Re: [lmi] UTF-16 output from automated GUI test?

[Greg Chicares]

On 2016-10-20 19:06, Greg Chicares wrote:
[...]
> 2) Don't revert. Whenever we would do:
>      ./wx_test.exe --data_path=/opt/lmi/data >eraseme 2>&1
>    instead do this:
>      ./wx_test.exe --data_path=/opt/lmi/data 2>&1 |tr --delete '\r' >eraseme

Incorrect. "Don't revert" means UTF-16 output [but see below]
with LF (not CRLF) delimiters, so "tr --delete '\r'" would fix
the wrong problem. Instead, we just want get rid of the NULs,
thus:
  ./wx_test.exe --ash_nazg --data_path=/opt/lmi/data 2>&1 |tr --delete '\000' 
>eraseme
until we can replace std::wcout with std::cout.

It might seem that we really ought to convert to UTF-8, thus:
  ./wx_test.exe --data_path=/opt/lmi/data 2>&1 |iconv -f UTF-16LE -t UTF-8 
>eraseme
That actually works in cygwin, but not in GNU/Linux. The
problem is that the output isn't actually UTF-16: it's a
mixture of UTF-16 and something else resembling ASCII. Maybe
that's just because 'wine' spits out some diagnostics that
aren't UTF-16.