Fix placeholders for password inputs

[Manishearth]

currently they show dots instead of the placeholder text

[eefriedman]

I think this messes up get_insertion_point_index_for_layout; it expects to see the value before the mapping to dots.

Can we reftest this? Maybe even just <input type=password placeholder=xxx> == <input type=text placeholder=xxx>.

[Manishearth]

The comment in get_insertion_point_index_for_layout mentions that we just need a 1:1 char mapping, which we have.

Yes, I will add a reftest.

[Manishearth]

Added a reftest

[Manishearth]

@bors-servo try

[bors-servo]

⌛ Trying commit 728ec11 with merge ec60957...

[bors-servo]

💔 Test failed - mac-rel-wpt

[Manishearth]

(failed due to an intermittent)

[eefriedman]

get_insertion_point_index_for_layout calls char_indices on the result of get_raw_textinput_value. See #8265.

[Manishearth]

Given that layout only gets the dots, shouldn't it be calculating the insertion point based on the dots and not the actual text? In case the text contains non-ascii values this could get messed up.

[Manishearth]

Pushed another commit which takes a different approach and moves the placeholder logic out of get_raw_textinput_value (it doesn't really belong there anyway?). It shouldn't affect get_raw_textinput_value since the placeholder is only there on an empty string, so the insertion point has to be zero.

[eefriedman]

Did you mean to make some other changes to get_raw_textinput_value, like removing the check for an empty string, or removing the mapping to dots?

[Manishearth]

My bad, updated.

[eefriedman]

Missing changes to get_raw_textinput_value?

[Manishearth]

Done

[eefriedman]

@bors-servo r+

[bors-servo]

📌 Commit cc5844a has been approved by eefriedman

[bors-servo]

⌛ Testing commit cc5844a with merge 7f156b8...

[bors-servo]

☀️ Test successful - android, gonk, linux-dev, linux-rel, mac-dev-ref-unit, mac-rel-css, mac-rel-wpt