Change behavior of --show-autofill-type-predictions

Change behavior of --show-autofill-type-predictions

Don't show prediction in placeholder text, because changing the placeholder text can alter the field's parsing.

BUG=445327
Committed: https://crrev.com/87f018f91276b9fa008699c6ed27640738672f98
Cr-Commit-Position: refs/heads/master@{#310113}

[Evan Stade, 2014-12-29 23:34:56]

estade@chromium.org changed reviewers:
+ isherman@chromium.org, thestig@chromium.org

[Lei Zhang, 2015-01-05 23:02:24]

lgtm

[Ilya Sherman, 2015-01-06 00:18:57]

I think the placeholder text is the most valuable part of the flag.  Why does it
mess up our parsing?  Isn't it just reiterating a previous parse of the form?

[Evan Stade, 2015-01-06 00:33:57]

On 2015/01/06 00:18:57, Ilya Sherman wrote:
> Isn't it just reiterating a previous parse of the form?

I don't know what this means. But when the form is parsed for the second time,
after the placeholder text has been set, the outcome is different because things
like ADDRESS_HOME_ZIP make it think it's a street address line.

[Ilya Sherman, 2015-01-06 00:42:36]

On 2015/01/06 00:33:57, Evan Stade wrote:
> On 2015/01/06 00:18:57, Ilya Sherman wrote:
> > Isn't it just reiterating a previous parse of the form?
> 
> I don't know what this means. But when the form is parsed for the second time,
> after the placeholder text has been set, the outcome is different because
things
> like ADDRESS_HOME_ZIP make it think it's a street address line.

Why does this make us think that the line is a street address line?  Shouldn't
we give more weight to "zip" in the text?  It doesn't seem unreasonable for a
site to have placeholder text like "Home address: Zip code", which seems
essentially equivalent to the placeholders that we generate.  Would that cause
the same issues?

[Evan Stade, 2015-01-06 00:51:03]

On 2015/01/06 00:42:36, Ilya Sherman wrote:
> Shouldn't we give more weight to "zip" in the text?

This is a philosophical question I don't want to delve into because there's no
site that actually uses "home address: zip" as placeholder or label text.
Besides, this is only one example. The bottom line is that we shouldn't be
tinkering with parts of the form that we use to parse types.

[Ilya Sherman, 2015-01-06 02:13:57]

On 2015/01/06 00:51:03, Evan Stade wrote:
> On 2015/01/06 00:42:36, Ilya Sherman wrote:
> > Shouldn't we give more weight to "zip" in the text?
> 
> This is a philosophical question I don't want to delve into because there's no
> site that actually uses "home address: zip" as placeholder or label text.
> Besides, this is only one example. The bottom line is that we shouldn't be
> tinkering with parts of the form that we use to parse types.

Eh, I do see where you're coming from, and I don't feel strongly about this. 
LGTM.

[Evan Stade, 2015-01-06 17:42:25]

On 2015/01/06 02:13:57, Ilya Sherman wrote:
> On 2015/01/06 00:51:03, Evan Stade wrote:
> > On 2015/01/06 00:42:36, Ilya Sherman wrote:
> > > Shouldn't we give more weight to "zip" in the text?
> > 
> > This is a philosophical question I don't want to delve into because there's
no
> > site that actually uses "home address: zip" as placeholder or label text.
> > Besides, this is only one example. The bottom line is that we shouldn't be
> > tinkering with parts of the form that we use to parse types.
> 
> Eh, I do see where you're coming from, and I don't feel strongly about this. 
> LGTM.

thanks

[Evan Stade, 2015-01-06 17:42:29]

The CQ bit was checked by estade@chromium.org

[commit-bot: I haz the power, 2015-01-06 17:43:09]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/833473003/1

[commit-bot: I haz the power, 2015-01-06 18:46:31]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-01-06 18:48:00]

Patchset 1 (id:??) landed as
https://crrev.com/87f018f91276b9fa008699c6ed27640738672f98
Cr-Commit-Position: refs/heads/master@{#310113}