If a trailing character is a thousands separator, VarParseNumFromStr()
only processes it if NUMPRS_THOUSANDS was specified; even if that
character is also a space and NUMPRS_TRAILING_WHITE was specified.
Signed-off-by: Francois Gouget <fgouget@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
As the test_VarParseNumFromStrMisc() tests show the converse is not
true.
Note also that the " 1 000" test was failing as expected because in
the French locale Wine was not recognizing regular spaces as thousands
separators. Now it does and Wine incorrectly allows them before the
first digit, just like it does for the English tests.
Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=51450
Signed-off-by: Francois Gouget <fgouget@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Windows <= 8.1 uses an obsolete currency symbol for the Swiss Franc and
Windows 8.1 incorrectly doubled the right-to-left mark in the Moroccan
dirham.
Signed-off-by: Francois Gouget <fgouget@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
The error codes are already defined with an HRESULT cast so there is
no need to cast them again.
Signed-off-by: Francois Gouget <fgouget@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Old Windows versions used 29 as the Y2K cutoff, that is is they mapped
two digit years 00-29 to 2000-2029 and years 30-99 to 1930-1999.
But starting with Windows 10 1903 the cutoff is now 49 by default.
So update Wine to match the current Windows versions and adjust the
tests to work for both.
Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=51453
Signed-off-by: Francois Gouget <fgouget@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Old Windows versions used 29 as the Y2K cutoff, that is they mapped
two digit years 00-29 to 2000-2029 and years 30-99 to 1930-1999.
But starting with Windows 10 1903 the cutoff is now 49 by default.
So update Wine to match the current Windows versions and adjust the
tests to work for both.
Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=51453
Signed-off-by: Francois Gouget <fgouget@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
This simplifies the code and fixes support for negative values.
Also trim unnecessary spaces from the EXPECT_DBL() failure messages.
Signed-off-by: Francois Gouget <fgouget@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Check how it handles currency signs and spaces before and after the
digits.
Check the combination of thousands separators, spaces and commas. Check
handling of regular vs. non-breaking spaces since the latter are often
used in some locales.
And check what happens in the French locale where spaces and thousands
separators are identical and the currency sign comes after the digits
and, quite often, a space.
Check handling of multi-character currency symbols.
Signed-off-by: Francois Gouget <fgouget@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
printf("%c Hello\n", 0) works but not TRACE("%c Hello\n", 0) because the
string is formatted before printing.
Work around the issue with wine_dbgstr_w().
Signed-off-by: Francois Gouget <fgouget@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
VariantCopy clears existing contents of pvargDest and thus requires
it contain a valid (possibly-empty) VARIANT, not uninitialized garbage.
If a failure still occurs, propgate the HRESULT to GetAll*CustData.
Signed-off-by: Kevin Puetz <PuetzKevinA@JohnDeere.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
GetFuncDesc removes parameters which are handled specially by Invoke;
GetNames should also consistently omit their names.
Signed-off-by: Kevin Puetz <PuetzKevinA@JohnDeere.com>
Signed-off-by: Huw Davies <huw@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
[lcid] parameters are supplied from IDispatch::Invoke's parameters,
rather than via DISPPARAMS::rgvargs[] and should also be removed
from the FUNC_DISPATCH translation.
Signed-off-by: Kevin Puetz <PuetzKevinA@JohnDeere.com>
Signed-off-by: Huw Davies <huw@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
A [retval] parameter overwrites any return type, not just HRESULT.
But HRESULT is special even without a [retval] to replace it;
it's translated into pExcepInfo and so the return type becomes void.
This rewriting should occur only for functions not originally
defined as FUNC_DISPACH (e.g. inherited or [dual]).
A FUNCDESC originally declared as a dispinterface is left as-is,
and might e.g. return HRESULT.
Signed-off-by: Kevin Puetz <PuetzKevinA@JohnDeere.com>
Signed-off-by: Huw Davies <huw@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
A dispinterface has no vtbl beyond IDispatch itself.
Signed-off-by: Kevin Puetz <PuetzKevinA@JohnDeere.com>
Signed-off-by: Huw Davies <huw@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
A NULL expected value is always a test failure, but printing
the unexpected "wrong" value is more helpful than crashing.
Signed-off-by: Kevin Puetz <PuetzKevinA@JohnDeere.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Always print initializer for funcs[] array, even when empty
Signed-off-by: Kevin Puetz <PuetzKevinA@JohnDeere.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Check if szFullPath ended with '\\' (was explicitly a directory).
Signed-off-by: Kevin Puetz <PuetzKevinA@JohnDeere.com>
Signed-off-by: Jacek Caban <jacek@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Francois Gouget