VOGONS

Common searches


Reply 301 of 343, by ykhwong

User metadata
Rank Oldbie
Rank
Oldbie

The existing SVGA patch seems not to work with a new implementation of ega machine.
Just a note.

2007-09-20 18:42  c2woody

* include/dosbox.h, src/dosbox.cpp, src/dos/dev_con.h,
src/dos/dos_keyboard_layout.cpp, src/hardware/dma.cpp,
src/hardware/gus.cpp, src/hardware/sblaster.cpp,
src/hardware/vga.cpp, src/hardware/vga_attr.cpp,
src/hardware/vga_crtc.cpp, src/hardware/vga_dac.cpp,
src/hardware/vga_draw.cpp, src/hardware/vga_gfx.cpp,
src/hardware/vga_memory.cpp, src/hardware/vga_misc.cpp,
src/hardware/vga_seq.cpp, src/hardware/vga_xga.cpp,
src/ints/bios.cpp, src/ints/int10.cpp, src/ints/int10_char.cpp,
src/ints/int10_memory.cpp, src/ints/int10_modes.cpp,
src/ints/int10_pal.cpp: add partial ega-only machine

Reply 306 of 343, by vasyl

User metadata
Rank Oldbie
Rank
Oldbie

Checked... ouch! This may take some time to integrate and test properly. Looks like I've been out of the loop for too long.
@wd: When do you expect the rest of your changes to go in? If it's less than a week, I'll wait. Otherwise I will risk doing the same work twice 😖

Reply 309 of 343, by DOSGuy

User metadata
Rank Newbie
Rank
Newbie

Hi vasyl.

First of all, great job on the SVGA plugin. I've been able to use an emulator called RIXAI to emulate an 8514/a adapter using the Tseng ET4000 svgachipset. To the best of my knowledge, I'm the first person to play Mah Jongg -8514- in DOSBox. That's a pretty amazing accomplishment for your plugin considering that you weren't even working on 8514 emulation.

I'm getting a ton of errors in the status window that look like:

Illegal write to c003364b CS:IP    c7ff:      1d,phys c003364b

Some kind of register issue? Just wanted to bring this to your attention. Links to Mah Jongg -8514-, RIXAI, and the discussion that led to all of this are at Accelerator board? In those old days?

"Today entirely the maniac there is no excuse with the article." Get free BeOS, DOS, OS/2, and Windows games at RGB Classic Games

Reply 310 of 343, by vasyl

User metadata
Rank Oldbie
Rank
Oldbie

I will take a look. I haven't seen that error message before, could be a bug in emulation but it might be in the game or RIXAI utility. I am quite curious about that utility now -- it supposedly does register-level emulation, that's unusual for DOS TSRs.

Reply 312 of 343, by DOSGuy

User metadata
Rank Newbie
Rank
Newbie

This plugin stores its variables in a section of dosbox.conf called "VGA". There are two variables: svgachipset and videoram. S3 is the default, but that's no guarantee of compatibility. Games that support SVGA resolutions will usually indicate which chipsets they support, and often how much video RAM they need. I used to get by with my Trident 1 MB card, but I couldn't display as many colors as I wanted at the resolutions I wanted in Windows 3.1. "2048" (2 MB) should be a safe minimum.

"Today entirely the maniac there is no excuse with the article." Get free BeOS, DOS, OS/2, and Windows games at RGB Classic Games

Reply 313 of 343, by Dominus

User metadata
Rank DOSBox Moderator
Rank
DOSBox Moderator

You have to edit dosbox.conf directly and you need to use a special build of Dosbox that has this patch integrated. Lock, stock Dosbox does not have this.

Windows 3.1x guide for DOSBox
60 seconds guide to DOSBox
DOSBox SVN snapshot for macOS (10.4-11.x ppc/intel 32/64bit) notarized for gatekeeper

Reply 314 of 343, by DosFreak

User metadata
Rank l33t++
Rank
l33t++

Should Hal's patch be included with this one? Is Vasyl still alive? I guess the question is, Is Hal's S3 patch a candidate to be included in the official version of DosBox?

How To Ask Questions The Smart Way
Make your games work offline

Reply 316 of 343, by vasyl

User metadata
Rank Oldbie
Rank
Oldbie

Is Vasyl still alive?

Yes, I am pretty much alive, just very busy lately -- not much time left between day job and family. There was also a mishap while upgrading main home PC, that took much more of my time than expected (and two motherboards sent to warranty repairs). I do remember about my promise to split my patch into manageable chunks... I will get to it when I can.

Reply 317 of 343, by DosFreak

User metadata
Rank l33t++
Rank
l33t++

heh, Just joking man hadn't seen you around here in awhile is all.

Too bad about the motherboards. Haven't had that happen to me yet. Been pretty lucky in that regard when building my computers.

How To Ask Questions The Smart Way
Make your games work offline

Reply 318 of 343, by vasyl

User metadata
Rank Oldbie
Rank
Oldbie

Yet another refresh. No functional changes, just making it to compile and run with recent CVS. The code is starting to show some staleness, e.g., there are a couple of S3-isms in vga_draw.cpp that I did not clean up yet. On the positive side, some things became more consistent between CVS and the patch, like those screen geometry calculations in SetupDrawing -- my patch had some fixes there before, now not needed.
Now, the plan. Looks like I may have a little time in a few coming weeks so I finally can break the patch down to pieces and present it to DOSBox devs so we may see at least some parts merged in.
1) there are a few generic bug fixes in the patch, nothing major but they don't belong to this patch. These will be taken care of separately;
2) a lot of changes (mostly in vga_memory and vga_draw) are due to the dynamic VRAM allocation. This can be done as a separate "second stage" patch;
3) large part of the patch is caused by moving mode lines from int10_modes to drivers. It is almost correct but not exactly. The "correct" layered solution would be to actually have driver-specific branches of int10_modes (int10_modes_s3, int10_modes_tseng, etc.) For the simplicity I am going to move those pieces back to int10_modes and add checks for SVGA type where needed (those checks are necessary anyway, now that we have VGANone in CVS);
4) currently, some lines are moved from other files to vga_s3.cpp. Strictly speaking, this is not necessary. I will try to rewrite relevant parts to minimize the diff;
5) I am going to move chipset control from [vga] section to machine= line for consistency with some recent changes.
Stay tuned.

Reply 319 of 343, by vasyl

User metadata
Rank Oldbie
Rank
Oldbie

Important status update. The outlined plan is complete and, big thanks to wd, the essential portion of the patch is already in CVS. Pretty much the only part of the patch that is still not in is the dynamic VRAM allocation and boundary checks. I will update the patch on SF shortly. The remainder is about 1/4 in size.
There is one important functional change. Instead of [vga] section, machine= line in [dosbox] section is used. In addition to existing values you can specify "vga_et4000", "vga_et3000", or "vga_pvga1a" with some alternative spellings ("svga" can be used instead of "vga"). The updated version of the patch will allow to specify VRAM size on the same line as a number; it will also understand "new" for ET4000 chipsets.