Add form validation message strings....

Add form validation message strings.

* webkit/glue/
  - Add overloads of queryLocalizedString() functions
    They will be added to WebKit::WebKitClient later.

  - Add implementations for WebLocalizedString::Validation*

  - Add message strings
    The English strings are different from WebKit's to avoid the brusque
    WebKit's strings.
    Port some messages from Mozilla.

* webkit/tools/test_shell/ and webkit/support/
  Hook queryLocalizedString() functions in order to return messages same as
  WebKit in layout test.

BUG=28264
TEST=LayoutTests/fast/forms/validationMessage.html

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=67382

[tkent, 2010-11-15 08:07:00]

Hi,

My English messages should be not good.  Would you give me advice on them
please?

I have concern about adding 3 overloads of queryLocalizedString().  Should we
merge them into queryLocalizedString(WebLocalizedString::Name, const
WebVector<WebString>)?

[tony, 2010-11-16 18:11:41]

Brian, who would be a good person to provide English strings for these error
messages?  I provided some guesses, but I'm not sure if it's the tone we want.

Kent, does any other browser have translations for this already?  It would be
nice to try to have similar wording.  I kind of think we should just have really
short and vague error messages to encourage site authors to provide their own
text.

I think the overloaded versions of queryLocalizedString are fine.

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.grd
File webkit/glue/webkit_strings.grd (right):

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:300: <message
name="IDS_FORM_VALIDATION_VALUE_MISSING" desc="Heading or short sentence shown
when a form control in a web page requires a value, but a user don't specify any
value.">
"but the user didn't specify"

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:301: Need to specify a value.
Your text sounds fine, but it's a bit wordy.  Maybe just: "
"Value required."

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:304: The value does not conform to the field
type.
Is it possible for this error message to be specific?  For example: "URL
required" or "Email address required".  Telling the user that the type is wrong
doesn't tell them how to fix.

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:307: Please specify an e-mail address like
foo@example.org.
Invalid e-mail address.

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:310: Please specify comma-separated e-mail
addresses.
Comma separated list of e-mail addresses required.  (Is one email address
allowed?)

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:313: Please specify a URL such as
http://www.example.com/foo .
Does this type of form field require http or https?  It's unfortunate that the
user will have to type that.  Maybe just "Invalid URL."

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:316: The value does not match to a character
restriction for this field.
The value does not match the character restrictions for this field. (This is
going to be very confusing for the user since they won't know how to fix it.)

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:319: The length of the value is <ph
name="CURRENT_LENGTH">$1<ex>101</ex></ph> characters. It shoud be shorter than
<ph name="MAX_CHARACTERS">$2<ex>100</ex></ph> characters.
Text is $1 characters too long.

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:322: Please specify a value equal to or greater
than <ph name="MINIMUM">$1<ex>0</ex></ph>.
Value must be greater than or equal to $1.

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:325: Please specify a value equal to or smaller
than <ph name="MAXIMUM">$1<ex>100</ex></ph>.
Value must be less than or equal to $1.

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:328: The value should be a multiple of <ph
name="STEP_VALUE">$2<ex>4</ex></ph>, based on <ph
name="BASE_VALUE">$1<ex>0</ex></ph>.
This is also really confusing.  Maybe just "Invalid value."

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkitclient_imp...
File webkit/glue/webkitclient_impl.cc (right):

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkitclient_imp...
webkit/glue/webkitclient_impl.cc:329: WebLocalizedString::Name name, int
numeric_value) {
Can this method just convert numeric_value to a string and call
queryLocalizedString(name, WebString&)?

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkitclient_imp...
webkit/glue/webkitclient_impl.cc:339: WebLocalizedString::Name name, int value1,
int value2) {
Can this method just call the WebString, WebString, version?

http://codereview.chromium.org/4678006/diff/7001/webkit/support/test_webkit_c...
File webkit/support/test_webkit_client.cc (right):

http://codereview.chromium.org/4678006/diff/7001/webkit/support/test_webkit_c...
webkit/support/test_webkit_client.cc:255: case
WebKit::WebLocalizedString::ValidationValueMissing:
Google C++ style says case should be indented:
http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Loops_and_Swit...

http://codereview.chromium.org/4678006/diff/7001/webkit/tools/test_shell/test...
File webkit/tools/test_shell/test_shell_webkit_init.cc (right):

http://codereview.chromium.org/4678006/diff/7001/webkit/tools/test_shell/test...
webkit/tools/test_shell/test_shell_webkit_init.cc:126: case
WebKit::WebLocalizedString::ValidationValueMissing:
Ditto.

[tkent, 2010-11-17 00:43:38]

Thank you for reviewing.  I'll update the patch.

On 2010/11/16 18:11:41, tony wrote:
> Brian, who would be a good person to provide English strings for these error
> messages?  I provided some guesses, but I'm not sure if it's the tone we want.
> 
> Kent, does any other browser have translations for this already?  It would be
> nice to try to have similar wording.  I kind of think we should just have
really
> short and vague error messages to encourage site authors to provide their own
> text.

Yes, Opera and Firefox 4.0b7 already have translations for these messages (at
least for English and Japanese).  

The English messages in Opera and Firefox are different. Their messages are
complete sentences unlike the default WebKit messages such as "type mismatch". 
The stepMismatch message in Opera consists of two sentences.

[tony, 2010-11-17 00:47:34]

On 2010/11/17 00:43:38, Kent Tamura wrote:
> Yes, Opera and Firefox 4.0b7 already have translations for these messages (at
> least for English and Japanese).  

I would just copy the Firefox strings (I think that the license allows it).  If
Brian doesn't like them, we can change them later.

[tkent, 2010-11-17 01:07:27]

On 2010/11/17 00:47:34, tony wrote:
> I would just copy the Firefox strings (I think that the license allows it). 
If
> Brian doesn't like them, we can change them later.

Firefox's strings are here:
http://hg.mozilla.org/mozilla-central/file/fe92d201e2c4/dom/locales/en-US/chr...
Unfortunately Firefox doesn't have messages for rangeUnderflow, rangeOverflow,
and stepMismatch yet.

[tkent, 2010-11-17 02:21:20]

What should we do for the license header of webkit_strings.grd if we copy the
Firefox strings?

[tony, 2010-11-17 16:27:17]

On 2010/11/17 02:21:20, Kent Tamura wrote:
> What should we do for the license header of webkit_strings.grd if we copy the
> Firefox strings?

We can put the license in an XML comment in the grd file.  You can list the
string IDs that it applies to.  Here's a similar example:
http://src.chromium.org/viewvc/chrome/trunk/src/webkit/glue/resources/README....

[tkent, 2010-11-22 10:19:05]

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.grd
File webkit/glue/webkit_strings.grd (right):

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:300: <message
name="IDS_FORM_VALIDATION_VALUE_MISSING" desc="Heading or short sentence shown
when a form control in a web page requires a value, but a user don't specify any
value.">
On 2010/11/16 18:11:41, tony wrote:
> "but the user didn't specify"

Done.

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:304: The value does not conform to the field
type.
On 2010/11/16 18:11:41, tony wrote:
> Is it possible for this error message to be specific?  For example: "URL
> required" or "Email address required".  Telling the user that the type is
wrong
> doesn't tell them how to fix.

This message is for types which have no specific localized messages.  So we
can't make it type-specific.
I changed this to just "Invalid value" like your STEP_MISMATCH message.

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:310: Please specify comma-separated e-mail
addresses.
On 2010/11/16 18:11:41, tony wrote:
> Comma separated list of e-mail addresses required.  (Is one email address
> allowed?)

One email address is allowed.  I updated this message for your comment and
Firefox's message.

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:313: Please specify a URL such as
http://www.example.com/foo .
On 2010/11/16 18:11:41, tony wrote:
> Does this type of form field require http or https?  It's unfortunate that the
> user will have to type that.  Maybe just "Invalid URL."

The value needs to start with the URL scheme like "http://" for now.
https://bugs.webkit.org/show_bug.cgi?id=29235

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:322: Please specify a value equal to or greater
than <ph name="MINIMUM">$1<ex>0</ex></ph>.
On 2010/11/16 18:11:41, tony wrote:
> Value must be greater than or equal to $1.

Done.

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:325: Please specify a value equal to or smaller
than <ph name="MAXIMUM">$1<ex>100</ex></ph>.
On 2010/11/16 18:11:41, tony wrote:
> Value must be less than or equal to $1.

Done.

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkit_strings.g...
webkit/glue/webkit_strings.grd:328: The value should be a multiple of <ph
name="STEP_VALUE">$2<ex>4</ex></ph>, based on <ph
name="BASE_VALUE">$1<ex>0</ex></ph>.
On 2010/11/16 18:11:41, tony wrote:
> This is also really confusing.  Maybe just "Invalid value."

Done.

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkitclient_imp...
File webkit/glue/webkitclient_impl.cc (right):

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkitclient_imp...
webkit/glue/webkitclient_impl.cc:329: WebLocalizedString::Name name, int
numeric_value) {
On 2010/11/16 18:11:41, tony wrote:
> Can this method just convert numeric_value to a string and call
> queryLocalizedString(name, WebString&)?

It's a good idea.
We should have
 - queryLocalizedString(Name)
 - queryLocalizedString(Name, const WebString&)
 - queryLocalizedString(Name, const WebString&, const WebString&)
and should remove queryLocalizedString(Name, int).  Integer->String conversion
should be done in WebKit APi layer.

http://codereview.chromium.org/4678006/diff/7001/webkit/glue/webkitclient_imp...
webkit/glue/webkitclient_impl.cc:339: WebLocalizedString::Name name, int value1,
int value2) {
On 2010/11/16 18:11:41, tony wrote:
> Can this method just call the WebString, WebString, version?

Removed this function.

http://codereview.chromium.org/4678006/diff/7001/webkit/support/test_webkit_c...
File webkit/support/test_webkit_client.cc (right):

http://codereview.chromium.org/4678006/diff/7001/webkit/support/test_webkit_c...
webkit/support/test_webkit_client.cc:255: case
WebKit::WebLocalizedString::ValidationValueMissing:
On 2010/11/16 18:11:41, tony wrote:
> Google C++ style says case should be indented:
>
http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml#Loops_and_Swit...

Done.

http://codereview.chromium.org/4678006/diff/7001/webkit/tools/test_shell/test...
File webkit/tools/test_shell/test_shell_webkit_init.cc (right):

http://codereview.chromium.org/4678006/diff/7001/webkit/tools/test_shell/test...
webkit/tools/test_shell/test_shell_webkit_init.cc:126: case
WebKit::WebLocalizedString::ValidationValueMissing:
On 2010/11/16 18:11:41, tony wrote:
> Ditto.

Done.

[tony, 2010-11-22 17:47:34]

LGTM!

[tkent, 2010-11-23 14:48:27]

Darin,
I'd like to change WebKitClient::queryLocalizedString().  This Chromium change
is a preparation for it.

1. Add the following two overloads to WebKitClient:
 - queryLocalizedString(Name, const WebString&)
 - queryLocalizedString(Name, const WebString&, const WebString&)

2. Change multipleFileUploadText(unsigned) in LocalizedStrings.cpp so that it
calls queryLocalizedString(Name, const WebString&) instead of
queryLocalizedString(Name, int).

3. Remove queryLocalizedString(Name, int)

Is this ok?

[darin (slow to review), 2010-11-23 16:53:41]

On Tue, Nov 23, 2010 at 6:48 AM, <tkent@chromium.org> wrote:

> Darin,
> I'd like to change WebKitClient::queryLocalizedString().  This Chromium
> change
> is a preparation for it.
>
> 1. Add the following two overloads to WebKitClient:
>
>  - queryLocalizedString(Name, const WebString&)
>  - queryLocalizedString(Name, const WebString&, const WebString&)
>

OK.  I presume that the WebString parameters are intended as substitution
parameters
to the localized string, right?

How about giving them more self-documenting names, or add a comment
explaining
what they are.

-Darin



>
> 2. Change multipleFileUploadText(unsigned) in LocalizedStrings.cpp so that
> it
> calls queryLocalizedString(Name, const WebString&) instead of
> queryLocalizedString(Name, int).
>
> 3. Remove queryLocalizedString(Name, int)
>
> Is this ok?
>
>
>
> http://codereview.chromium.org/4678006/
>

[tkent, 2010-11-25 04:50:28]

On 2010/11/23 16:53:41, darin wrote:
> OK.  I presume that the WebString parameters are intended as substitution
> parameters
> to the localized string, right?

Right.

> How about giving them more self-documenting names, or add a comment
> explaining
> what they are.

I see.  I'll add a comment.