tree d43f95cad1c8b07ecfb9c906a327bdd03e7668b8
parent 53168db99a1a61bcc3b9d183d86cede24f618871
author Greg Spencer <gspencergoog@users.noreply.github.com> 1566594484 -0700
committer GitHub <noreply@github.com> 1566594484 -0700
gpgsig -----BEGIN PGP SIGNATURE-----
 
 wsBcBAABCAAQBQJdYFW0CRBK7hj4Ov3rIwAAdHIIAAu7dJCmEkjkY5lAVpgeVGlN
 l9s2YPQ16YZ3KyFEwt8pQ91SwkhoW0j3nFWL3Ptql2RTNShzAVlDZU+HRXJb+65p
 fGbr4TyEJc0y1a7TKBgZm4N6yrFwsfq0hRWJB7YLtcnZ1JVtflyKO2l3VGTjXq4Y
 yvukgNMu2PvqlSz6kivbL+TWlxvBOjCuMi3iTiMNglz/mLGQBjZ0MBhUYAwmZ6U8
 Pm4FIntisKAkwqOS6JxVywVZsWqCciAUQSqTFpYs88jtXPkRtisEc/bfqWoyB0Oz
 HqRtAxap21c1hkjr4oPsqj8aCSCXLg4TwPyoGiFrx/IV8tCOCs48UPnP/8ZVtu8=
 =Jsbu
 -----END PGP SIGNATURE-----
 

Normalize assert checking of clipBehavior (#38568)

I noticed that we were pretty inconsistent with the way that we checked the value of clipBehavior in the framework, so I normalized the usages and updated docs where necessary.

This is a breaking change if you used to pass null explicitly to FlatButton, OutlineButton or RaisedButton constructors, expecting to get Clip.none. It will now assert if you do that. Existing implementations that pass null implicitly by not specifying clipBehavior won't need to change their call sites. It always implicitly defaulted to Clip.none before, and it will continue to do that, it's only places where it was explicitly set to null in order to get the implicit default that it will fail.