|
Re: End of 32-bit
10/28/19 05:14 AM
|
|
|
> Microsoft is probably the sole reason why support for 32-bit (x86) PCs has been > around for so long - making $$$$$$$$$$ on upgrading all those XP and Vista machines > when Windows 7 could have easily been a 64-bit only release, given that 32-bit x86 > CPUs were practically dead in the water even in 2009 when Windows 7 was first > released (as in, PCs for the general public, not embedded/industrial machines where > older operating systems tend to remain virtually untouched anyway e.g. ATMs and > point-of-sales machines that still use Windows XP).
Windows 8 and Windows 10 support 32-bit x86 hosts because of relatively recently released Intel Atom notebooks and tablets. Microsoft would have liked to ditch 32-bit host support earlier.
> That said, I wouldn't be surprised if the last 32-bit versions are given an > unofficial fork just to keep cabinet owners and other PC hardware laggards in the > loop (e.g. when more ROM sets and decapped/emulated protection devices are inevitably > added to the upcoming 64-bit only builds, they may be able to be back-ported to the > older MAME versions if MAME's code hasn't changed significantly), much like how > RetroArch etc. maintains 20-year old versions of MAME solely for compatibility with > potato-powered toasters.
You'll still be able to build for 32-bit targets. We're just not going to be distributing 32-bit Windows binaries ourselves. There's no reason to fork the last release that has pre-built 32-bit binaries, just like there was no reason to fork the last release that had pre-built DEBUG=1 binaries.
Dropping GCC5/GCC6 support is a year overdue. For various reasons, I've been two weeks of actual work away from actually being able to push the change for a year.
|
|