is necessary, and making x-platform CI harder for everyone. That IMO simply doesn't make sense. Not meaning to vent - trying to be constructive, please take it that way :)

@lowenna
Copy link
Member Author

lowenna commented Jun 12, 2015

libnetwork
Waiting for CI

@mrjana
Copy link
Contributor

mrjana commented Jun 12, 2015

LGTM

Signed-off-by: John Howard <jhoward@microsoft.com>
@mavenugo
Copy link
Contributor

LGTM.

mavenugo added a commit that referenced this pull request Jun 12, 2015
Windows: Fix compile after SB interface
@mavenugo mavenugo merged commit 3be4889 into moby:master Jun 12, 2015
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants