dojox.date.*.locale prints console noise when empty string is passed

[Rob Retchless]

dojox.date.*.locale should behave just like dojo.date.locale with the same input. This is not the case with falsy input.

At the moment, here's the behavior:

dojo.date.locale.parse() returns null. dojo.date.locale.parse("") returns null. dojox.date.buddhist.locale.parse() throws an exception. dojox.date.buddhist.locale.parse("") returns null, but spews console noise about invalid format dojox.date.islamic.locale.parse() throws an exception. dojox.date.islamic.locale.parse("") returns null, but spews console noise about invalid format dojox.date.hebrew.locale.parse() throws an exception. dojox.date.hebrew.locale.parse("") returns null, but spews console noise about invalid format

These dojox locale classes should not print to the console for falsy input.

[Adam Peller]

I would prefer that we patch the dojox.date code so that it matches dojo.date which has no console.debug (also in better compliance with the style guide)

[Adam Peller]

patch from retchless (IBM, CCLA)

[Adam Peller]

In [28106]:

Remove debug message. Fixes #14969

[haysmark]

In [28135]:

Backport [28106] to 1.6 branch. Refs #14969.

[Adam Peller]

@haysmark, this probably doesn't hit the threshold to backport in the public repo, but if we do backport to 1.6, we should probably backport to 1.7, and change the milestone accordingly. My vote is not to backport.

[haysmark]

Right, I will backport to 1.7.

You will backport major changes to the ways dates are interpreted, but won't remove a few debug statements? I think it is a pretty harmless change.

[haysmark]

In [28136]:

Backport [28106] to 1.7 branch. Refs #14969.

[Adam Peller]

@haysmark, if you're referring to #14158 that caused a regression where the calendar would fail to open. I think that's a bit different, don't you?