-
-
Notifications
You must be signed in to change notification settings - Fork 1.6k
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
Framework-agnostic container & test lifecycle #702
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.
Although it seems like a bigger change, the actual changes aren't that big and we aren't breaking any compatibiltiy I assume, so I think we can merge it soon?
@@ -0,0 +1,13 @@ | |||
package org.testcontainers.lifecycle; | |||
|
|||
public interface Startable extends AutoCloseable { |
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.
Why to we need this interface in this PR? It seems it's not used anyhwere (besides being implemented).
To use it in the testframework specific implementation?
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.
Yes, it defines a unified set of methods for both GenericContainer
and DockerComposeContainer
|
||
import java.util.Optional; | ||
|
||
public interface TestDescription { |
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.
I wonder if we need all methods as the public interface.
The way BrowserWebDriverContainer
works with it is kind of awkward, implementing the fallback if getFilesystemFriendlyName()
is not implemented itself?
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.
What would you suggest? :)
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.
I think it makes sense to not have getFilesystemFriendlyName
in here, since it seems oddly use case specific?
Or if we want to make it convenient, we could maybe have the default implementation here?
return getNameParts().map(it -> String.join("-", it));
And then the default implementation of getNameParts()
would also become
return Optional.of(new String[]{
getTestId()
});
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.
oddly use case specific
But we have this use case at least in one place.
default implementation
I don't think this is a good default implementation to be honest, I would rather keep it explicit and let framework adapters think about it, what's the best way to provide a FS friendly name of the test
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.
What about omitting the default implementation completely then?
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.
fine by me :)
|
||
public interface TestLifecycleAware { | ||
|
||
default void beforeTestBlock(TestDescription description) { |
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.
Why beforeTestBlock
and not beforeTest
? What are the semantics of block?
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.
Just in case if some Test frameworks define Test
as something more bigger and an atomic execution unit is some block of code. Arguable :)
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.
I think I would prefer removing block
in order to remove the ambiguity, but your explanation is also fine, so that's a IMO thing.
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.
ok, I summon @rnorth as a tiebreaker 😄
try { | ||
imageName = image.get(); | ||
} catch (Exception e) { | ||
imageName = "<unknown>"; |
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.
I see it's just copied from the old implementation, but does this acutally happen?
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.
well, to give you a bit of history here:
When I was implementing https://github.com/testcontainers/testcontainers-java/pull/106/files, we wanted to keep the backward compatibility as much as possible, and one of such is Logger based on image's name :)
See #87