Infrastructure for mounting more than one webview at a time #12382
Conversation
rchiodo
requested review from
DavidKutu,
DonJayamanne,
greazer,
IanMatthewHuff,
joyceerhl and
techwithtim
|
Forgot to mention, I also upgraded Mocha. This was necessary to fix this bug: Skipped tests were causing the mounted web panels to not be cleaned up because the setup function for the skipped tests would run but the teardown wouldn't. |
| @@ -75,11 +75,11 @@ suite('Daemon', () => { | |||
| ); | |||
| }); | |||
| teardown(() => { | |||
| pythonProc.kill(); | |||
| pythonProc?.kill(); | |||
There was a problem hiding this comment.
| pythonProc?.kill(); | |
| pythonProc?.kill(); // NOSONAR |
| if (connection) { | ||
| connection.dispose(); | ||
| } | ||
| pythonDaemon.dispose(); | ||
| pythonDaemon?.dispose(); |
There was a problem hiding this comment.
| pythonDaemon?.dispose(); | |
| pythonDaemon?.dispose(); // NOSONAR |
There was a problem hiding this comment.
Hmm, guess we have ignored Sonar analysis for some files
| const response = waitForMessageResponse(ioc, () => exportButton!.simulate('click')); | ||
| await waitForPromise(response, 100); | ||
| exportButton!.simulate('click'); | ||
| await sleep(100); |
There was a problem hiding this comment.
This change and the one below (copyToSource) seem like ordering changes. Were the waits just not needed?
There was a problem hiding this comment.
I got rid of waitForMessageResponse. It was rather hacky. However in this case the wait for not doing export still has to be there.
In reply to: 441176542 [](ancestors = 441176542)
|
Kudos, SonarCloud Quality Gate passed!
|
For #11445
In order to get a functional test that opens an interactive window and a notebook, I had to refactor all of our webview stuff in the tests to not assume a single instance.
This PR is about creating just this change (test isn't there yet for testing both at the same time).
This was accomplished by creating a new service that mounts the webview.