fix(panel): loosen up bounds assessment

[oliversalzburg]

getBoundingClientRect() will return a width of 0 in certain scenarios.
Retrieving the width through clientWidth will hold the desired width for the position calculation.

As this can be reversed for other situations, we take the max of both values.

Closes #10543

[oliversalzburg]

The mdChips failure is already in master AFAIK. I'd be happy to wait and rebase later if needed.

[EladBezalel]

@ErinCoughlan care to take a look?

[ErinCoughlan]

LGTM

@DerekLouie I think this is the cause of the sliding animation bug as well.

[fsmeier]

Any updates here @ThomasBurleson ?

[ThomasBurleson]

@oliversalzburg - lgtm

[oliversalzburg]

@ThomasBurleson Awesome :) What does the needs: presubmit tag refer to? Anything I need to take care of?

[ThomasBurleson]

@oliversalzburg - No, that is a flag used to indicate readiness for strict internal testing; the work is on our end now.

[fsmeier]

@ThomasBurleson what is the timeline for the next version and the pull/merge requests? This request is there for 3 months already and not merged yet and no new release.
You even have persons who contribute like @oliversalzburg but yet his afford is just meaningless if it is not going to be released.

At our project we are still on 1.1.1 which also has exceptions for some components but updating is no choice since the last "bugfix" releases introduced new features and new bugs.

So pls.: focus on bugfixes, close every new feature and bring this project to a stable version for the "end".

Best,
Florian

[fsmeier]

@ThomasBurleson why is this now marked with 1.1.6 instead of 1.1.5? Without this fix tooltips are not usable in the described cases.

[fsmeier]

@ErinCoughlan @EladBezalel can you push it somehow forward to version 1.1.5? what is missing here to have it merged? with debugInfoEnabled(false) the tooltips are not usable so an update is not possible for us.

[EladBezalel]

@IMM0rtalis as i said in - #9379 (comment)

[fsmeier]

@Splaktar Heyho, I wish you a happy new year!
So how is it continuing with this PR? What are the google internal failed tests? How can we finally merge this? What can we do to help?

Again another month passed without any change here. It is really frustrating. This PR is RDY since may!!! If every bugfix needs a year of merging-time we will be doomed! :(

[oliversalzburg]

I'm assuming this means that Google is using angular-material is some of their own products and, with this change, there are issues in their product.

So there is really nothing to be done by outsiders, except waiting or forking. Let me know if you're forking! 😄

[Splaktar]

Sorry for the delay. I am working to get more details on what kind of failures have been encountered. Obviously I can't just get access to the test results as there may be confidential information in them.

As @oliversalzburg mentioned, there is nothing that the community can do at this time. Google is very hesitant to make breaking changes to this library. It is in production in many internal and external applications around the world. These applications may depend upon the current behavior (even if erroneous).

For now, my focus is on getting all of the PRs merged that don't cause any breakages. Once those are all handled, I will push hard to get PRs merged that fix bugs and seem to be caught by tests that are too specific or false positives.

[Splaktar]

@IMM0rtalis it looks like we've managed to get this through the presubmit phase. Hoping to be able to merge very soon.