Change-Id: I2e5462544d83572e8206bb1db92a2b9df032281b

Change-Id: I27985e456447ef3d71cae39c2acaf59b3aa60b83

Change-Id: I213722cc98490430378014290cb09cca9e469bbb

Instead of a bare NAN.

There are still many places in this function where the code bluntly returns a
bare NAN. That is always the wrong thing to do. However, it is not certain
that the #DIV/0! is the right error in all cases. One would need to check in
each case how to get there, and what the reference C++ implementation and
other spreadsheet products do in such a case.

Change-Id: I5454609082a92144b1afc3c1745586bfe8ef87f8

Returning a NaN with no "double error" semantic payload as such does not make
sense. "Pure" NaNs are displayed in Calc as "nan" which probably is not what
we want.

Change-Id: I85dc55328669bd27478be438dd5a9970d24cf7f9

Change-Id: Ia3e978e3fd9c0ce19ec679c41d40a34338147c37
Reviewed-on: https://gerrit.libreoffice.org/14766Reviewed-by: Lionel Elie Mamane <lionel@mamane.lu>
Tested-by: Lionel Elie Mamane <lionel@mamane.lu>

Returning -DBL_MAX doesn't make sense. The traditional C++ implementation and
other spreadsheet products return an error.

Change-Id: I71117c0b1f113d31ad6eb0c1798b74f96b3a75de

Change-Id: Ia58f19c7121dc8a835436bcefc7fd1e3789407db

Returning DBL_MAX doesn't make sense. The traditional C++ implementation and
other spreadsheet products return an error.

Change-Id: I8189c3ecf4aa7b1df00ec2fd8b91030085848bf4

Surely it doesn't make sense, but we should return zero instead?

At the two call sites, there are tests against a zero having been returned,
but not against a negative value. And the return value is even passed as the
nArrayLength parameter to a formula::DoubleVectorRefToken constructor, which
is of type size_t, thus unsigned. Passing for instance -4 to it ends up being
interpreted as 18446744073709551612, which has fun consequences.

Change-Id: Ifca97cd4bae8e921345bcf9d0ca6bb040e3c3b79

Change-Id: I252987d6c5e5da56a83742c96b4e86abbb7108c6

Having these statistical functions perform well is essential in many cases,
and their implementations seem to be correct.

Change-Id: I30afa096295cc163f8c0146e916a77aa411dd07b

Change-Id: Ib71d74d4b66b02a6a680b5ab0c905b987d59cf7f

Change-Id: I43b59b4359a010e7b64df5825ae3ce845e47fc4b

The initialism “TDF” is rather obscure still.

Change-Id: I987e96d869a5126f5d137d188a34b4ae7d4888f8

Change-Id: I11c7c94a52e4be7286a919e61818d08d5b9ad623

Change-Id: I3aacb177d209db74ffadf5284550cdd4972554de

Change-Id: Ie063071cc57d756fd5da20a15533a6857940f676

Change-Id: I9d7d1e16c34eac75652a1275b6e4ac3f7973cb56

Change-Id: I33a740af4ccbcaa9ed4e3ffd79c4f8cda9b37dc4

Change-Id: I7569a20877411a5ceb4a7475d82cb634eb6cb114

In general it should work like this..

Change-Id: Ic45ca6138c1b252ce9d216ce61c7129c055342ef

... similar to ocArrayOpen,ocArrayClose

Change-Id: I10b4cd2f74167cc57ef7a2ed96e0e98f0cd86805

Change-Id: Icfbf088ee8d91d90779dc804b7dff299df1170a4

Change-Id: Ic82ccfcba08ce276eac08609de215ba8edf8e813

Change-Id: Ia9d54fd582f8e195edaa05937ee0f0290e56875f

Change-Id: I62cce7321dfe038d1723c73ba9deb0450a03f0e5

Change-Id: I63b41ea4619098390824f335859d62b4fa8fcc66

Change-Id: I48c716d87ce3e46f61393ae140bf2e05d9fd6241

in order to get the font colors for menus right.

includes revert "set parent styles to fall back to"
This reverts commit 710d0344.

Change-Id: Ica73a30a44a5dd61a8310f2bee74068c406f470a

Change-Id: I187c4d1bef306ec05985cd6e8072d044e12a993f

... makes it hard to clean up these problems.

But the warnings on almost every std::copy and std::transform are
obnoxious, so disable them with _SCL_SECURE_NO_WARNINGS

Change-Id: Ieea611175fb16b78ff972d194767d3d192992de5

Change-Id: Id70289b0c88dd203cac6301303fe7452088f60f9

Change-Id: I07e6757f447ad1c4514f2ddcdf26589190445870

inet_ntoa, gethostbyname, gethostbyaddr are all deprecated in the
Windows SDK 8.1; use the replacement functions unless targetting WinXP.

Change-Id: Ic290ce925835c97f0ca0c611d1fd2675d0045ec7

... i.e. decide based on target instead of SDK version.

Change-Id: I225588d34f961c19e223ed9c1b9a934cd5196cf7

The value set here does not correspond to any valid value in
sdkddkver.h; it is a mystery what setting an invalid value is supposed
to achieve.

Change-Id: I76b1da148806e249870154efda3e6c4fa20010bf

Since with the 8.x SDK it is not possible to target Windows XP anyway
according to MSDN, just set _WIN32_WINNT to _WIN32_WINNT_VISTA so that
we can at least work on cleaning up API calls deprecated in SDK 8.1.

Change-Id: Ic433caf454f57187b91ef6fdec3f3bcde3f1af79

... and remove the definitions in makefiles, except in fpicker, which
defines a different value for the Vista file picker.

The WINVER value is derived automatically from _WIN32_WINNT by SDK
header sdkddkver.h.

Change-Id: I73358a9bf5d070f71821654cae2a29335a754a21

... so that -D_WIN32_WINNT=0x0502 does not break the build.

Change-Id: I423ba9c194dc4928d59e1713255bce9c3150bcc2