Viewing Issue Advanced Details
[ Jump to Notes ]
capcom/cps2.cpp
ID Category [?] Severity [?] Reproducibility Date Submitted Last Update
01653 Interface Trivial N/A Mar 27, 2008, 18:14 Aug 2, 2008, 06:23
Tester Gaius_4 View Status Public Platform MAME (Official Binary)
Assigned To Resolution No change required OS
Status [?] Closed Driver
capcom/cps2.cpp
Version 0.124a Fixed in Version Build Normal
Fixed in Git Commit Github Pull Request #
Summary 01653: sfa, xmcota, ssf2t, (etc.): Control configuration not being loaded correctly.
Description The control configuration for PL1 buttons 4, 5, and 6 seem to have lost their mapping if changed from thier original setting.

I have a Saturn pad and set (past tense) P1 buttons as follows:
PL 1 Button 1...... Joy 1 Button 3
PL 1 Button 2...... Joy 1 Button 4
PL 1 Button 3...... Joy 1 Button 5
PL 1 Button 4...... Joy 1 Button 0
PL 1 Button 5...... Joy 1 Button 1
PL 1 Button 6...... Joy 1 Button 2

and after I updated to .124a they changed to:

PL 1 Button 1...... Joy 1 Button 3
PL 1 Button 2...... Joy 1 Button 4
PL 1 Button 3...... Joy 1 Button 5
PL 1 Button 4...... Shift or Joy 1 Button 3
PL 1 Button 5...... Z or Joy 1 Button 4
PL 1 Button 6...... X or Joy 1 Button 5

Tafoid:
 Something changed between the two that old .CFG are not compatible with new .CFG? I compared a 0.123 cfg and a 0.124 cfg and noticed this:

0.123:
            <port type="P1_BUTTON4" mask="1" index="18" defvalue="1">
                <newseq type="standard">
                    JOYCODE_1_BUTTON1
                </newseq>
            </port>
            <port type="P1_BUTTON5" mask="2" index="19" defvalue="2">
                <newseq type="standard">
                    JOYCODE_1_BUTTON4
                </newseq>
            </port>
            <port type="P1_BUTTON6" mask="4" index="20" defvalue="4">
                <newseq type="standard">
                    JOYCODE_1_BUTTON5
                </newseq>
            </port>

0.124:
            <port type="P1_BUTTON4" mask="1" index="19" defvalue="1">
                <newseq type="standard">
                    JOYCODE_1_BUTTON1
                </newseq>
            </port>
            <port type="P1_BUTTON5" mask="2" index="20" defvalue="2">
                <newseq type="standard">
                    JOYCODE_1_BUTTON4
                </newseq>
            </port>
            <port type="P1_BUTTON6" mask="4" index="21" defvalue="4">
                <newseq type="standard">
                    JOYCODE_1_BUTTON5
                </newseq>
            </port>

The index increased for the game by one on each of the suspect inputs... why?
Steps To Reproduce Tafoid:

- Delete .CFG file for game - (in this case, SFA) (CFG\sfa.cfg)
- Start mame 0.123 (mame123 sfa -joystick)
- Goto Inputs and assign buttons 1-6, Player one
- Exit MAME
- Start mame 0.124a (mame sfa -joystick)

If you notice when you check the input settings, #1-3 Player one buttons are what you set, #4-6 Play one buttons seems to have reverted to original (default) settings. If I remap after this - then exit and restart mame - the bindings seem to stick.
Additional Information
Github Commit
Flags
Regression Version
Affected Sets / Systems sfa, xmcota, ssf2t, (etc.)
Attached Files
  
Relationships
There are no relationship linked to this issue.
Notes
4
User avatar
No.00441
Tafoid
Administrator
Mar 27, 2008, 18:41
 As strange as it might sound, I can confirm this. It might not be just this driver or these games.. it might be a wider issue with .CFG files in general.
User avatar
No.01874
etabeta
Developer
Aug 1, 2008, 22:34
 I don't see this behavior in 126u3, but I'd like to have a confirmation before closing it
User avatar
No.01875
Tafoid
Administrator
Aug 2, 2008, 02:58
 I have to suspect something was simply added which threw off the .CFG files for some people - I remember testing it and all I noticed as I mentioned above was the index increasing. What makes the index increase.. added inputs to the driver or something else? Just like everything else - I'd assume that .CFG files are not guaranteed to work between versions.
User avatar
No.01877
etabeta
Developer
Aug 2, 2008, 06:23
 yes, in 123u1 stephh has improved a lot cps2.c inputs, verifying how many buttons were present for each game in the test mode (independently from the number of used buttons in-game).
this was probably the reason for the increased index number.

anyway, what I meant (sorry, it was late when I posted and I was tired) was that:

1. there has been no other cfg problem in MAME -> no general issue with cfg handling
2. .cfg are not always supposed to be compatible between versions (otherwise we would not have the rule to delete them before testing a bug ;) )

therefore, I close this bug