unexpected lowercase(Char) result

[catawbasam]

Expected to see 'δ'--

julia> lowercase('Δ')
'Δ'

              _
  _       _ _(_)_     |  A fresh approach to technical computing
 (_)     | (_) (_)    |  Documentation: http://docs.julialang.org
  _ _   _| |_  __ _   |  Type "help()" for help.
 | | | | | | |/ _` |  |
 | | |_| | | | (_| |  |  Version 0.3.0-rc1+284 (2014-07-30 21:54 UTC)
_/ |\__'_|_|_|\__'_|  |  Commit da17f92 (5 days old master)
__/                   |  x86_64-w64-mingw32

[catawbasam]

Looks like it is Windows only:

   _       _ _(_)_     |  A fresh approach to technical computing
  (_)     | (_) (_)    |  Documentation: http://docs.julialang.org
   _ _   _| |_  __ _   |  Type "help()" for help.
  | | | | | | |/ _` |  |
  | | |_| | | | (_| |  |  Version 0.3.0-rc1+258 (2014-07-29 15:42 UTC)
 _/ |\__'_|_|_|\__'_|  |  Commit 15f5aed* (6 days old master)
|__/                   |  x86_64-redhat-linux

julia> lowercase('Δ')
'δ'

[tkelman]

msvcrt.dll's towlower doesn't speak Greek, apparently. Neither does msvcr120.dll.

According to http://msdn.microsoft.com/en-us/library/8h19t214.aspx, it only works for ascii.
I wonder whether utf8proc could give us upper/lowercase mappings?

[tkelman]

I do get the right answer if I use _towlower_l though (I think because we set the C locale somewhere - http://stackoverflow.com/questions/15889784/why-does-the-towlower-function-not-convert-the-%D0%AF-to-a-lower-case-%D1%8F):

julia> convert(Char, ccall(:_towlower_l, Cwchar_t, (Cwchar_t,), 'Δ'))
'δ'

I don't think that exists in Windows XP.

[jiahao]

I wonder whether utf8proc could give us upper/lowercase mappings?

utf8proc does not expose Unicode letter casings data except for a casefolding function (for case-insensitive string comparisons).

Case mappings cannot be defined uniquely without a locale. The canonical example is small 'i', whose upper-cased image is capital 'I' in English but capital 'İ' (with overdot) in Turkish.

Since utf8proc is locale-agnostic, it wouldn't be possible to implement this without also implementing all the ancillary code for locale handling. This problem is better suited for a library like libicu,.

[jiahao]

In fact, this manifestly returns the wrong result for terminal sigma in Greek:

julia> lowercase("ΛΌΓΟΣ") #logos; λόγος
"λόγοσ"

This is a separate issue, though.

[JeffBezanson]

The setlocale calls we make are

        setlocale(LC_ALL, ""); // set to user locale
        setlocale(LC_NUMERIC, "C"); // use locale-independent numeric formats

[vtjnash]

I think I succeeded in removing those calls for all of 2 days, then someone noticed that it broke printing of bignums.

[catawbasam]

closed by PR #8233