Log Message:
tex-polyglossia{,.doc}: update to 1.63

Bug fixes:
‣ Language options are now lowercased before they are handled.
  This fixes errors when using the options in \MakeUppercase
  context (e.g., in headings).
‣ Fix casing value for medieval Latin (this concerns ⟨u⟩ and
  ⟨V⟩ casing).
‣ Remove hacks in some gloss files where \MakeUppercase has been
  redefined to a no-op since the respective scripts do not have
  uppercasing. As \MakeUppercase is now locale-aware, these hacks
  are no longer needed. This concerns Arabic, Divehi, Hebrew,
  Hindi, Kurdish, Persian, Punjabi, Syriac, and Urdu.
‣ Fix Eastern (Mashriq) Arabic numbering direction with LuaLaTeX
  for Arabic, Kurdish, Persian, Syriac, and Urdu.
‣ French variant=acadian refers to the Canadian French subvariety
  as spoken in the Acadian region (as opposed to Quebec French).
  To differentiate it we now use a dedicated subtag which can also
  be used to access the language (fr-CA-u-sd-canb).
  Note that Quebec French is currently connected with fr-CA,
  not the more specific fr-CA-u-sd-caqc, although the u extension
  is returned if requested via \BCPdata. In practice, the Canadian
  subvarieties of French are identical currently, anyway.
‣ Do not blindly de-activate the quote character in languages that
  provide shorthands if the shorthands are not used. This allows
  to activate the character with other packages such as csquotes.
‣ Do not flood the LaTeX hooks with redundant French itemlabels
  and footnote settings.
‣ Fix Korean patching of part format with hyperref.
‣ Make Hebrew numerals expandable.
‣ Avoid superfluous writing of language switches to aux file.

New features:
‣ New option capitalyiwn for Armenian.
‣ New option capitaleszett for German.
‣ New option capitaliota for Greek.
‣ New option datei for Serbian.
‣ New option sectionsep for Arabic, Kurdish, Persian, Syriac,
  and Urdu.
‣ Add aliases eastern (= mashriq) and western (= maghrib) to
  Arabic numerals.
‣ New date commands \today*, \todayGen, \todayArabic, and
  \todayRoman (all with starred counterparts) for Serbian.
‣ Implement basic support for variant=ijekavian in Serbian.
‣ Dutch now hyphenates properly on letters with tremata.
  The new option, tremahyphenation, is provided to opt-out
  this feature.
‣ Add shorthands "a, "e, "i, "o, "u, "A, \ 
"E, "I, "O, "U,
  "y, and "Y to Dutch.

Interface changes:
‣ All language options are now lowercase by default (though the
  mechanism is case-insensitive anyway). Changes cover Kurdish,
  Sanskrit, Serbian.