ruthan wrote:In long term its solution, in short term, i dont believe in it.. […]
Show full quote
In long term its solution, in short term, i dont believe in it..
Because its very long time so Virtual box is available and only missing thing for Windows 98 is write good 3D acceleration driver, same with Dos SB drivers.
QEMU is mess, its very ambitious, it would be ultimate piece of SW which could fix all HW incompatibilities (x86 on ARM/PowerPC.. PowerPC on x86 etc.).. but its incomplete, slow and buggy.. for year and i dont see too much progress.
At least on Windows if im not wrong is there not good GUI frontend for 5+ years. QEMU is would be solution for Mac emulation too, but its the same no 3D video and sound driver for Motorola a PowerPC machines. At least some Xbox 360 emulators tried to use QEMU too but it wasnt success, same as some Android emulator, which even on x86 Android device arent able run Windows 98 / XP on useful level.
Don't underestimate the current capabilities and the new interesting stuff happening with QEMU. In fact, kjliew recently managed to hack 3Dfx Glide pass-through support into QEMU. See this thread: Topic 60950
Since QEMU and DOSBox are both licensed under GPLv2, code can be shared between both projects with no problems. In fact, I believe this has already happened on occasion. However, I would still very much like to see the superior sound card emulation code from DOSBox to be ported over to QEMU. QEMU already supports Sound Blaster 16 and GUS emulation, and it may well be based on older DOSBox code, but as far as I know, it's outdated now, compared to the latest code in DOSBox trunk. I'd especially like to see the recent high quality OPL2 and OPL3 emulation code ported over to QEMU.
That, combined with the Glide pass-through support, would make QEMU a very compelling solution for running DOS/4GW and Win9x games on modern Intel and AMD systems.
As for the lack of a proper cross-platform user-friendly GUI for QEMU, especially on Windows and macOS, I absolutely agree with you. In fact, I think that's the last remaining real advantage that VirtualBox still has going for it over QEMU at this point, at least for non-Linux hosts. And I consider a proper and easy (un)installer to be part of such a user-friendly GUI. But for someone with proper GUI programming skills, this shouldn't be too hard a problem to solve. Perhaps we could work on a cross-platform web-based UI for QEMU. And lacking any other solution, we could even just steal the VirtualBox qt5-based GUI and port that to a cross-platform QEMU distribution. VirtualBox is also GPLv2-licensed, so why not? 😀
Yes, VirtualBox also still has the other advantage of having an integrated hypervisor, allowing it to offer hardware-assisted virtualization even on operating systems that don't include their own hypervisor, such as older versions of macOS, Windows 7 and 32-bit and/or cheaper editions of Windows 10 that don't include Hyper-V and therefore don't offer WHPX. But for Windows users with an Intel CPU, the Intel HAXM hypervisor offers an open source alternative to WHPX, which QEMU also supports. Also, this is becoming less of an issue, since the number of people running any host OS without an integrated hypervisor is dropping fairly quickly these days. And QEMU has supported KVM in Linux for a long time, and as of QEMU 2.12, it now also supports WHPX on Windows 10 and Hypervisor.framework on macOS.
So I'm trying to think what remaining advantages VirtualBox still offers over QEMU, but other than a fairly easy to use and consistent cross-platform GUI and installer, I can't really think of anything. But perhaps I'm forgetting something obvious here?
Long story short: if someone (we perhaps?) would port the most recent sound hardware emulation code from DOSBox to QEMU and if we were then to combine that with the kjliew's 3Dfx Glide pass-through patch, then I believe we will indeed have a very compelling retro game virtualization platform for running DOS/4GW and Win9x games on modern Intel and AMD systems.