Requiring SSE2 on Windows

[Stephan Bergmann]

On 05/05/2020 15:48, Luboš Luňák wrote:
>   I suggest that we remove the '-arch:SSE' added on Windows, i.e. that we
> require at least SSE2-capable systems.
> 
>   This was added by 8bd6bf93b7 in 2014, but at this point there's basically
> nobody who would be running LO but wouldn't be capable of SSE2.
> 
>   Some relevant facts:
> - GCC, Clang and MSVC all default to SSE2+
> - The last Windows to work without at least SSE2 was Windows 7, which is EOL
> (and in fact it in practice required SSE2+ too, since an update in 2018).
> - Windows 8, MSOffice 2013, Chromium require SSE2+.
> - I checked all ~240k crash dumps reported in the last month and there's not a
> single one from a machine not capable of SSE2.

(FWIW, when -msse2 was discussed for 32-bit Linux x86 in the context of 
<https://lists.freedesktop.org/archives/libreoffice/2016-September/075211.html> 
"Re: Test File: sc/qa/unit/data/functions/fods/chiinv.fods: fails with 
Assertion", it was turned down in replies 
<https://lists.freedesktop.org/archives/libreoffice/2016-September/075324.html> 
and 
<https://lists.freedesktop.org/archives/libreoffice/2016-September/075316.html>.)