DateTime.ToShortDateString on Linux

[LouisBernath]

I already saw the issue #18345, and i respect that .NET core uses the operating system cababilities and that they may differ from what we see on windows.

Problem:
The DateTime.ToShortDateString produces a different string on a Linux (cflinuxfs2 / Ubuntu 14.04) or a Windows 10.
E.g. for "de-CH"
On Windows "01.01.2016"
On Linux "01.01.16"

It seems to me that ToShortDateString uses the "SHORT" ICU format. It is really unfortunate, that the ICU "de_CH" (en_US as well) locale does define the year only as a two digit. If you have dates from the 19th century, this is really a bummer when formatted with only two digits.

Proposal:
Could you at provide a way to use a standard format specifier to use the medium ICU format? (I know, Windows only has short or long)
Or a way to control if the short or medium ICU format is used. (Maybe the short format is not the right choice for some locales and ToShortDateString.)

Because of this i ended up using a custom format string. So I esentially loose the globalization feature for this particular case.

From http://userguide.icu-project.org/formatparse/datetime

SHORT is numeric, such as 12/13/52 or 3:30pm
MEDIUM is longer, such as Jan. 12, 1952
LONG is longer, such as January 12, 1952 or 3:30:32pm
FULL is completely specified, such as Tuesday, April 12, 1952 AD or 3:30:42pm PST

[tarekgh]

@Burnaster I am still seeing a gap between short and Medium. for example, if you want 12/13/1952, this is not short nor medium.

Would it be better to suggest the change to the ICU library, to update the short formats for such locales? the only concern I have, if we expose a new API, we'll still have a gap and will need to expose more APIs to fill such gaps. so fixing the ICU would make more sense.

[tarekgh]

The other idea could be, when we read the short date format from ICU we always force the 4-digit year instead of a 2-digit year.

@Burnaster does this will be better for you?

[LouisBernath]

@tarekgh 4 year digits would be a lot better from my perspective and solve my problem.

[LouisBernath]

@tarekgh As far as i understood the ICU relies on the CLDR project. I created a ticket stating my concern about the short date pattern; even though i don't see much chance for success in the near future due to the nature of the project and the unicode consortium: https://unicode.org/cldr/trac/ticket/11090

[tarekgh]

@Burnaster thanks. I'll try to look if we can force a 4-digit year in the short date format.

[tarekgh]

This should be fixed by the linked PR.

[effyteva]

Hi,
For anyone suffering from this issue, I've create a simple solution.
https://github.com/effyteva/Teva.Common.Cultures