On 04/05/2021 17:44, Eike Rathke wrote:
...
If we would really introduce DateOrder::YDM the extra check for
rCode.indexOf('g') as an lv-LV speciality wouldn't be necessary.
(btw, it should be "nDay <= nMonth" instead of "nDay << nMonth", and the
condition be made another else if like the others).
My goal was indeed to be the most specific possible to not guess
dateformat at random in other cases than Latvian.
Perhaps these other cases should be considered as wrong.
Ugh.. there are quite some places that assume only the three values
exist. Anything using those would have to be inspected what happens with
the new value, and possibly adapting. ( git grep DateOrder:: )
I haven't encountered YDM in short or medium date formats, so maybe that
should be reserved for long date formats. Which then would be an
additional LongDateOrder enum used only with
LocaleDataWrapper::getLongDateOrder(), narrowing down the having to
be adjusted cases quite by some and avoiding unnecessary code and
complications.
Ok it's more complicate to deal with than I thought.
No need to spend some time and make the code more complicated because of
this slight warning then.
Thank you for your feedback! :-)
_______________________________________________
LibreOffice mailing list
LibreOffice@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/libreoffice
