openMSX post-0.7.0 build available for beta testing

Ascended (19298)

02-05-2009, 23:48

OK folks, here's the latest builds, which should be very similar to the next release. We really need you people to help us find those last bugs.... Check those out:

http://openmsx.sf.net/temp/openmsx-0.7.1-dev9618-VC-x86.msi

or, for 64-bit Windows:

http://openmsx.sf.net/temp/openmsx-0.7.1-dev9618-VC-x64.msi

We really need you to test this now and not have situations like:
- "I hope this and that bug will be fixed in next release" -> don't hope, make sure it is, by telling us about the problem!
- "I just installled 0.7.1 and it bugs all over the place" -> prevent this by doing this now (it really won't hurt!) and tell us your experiences!

Thanks in advance!

PS: replace .msi with .zip if you'd rather not use the brand new shiny installer.

By Metalbrain

Expert (67)

03-05-2009, 01:56

The wrong C-BIOS roms are still included.

By Manuel

Ascended (19298)

03-05-2009, 11:58

MetalBrain: do you mean there's still a bug in the C-BIOS ROMs? If so, that could be correct, because there has been no new release of C-BIOS, so the latest release has been included. I did file a bug report on C-BIOS for this, please check if it's correct: https://sourceforge.net/tracker/?func=detail&aid=2786007&group_id=123693&atid=697337

By Metalbrain

Expert (67)

03-05-2009, 16:30

The C-BIOS 0.23 can be compiled with the bit for either 50Hz or 60Hz, by default they are compiled with this:

; -- generic european config (50Hz, US keyboard layout)
LOCALE_CHSET:   equ     LOCAL_CHSET_US
LOCALE_DATE:    equ     LOCAL_DATE_DMY
LOCALE_INT:     equ     LOCAL_INT_50HZ
LOCALE_KBD:     equ     LOCAL_KBD_US
LOCALE_BASIC    equ     LOCAL_BASIC_US

but the japanese versions are compiled with this:

; -- japanese config
LOCALE_CHSET:   equ     LOCAL_CHSET_JP
LOCALE_DATE:    equ     LOCAL_DATE_YMD
LOCALE_INT:     equ     LOCAL_INT_60HZ
LOCALE_KBD:     equ     LOCAL_KBD_JP
LOCALE_BASIC    equ     LOCAL_BASIC_JP

In the main code, the byte is generated as:

idbyt1:
; Basic ROM version
; 7 6 5 4 3 2 1 0
; | | | | +-+-+-+-- Character set
; | | | |           0 = Japanese, 1 = International (ASCII), 2=Korean
; | +-+-+---------- Date format
; |                 0 = Y-M-D, 1 = M-D-Y, 2 = D-M-Y
; +---------------- Default interrupt frequency
;                   0 = 60Hz, 1 = 50Hz
                db      LOCALE_CHSET + LOCALE_DATE + LOCALE_INT

In version 0.21 of C-BIOS, that byte had a fixed value of 21h.

The japanese ROMs aren't included in the download, but if you manually compile them (e.g.: typing "make pasmo" in the src directory) they will be generated.

Now I don't know about the specifications of a C-BIOS machine, or how the 50Hz/60Hz difference is specified for the different machines. If the C-BIOS machines are supposed to run at 60Hz, the proper (japanese) ROMs should be included. If they're supposed run at 50Hz, then it's a problem with openMSX (in their machine definition). If they can be either, and the emulator should auto-select the refresh speed, then it's a problem with openMSX (the autodetection). If they can be either, but the refresh rate must be initialized in ROM, then it's a bug in the 50Hz C-BIOS (and therefore I'd recommend including the japanese versions).

I've just seen that the new blueMSX 2.8.1 suffers the same problem.

I've discovered this because my game I Need Speed trusts that bit to choose between 50/60Hz mode, and the music plays too fast if it runs with 50Hz code on a 60Hz computer. Maybe there's a better way to detect the vsync rate?

By hap

Paragon (2042)

03-05-2009, 16:58

The European C-BIOS is meant to be 50Hz.

Maybe there's a better way to detect the vsync rate?Yes! Create a simple 1/55s waitloop at the start of vblank, and then check the vblank bit.

By Manuel

Ascended (19298)

05-05-2009, 23:10

Any other comments from other people? Unfixed bugs?? Anything? Is the new installer OK? Cool? Uncool? Suck?

By Alex

Master (205)

06-05-2009, 00:10

I have tested the new build under Vista and under Xp.

It works fine now. The problem with Catapult and special characters that I reported earlier is now resolved.

By Manuel

Ascended (19298)

07-05-2009, 20:46

Thanks Alex. Any other people who want to give it a spin?

By Manuel

Ascended (19298)

27-05-2009, 20:50

Latest binaries: http://openmsx.sf.net/temp/openmsx-0.7.1-dev9895-VC-x86.msi or http://openmsx-0.7.1-dev9895-VC-x64.msi
(zip also available)

We're now very close to a release, so here's one of your last chances to find bugs before releasing! Check those out.

Also, if you are interested, try building openMSX from source on your own platform. Much of the probing part of the build system has been rewritten and could use more testing.

By RyJuZo

Master (236)

28-05-2009, 20:12

Latest binaries: http://openmsx.sf.net/temp/openmsx-0.7.1-dev9895-VC-x86.msi or http://openmsx-0.7.1-dev9895-VC-x64.msi
(zip also available)

We're now very close to a release, so here's one of your last chances to find bugs before releasing! Check those out.

Also, if you are interested, try building openMSX from source on your own platform. Much of the probing part of the build system has been rewritten and could use more testing.

had some issues with the previous version in screen 0 width 80 and blinking...
my scc tracker uses this mode and when I run it the blinkpositions are all messed up...but when running release 0.63 all is fine...
I will check if this version has the same issues...
I'll be back...;)