[ Config | Root Config ] Clarify readonly #476
Conversation
Clarifies how root config of readonly==true should be handled if not supported by the platform. Signed-off-by: Rob Dolin <RobDolin@microsoft.com>
| @@ -27,6 +27,7 @@ Each container has exactly one *root filesystem*, specified in the *root* object | |||
|
|
|||
| * **`path`** (string, required) Specifies the path to the root filesystem for the container. A directory MUST exist at the path declared by the field. | |||
| * **`readonly`** (bool, optional) If true then the root filesystem MUST be read-only inside the container. Defaults to false. | |||
| If readonly is present and set to true, but not supported by the platform, the platform MAY ignore. | |||
There was a problem hiding this comment.
I think that failing may be a better option.
There was a problem hiding this comment.
I agree failing seems better.
There was a problem hiding this comment.
In general with configs, failing is much better than silently ignoring errors. A readonly path means that the user has made certain assumptions about the setup and might be quite confused if suddenly things break underneath then.
For the Microsoft case, just make it so that your runtime errors out if a user specifies readonly (though we should make it clear somewhere in the spec whether or not all runtimes need to implement all optional options).
There was a problem hiding this comment.
+1; @RobDolinMS can you either update, close, or make a good argument for why you think ignoring is preferable? 😄
There was a problem hiding this comment.
On Thu, Jun 09, 2016 at 10:25:55PM -0700, Aleksa Sarai wrote:
For the Microsoft case, just make it so that your runtime errors out
if a user specifies readonly (though we should make it clear
somewhere in the spec whether or not all runtimes need to implement
all optional options).
I've taken a stab at part of the general behavior in #559. I'm not
quite sure how to flag options which a runtime can not support while
still being certified. Some thoughts on that in #559 too.
|
@RobDolinMS was this change covered by #559? (ie, can this one be closed or updated? 😄 🙏) |
|
@tianon I think this needs to be updated to specify that not being able to implement |
|
On Sat, Nov 05, 2016 at 08:35:58PM -0700, Aleksa Sarai wrote:
If we add optional runtime support for a feature, we need to declare But ideally we can just require runtime support. It sounds like |
|
Closing this. @RobDolinMS please re-open if you disagree. |
Clarifies how root config of readonly==true should be handled if not supported by the platform.
Signed-off-by: Rob Dolin RobDolin@microsoft.com