Date: Olson-timezone-support (real 'z', 'v', 'V') by rajavelmani · Pull Request #687 · globalizejs/globalize

rajavelmani · 2017-02-09T23:16:01Z

#340

Supports z, v, V format with timezone names

Added by @rxaviers:

TODO:

Follow up items:

Support the historic metazones based on the given date.

jsf-clabot · 2017-02-09T23:16:02Z

All committers have signed the CLA.

rxaviers

@rajavelmani, thanks for your PR. I left some comments above and updated the description with some TODOs. Thanks so far!!

rxaviers · 2017-02-10T13:30:37Z

 Globalize.dateFormatter =
 Globalize.prototype.dateFormatter = function( options ) {
-	var args, cldr, numberFormatters, pad, pattern, properties, returnFn;
+	var args, cldr, numberFormatters, pad, pattern, properties, returnFn, tzId;


CLDR and IANA refers to them as timezone names. Let's rename tzId into timezoneName?

i can use timeZone as timezoneName is locale dependent

rxaviers · 2017-02-10T13:39:55Z

 	pattern = dateExpandPattern( options, cldr );
 	properties = dateFormatProperties( pattern, cldr );
 	cldr.off( "get", validateRequiredCldr );
+	properties.tzId = tzId;//TODO validate


This variable is being declared and then set here (i.e., always undefined). I believe this was here as a placeholder to accept a potential options.timeZone value. Therefore, I think this can be safely removed for now.

rxaviers · 2017-02-10T13:44:30Z

 cldr = new Cldr( "en" );

-QUnit.assert.dateFormat = function( date, pattern, cldr, expected ) {
+QUnit.assert.dateFormat = function( date, pattern, cldr, expected, tzId ) {


Nitpick comment, let's leave cldr, expected as the last arguments. We could place tzId (actually, timezoneName) after pattern.

Just to confirm, I added the tzId to the end to avoid the changes to rest of the test cases. if your are ok to change all the test then, i can change all the tests.

I have changed the order of the signature as you mentioned.

rxaviers · 2017-02-10T15:51:03Z

+
+	//fall through to 'O' format
+	date = new FakeDate( 0 );
+	assert.dateFormat(date, "z", cldr, "GMT", "Etc/GMT");


Where that is unavailable, falls back to the short localized GMT format ("O").
http://www.unicode.org/reports/tr35/tr35-dates.html#dfst-zone

According to the above, I expect a test where it's used an unavailable metaZone, not "Etc/GMT" that actually exists and is equal to O format as a coincidence.

i used "America/Argentina/Buenos_Aires" which is not supported in CLDR json yet

Nice! Though, I'm wondering... Should we use an invalid stuff that will never be supported like "Foo/Bar"?

rxaviers · 2017-02-10T15:52:15Z

+							);
+						properties.regionFormatDaylight = cldr.main(
+							"dates/timeZoneNames/regionFormat-type-daylight"
+							);


Missing break here, i.e., in the case where it's found.

rxaviers · 2017-02-10T16:16:06Z

+		format = ( standardTzName ) ? standardTzName :
+		( regionFormatStandard ) ? regionFormatStandard.replace( /\{0\}/, exemplarCity ) :
+		undefined;
+	}


It seems to me that at this runtime point we could have done more processing on these properties... I mean, standardTzName (or exemplarCity) could have been incorporated into regionFormatStandard, and daylightTzName (or exemplarCity) could have been incorporated into regionFormatDaylight.

Other points:

The parentheses on ( standardTzName ) ? is unnecessary;

Why the undefined case?

It seems to me that at this runtime point we could have done more processing on these properties... I mean, standardTzName (or exemplarCity) could have been incorporated into regionFormatStandard, and daylightTzName (or exemplarCity) could have been incorporated into regionFormatDaylight.

I have optimized the properties please take a look and let me know your thoughts

rxaviers · 2017-02-10T16:18:34Z

 				"pattern": "h:mm:ss a",
-				"timeSeparator": ":"
+				"timeSeparator": ":",
+				"tzId": undefined


If it's undefined, let's simply not include it.

rxaviers · 2017-02-10T16:19:47Z

+	assert.dateFormat( date, "OO", cldr, "GMT+11" );
+	assert.dateFormat( date, "OOO", cldr, "GMT+11" );
+	assert.dateFormat( date, "OOOO", cldr, "GMT+11:00" );
+});


There's already a "should format timezone (O)" section above. Instead of duplicating it, let's simply incorporate the missing pieces there please.

rxaviers · 2017-02-10T16:22:11Z

+				// z...zzz: "{shortRegion}", eg. "PST" or "PDT".
+				// zzzz: "{regionName} {Standard Time}" or "{regionName} {Daylight Time}",
+				// eg. "Pacific Standard Time" or "Pacific Daylight Time".
+				if ( properties.tzId ) {


Checking for properties.tzId isn't valid, because there's a case where properties.tzId is truthy and metaZone is unavailable and therefore the properties below are invalid (by looking at your code at date/format-properties.js). I think you should check for something else, e.g., one of the set properties.

ok let me change this to check properties.standardTzName and properties.daylightTzName

rxaviers · 2017-02-10T16:23:59Z

+					);
+					if ( ret ) {
+						break;
+					}


Since the properties exist, this should always be truthy.

Yes, this will not be required with check for properties.standardTzName and properties.daylightTzName

rxaviers · 2017-02-13T18:50:32Z

+							length < 4 ? "short" : "long",
+							"daylight"
+						]);
+						if ( !standardTzName && !daylightTzName ) {


What should happen for !standardTzName || !daylightTzName? It won't fall either in here or in the next if. Should we use examplarCity if !standardTzName || !daylightTzName. Probably, this would only happen in a corrupted CLDR data where there is one content, but not the other.

this will help on the scenario like https://github.com/globalizejs/globalize/pull/687/files#diff-df6075946ae91cbb093f63c7597cd746R460 where the timezone id is valid not supported in CLDR.

Note I'm referring to using || instead of && in here, so we also catch the additional case where part of the data exists.

rxaviers

Some more changes please

rxaviers · 2017-02-13T19:04:12Z

+						]);
+						if ( !standardTzName && !daylightTzName ) {
+							var exemplarCity = cldr.main(
+								"dates/timeZoneNames/zone/",


remove trailing /

rxaviers · 2017-02-13T19:08:25Z

+							standardTzName = cldr.main(
+								"dates/timeZoneNames/regionFormat-type-standard"
+							)
+							.replace( /\{0\}/, exemplarCity );


Instead of replacing {0} yourself, you can use src/common/format-message (example).

FWIW, it's an internal utility to help in cases like that. Note, it's not ICU message format :)

i just followed from the format.js. but, i can use the format-message as you suggested. Do i change the format.js code too?

rxaviers · 2017-02-13T19:16:40Z

 			pattern: pattern,
-			timeSeparator: numberSymbol( "timeSeparator", cldr )
+			timeSeparator: numberSymbol( "timeSeparator", cldr ),
+			timeZone: timeZone


Let's skip adding timeZone to the properties. We can add it later if required.

If we don't add timeZone from option.timeZone to properties.timeZone we may not be able to test time zone implementation. Can you clarify why we need to skip timeZone from properties?

rxaviers · 2017-02-13T19:22:40Z

+	assert.dateFormat(date, "zz", "America/Los_Angeles", cldr, "PST");
+	assert.dateFormat(date, "zzz", "America/Los_Angeles", cldr, "PST");
+	assert.dateFormat(date, "zzzz", "America/Los_Angeles", cldr, "Pacific Standard Time");
+


We need to test the "exemplarCity" case.

rxaviers

It's looking great!! I have left some comments for your consideration please.

rxaviers · 2017-03-02T01:06:15Z

+				}
+
+				//fall through "{0} {Time}""
+				//for VVVV format


nitpick: space after //, like // for ...

rxaviers · 2017-03-02T01:06:58Z

+				//fall through "{0} {Time}""
+				//for VVVV format
+				if ( timeZone ) {
+					var TzName;


nitpick: variable names should start with lowercase, i.e., tzName. Preferably using full words, like timezoneName.

rxaviers · 2017-03-02T01:16:47Z

+			]),
+			exemplarCity = cldr.main([
+				"dates/timeZoneNames/zone", timeZone, "exemplarCity"
+			]);


If timeZone is passed to format anything but VVVV this is unnecessary, so let's move this assignment below in the only place where it's used.

rxaviers · 2017-03-02T01:21:40Z

+			metaZone = cldr.supplemental([
+				"metaZones/metazoneInfo/timezone", timeZone, 0,
+				"usesMetazone/_mzone"
+			]),


If timeZone is passed to format anything but z* or v* this is unnecessary, so (considering we move exemplarCity below away) let's update the if condition:

- if ( timeZone ) { + if ( timeZone && ( chr === "v" || chr === "z" ) ) {

ok let me move this condition into https://github.com/globalizejs/globalize/pull/687/files#diff-a8a17cd910e21f0b9d3e8ffe4fcea2a8R50 to access chr

rxaviers · 2017-03-02T02:00:52Z

+			assert.equal( "Pacific Time", properties( pattern, cldr, timeZone ).genericTzName );
+		}
+	});
+});


Let's add a test to assert vv and vvv throw.

In here, or where it's possible to test this...

rxaviers · 2017-03-02T02:01:01Z

+		}
+	});
+});
+


Let's add a test to assert V throws.

rxaviers · 2017-03-02T02:02:11Z

 			case "z":
+				properties.standardTzName = standardTzName;
+				properties.daylightTzName = daylightTzName;
+				break;


nitpick: let's add an empty line after break and before the next case.

rxaviers · 2017-03-02T02:12:40Z

+							),
+							[ unKnownExemplarCity ]
+						);
+					}


Can we simplify the the above as:

if ( length === 2 ) { TzName = timeZone; - } - - if ( exemplarCity ) { - if ( length === 3 ) { - TzName = exemplarCity; - } else if ( length === 4 ) { - TzName = formatMessage( - cldr.main( - "dates/timeZoneNames/regionFormat" - ), - [ exemplarCity ] - ); + } else if ( length === 3 ) { + TzName = exemplarCity; + } else if ( length === 4 ) { + if ( !exemplarCity ) { + exemplarCity = cldr.main([ + "dates/timeZoneNames/zone/Etc/Unknown/exemplarCity" + ]); } - } else if ( !exemplarCity && length === 4 ) { - var unKnownExemplarCity = cldr.main([ - "dates/timeZoneNames/zone/Etc/Unknown/exemplarCity" - ]); TzName = formatMessage( cldr.main( "dates/timeZoneNames/regionFormat" ), - [ unKnownExemplarCity ] + [ exemplarCity ] ); } if ( TzName ) {

https://gist.github.com/d6ca0c74e17c203da094a1167b875b9b

rxaviers · 2017-03-02T02:16:51Z

+	var date = new FakeDate( 0 );
+	assert.dateFormatWithTimezone( date, "VV", "America/Los_Angeles", cldr, "America/Los_Angeles" );
+	assert.dateFormatWithTimezone( date, "VVV", "America/Los_Angeles", cldr, "Los Angeles" );
+	assert.dateFormatWithTimezone( date, "VVVV", "America/Los_Angeles", cldr, "Los Angeles Time" );


Can we include VVV and VVVV tests that use unkown city? It could re-use the above America/Argentina/Buenos_Aires, but it's good to have it explicitly tested here (in addition to the above fallback-test). Thanks

rxaviers · 2017-03-06T18:30:09Z

+ * @json [JSON]
+ *
+ * Load IANA data.
+ */


Pros:

Letting users to load this data similarly to the way they do with CLDR is good. Therefore we don't embed such in the library.

Cons:

Differently from CLDR, IANA doesn't provide with an official JSON binding. The one used here is non-official, it is using the JSON bindings from moment-timezone, and could be subject to changes. Ways to mitigate risks here could be 1) to get moment-timezone data put into a separate iana-json library; 2) check if IANA is interested in making it official? Ideas?

rxaviers · 2017-03-06T18:33:41Z

 	validateDefaultLocale( cldr );

+	timeZone = options.timeZone;
+	validateParameterTypeString( options.timeZone, "options.timeZone" );


Now that we the additional option options.timeZone, we need to update the above options = options || { skeleton: "yMd" }. For example, dateFormatter({timeZone: "something"}) should also set the default skeleton: "yMd". This default doesn't need to be set if either one is used: skeleton, date, or datetime.

rxaviers · 2017-03-06T18:34:41Z

 	"../number/symbol",
-	"../util/string/pad"
+	"../util/string/pad",
+	"../common/format-message"


nitpick: sort these filepaths alphabetically please, therefore common/format-message should be included after the common/.. one above.

rxaviers · 2017-03-06T18:39:00Z

+		"zones": [
+			{
+				"name": "America/Los_Angeles",
+				"abbrs": [


Should we remove this?

rxaviers · 2017-03-06T18:39:27Z

+				420,
+				480
+				],
+				"population": 15058000


Should we remove this?

rxaviers · 2017-03-06T18:40:27Z

+});
+
+QUnit.test( "should format timezone (v)", function( assert ) {
+	try {


It seems like this was used for debug and was left?

rxaviers · 2017-03-08T13:26:55Z

Closed in favor of #701

Ref #340 Ref #453 Ref #687

Fixes #340 Fixes #453 Closes #687 Closes #701

rajavelmani force-pushed the master branch 2 times, most recently from 212e748 to 8e074bf Compare February 10, 2017 01:31

rxaviers requested changes Feb 10, 2017

View reviewed changes

rajavelmani force-pushed the master branch 2 times, most recently from a0f3c96 to be210ba Compare February 11, 2017 01:38

rxaviers reviewed Feb 13, 2017

View reviewed changes

rxaviers requested changes Feb 13, 2017

View reviewed changes

rajavelmani force-pushed the master branch 4 times, most recently from 5b0201e to 028ae5f Compare February 15, 2017 01:24

rxaviers added date module new feature labels Feb 15, 2017

rajavelmani force-pushed the master branch 3 times, most recently from 9599fe1 to b46372b Compare February 22, 2017 00:24

rajavelmani force-pushed the master branch from b46372b to 32e1e9e Compare February 23, 2017 23:33

rajavelmani changed the title ~~Date: Olson-timezone-support (real z)~~ Date: Olson-timezone-support (real 'z', 'v', 'V') Feb 23, 2017

rajavelmani force-pushed the master branch 3 times, most recently from 801152a to 3c8928c Compare March 1, 2017 00:01

rxaviers requested changes Mar 2, 2017

View reviewed changes

rajavelmani force-pushed the master branch 2 times, most recently from ef85e52 to 8fd175f Compare March 2, 2017 21:17

rxaviers reviewed Mar 6, 2017

View reviewed changes

rajavelmani force-pushed the master branch from 85305e9 to 6c7373e Compare March 7, 2017 17:27

Date: Olson-timezone-support (real 'z', 'v', 'V')

ab7566c

rajavelmani force-pushed the master branch from 6c7373e to ab7566c Compare March 7, 2017 22:59

rxaviers mentioned this pull request Mar 8, 2017

Date: Olson-timezone-support (real 'z', 'v', 'V') and options.timeZone #701

Closed

13 tasks

rxaviers closed this Mar 8, 2017

rxaviers pushed a commit that referenced this pull request May 18, 2017

Date: Olson-timezone-support (real z, v, V) and options.timeZone (1/2)

7eda86c

Ref #340 Ref #453 Ref #687

rxaviers added a commit that referenced this pull request May 18, 2017

Date: Olson-timezone-support (real z, v, V) and options.timeZone (2/2)

4fcd2fe

Fixes #340 Fixes #453 Closes #687 Closes #701

Uh oh!

Conversation

rajavelmani commented Feb 9, 2017 • edited by rxaviers Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jsf-clabot commented Feb 9, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rxaviers left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rajavelmani Feb 10, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rxaviers left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rajavelmani Feb 14, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rxaviers left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

rajavelmani commented Feb 9, 2017 •

edited by rxaviers

Loading

jsf-clabot commented Feb 9, 2017 •

edited

Loading

rajavelmani Feb 10, 2017 •

edited

Loading

rajavelmani Feb 14, 2017 •

edited

Loading

rajavelmani Mar 2, 2017 •

edited

Loading