Handle URL ellipsizing using CSS

[finkef]

NOTE: This PR depends on ~#2077~ #2081.

Fixes #1712, #2039, #2040, #2041, #2052, #2053, #2084, #2085.

This PR attempts fixing problems related to ellipsizing and placing of the url icon. It does so by:

• Making sure thought annotations match the actual thought (f1bb087d5e52b872435d5f0bce4e18dac9f7d5bb) Moves margin and padding out of .editable-annotation-text to its container to prevent element flow issues in inline-block layouts and positions URL and email icons absolutely.

• Ellipsizing URLs using CSS (ed047406af5631e41bc341c616168c61d6165bea) Adds a .url class to .editable and .editable-annotation-text if the displayed value is a valid URL forcing to ellipsize when overflowing. Note that this only works when also setting width: 100% on the parent. To not affect other thoughts, we only apply this when a .url child is present. We also use a placeholder value in thought annotations if necessary to prevent the URL icon rendering above the placeholder if the thought is still empty.

• Update URL Icon styles (476fa1b6279c0dea81b45c4071a428b1c18d4d5c) Removes margin and padding around URL icons that seem no-longer necessary.

Added snapshot tests that cover the cases:

• URL should not be ellipsized
• URL should be ellipsized when overflowing
• URL icon should respect placeholders
• URL should not be ellipsized when editing

Since this PR positions icons always immediately after the last character, it makes #2052 no longer applicable.

[finkef]

Can confirm that this PR also resolves #2084 and #2085.

[finkef]

Behavior on mobile when approaching screen edge:

[finkef]

@raineorshine I've rebased on the linewrap-fabian branch, resolved all outstanding comments and went through all affected issues. As we're standing right now, this PR fixes all of https://github.com/cybersemics/em/issues/1712, https://github.com/cybersemics/em/issues/2039, https://github.com/cybersemics/em/issues/2040, https://github.com/cybersemics/em/issues/2041, https://github.com/cybersemics/em/issues/2052, https://github.com/cybersemics/em/issues/2053, https://github.com/cybersemics/em/issues/2084, https://github.com/cybersemics/em/issues/2085.

The also related issue #986 remains unresolved, but will tackle that in a separate PR once we have this merged.