JSON is always UTF-8 and there is absolutely no need to escape all non-ASCII
characters in output strings.
Signed-off-by: Simon Rozman <simon@rozman.si>
wchar_t is not char16_t on MSVC, requiring a lot of typecasting when
interfacing ZRCola database strings with GUI.
Signed-off-by: Simon Rozman <simon@rozman.si>
ZRCola is using UTF-16LE strings internally (thanks to Windows).
However, wchar_t and std::wstring are UTF-32 on other platforms.
Signed-off-by: Simon Rozman <simon@rozman.si>
gcc precompiles templates. When << and >> operators of our datatypes are
not overloaded at the time <stdex/idrec.h> is #included yet, gcc will
seek/look for currently available << and >> operators when reaching
std::ostream and std::istream templates.
Signed-off-by: Simon Rozman <simon@rozman.si>
This allows ZRCola to Unicode translation to use precomposed Unicode
characters and provide better coverage for ZRCola to Unicode
translation.
Signed-off-by: Simon Rozman <simon@rozman.si>
Thou linker can locate the output .lib file of referenced projects
wherever .lib is just fine, this helps us to gather all .pdb files in
the output folder.
Signed-off-by: Simon Rozman <simon@rozman.si>
Windows.h must be included first and must be included with care
regarding WinSock.h and WinSock2.h affair.
Signed-off-by: Simon Rozman <simon@rozman.si>
The initial support compares characters in ZRCola font Private-Use-Area
to the characters obtained using Unicode combining characters. Those
which match sufficiently are arranged into a new transliteration.
As the ZRCola Composed to Unicode transliteration requires to be applied
_after_ ZRCola composition but UI applies additional transliterations
_before_ ZRCola composition, the ZRCola composition was reintroduced as
one of the transliterations. This allows configuring a custom
transliteration sequence.
Signed-off-by: Simon Rozman <simon@rozman.si>
