feat(material-experimental): add test harness for mat-tooltip #16676
Conversation
googlebot
added
the
cla: yes
crisbeto
force-pushed
the
tooltip-test-harness
branch
from
c878a24 to
0b1e1fb
Compare
crisbeto
changed the title
Sets up a test harness for the current Material tooltip.
crisbeto
force-pushed
the
tooltip-test-harness
branch
from
0b1e1fb to
230d552
Compare
crisbeto
added
P3
crisbeto
requested review from
andrewseguin,
jelbourn and
mmalerba
as code owners
|
|
||
| // TODO(crisbeto): this fails because we have a setTimeout inside the tooltip which | ||
| // can't be flushed. Talk to Miles about how we should approach this in harnesses. | ||
| // it('should be able to close a tooltip', async () => { |
There was a problem hiding this comment.
@mmalerba In the tooltip's hide method we do something like setTimeout(/* some close logic */, 0) which I wasn't able to flush during the unit tests. I tried awaiting a setTimeout, but it didn't help. I could probably get it to work if I use fakeAsync with a tick for the test, but that means that whoever is consuming the harness would have to know to that they have to use it. How should we handle this case? I'm guessing that we don't want to lock ourselves into using fakeAsync or async around some tests.
There was a problem hiding this comment.
When you say awaiting a setTimeout, what did that code look like?
There was a problem hiding this comment.
I had something like this:
function wait(time: number) {
return new Promise(resolve => {
setTimeout(resolve, time);
});
}Then I had await wait(20000) right before the assertions, but the timer still hadn't fired.
| */ | ||
|
|
||
| export type TooltipHarnessFilters = { | ||
| id?: string; |
| /** Gets an attribute value from the tooltip trigger. */ | ||
| async getAttribute(attributeName: string): Promise<string|null> { | ||
| return (await this.host()).getAttribute(attributeName); | ||
| } |
There was a problem hiding this comment.
I think we should just let people do host().getAttribute themselves if they really need to get this level of flexibility
|
|
||
| // TODO(crisbeto): this fails because we have a setTimeout inside the tooltip which | ||
| // can't be flushed. Talk to Miles about how we should approach this in harnesses. | ||
| // it('should be able to close a tooltip', async () => { |
There was a problem hiding this comment.
When you say awaiting a setTimeout, what did that code look like?
|
Closing since this too out of date now. I'll take whatever I can from here and resubmit. |
This is redo of angular#16676. Adds a test harness for `mat-tooltip`, as well as an API to move the pointer away from an element, as an opposite to `TestElement.hover`.
This is redo of angular#16676. Adds a test harness for `mat-tooltip`, as well as an API to move the pointer away from an element, as an opposite to `TestElement.hover`.
This is redo of angular#16676. Adds a test harness for `mat-tooltip`, as well as an API to move the pointer away from an element, as an opposite to `TestElement.hover`.
This is redo of #16676. Adds a test harness for `mat-tooltip`, as well as an API to move the pointer away from an element, as an opposite to `TestElement.hover`.
|
This issue has been automatically locked due to inactivity. Read more about our automatic conversation locking policy. This action has been performed automatically by a bot. |
Sets up a test harness for the current Material tooltip.