Add MAME configs for UNIX workstations and servers

[ajacocks]

Would you consider adding the configurations for the following UNIX workstations and servers to Ample? It's such a fantastic front end, and it'd be great to do all of my non-gaming MAME emulation using it:

1. ibm/rtpc
2. ibm/rs6k7009
3. hp/hp9k310
4. hp/hp9k320
5. hp/hp9k330
6. hp/hp9k332
7. hp/hp9k340
8. hp/hp9k360
9. hp/hp9k370
10. hp/hp9k380
11. hp/hp9k382
12. dec/d2k300axp
13. dec/d2k500axp
14. dec/dpcaxp150
15. dec/ds2100
16. dec/ds3100
17. dec/ds5k133
18. dec/pdp11qb
19. dec/pdp11ub
20. dec/pdp11ub2
21. intergraph/ip2000
22. intergraph/ip2400
23. intergraph/ip2500
24. intergraph/ip2700
25. intergraph/ip2800
26. intergraph/ip6000
27. intergraph/ip6400
28. intergraph/ip6700
29. intergraph/ip6800
30. mips/rc2030
31. mips/rc3230
32. mips/rs2030
33. mips/rs3230
34. sgi/crimson
35. sgi/indigo2_4415
36. sgi/indy_4610
37. sgi/indy_4613
38. sgi/indy_5015
39. sgi/indigo
40. sgi/pi4d30
41. sgi/pi4d35
42. sgi/indigo_r4000
43. sgi/indigo_r4400
44. sgi/pi4d20
45. sgi/pi4d25
46. sgi/iris3130
47. sgi/4d410
48. sgi/o2
49. sgi/octane
50. sgi/pi4d50
51. sgi/pi4d70
52. sony/nws3860
53. sony/nws1580
54. sony/nws3260
55. sony/nws3410
56. sony/nws5000x
57. sun/sun1
58. sun/sun2_120
59. sun/sun2_50
60. sun/sun3_110
61. sun/sun3_150
62. sun/sun3_260
63. sun/sun3_50
64. sun/sun3_60
65. sun/sun3_e
66. sun/sun3_460
67. sun/sun3_80
68. sun/sun4_110
69. sun/sun4_20
70. sun/sun4_25
71. sun/sun4_300
72. sun/sun4_400
73. sun/sun4_40
74. sun/sun4_50
75. sun/sun4_60
76. sun/sun4_65
77. sun/sun4_75
78. sun/sun_s10
79. sun/sun_s20

This is a lot of drivers, I know. So, would you alternatively consider making the interface driven from config files, so that I could create a UNIX drivers list, along with their associated configs (RAM, boards, etc.)?

[ksherlock]

That is a big list.

The UI is driven by config files (example) which are pre-generated from the mame -listxml output because -listxml is often inconsistent and lies about some things so I don't want to do it at runtime.

However, there is a hardcoded list of slots within Ample so that needs to be changed before it will work for arbitrary machines. That is something I've been meaning to address so I'll work on that.

[ksherlock]

Do any of them work? Most are flagged as MACHINE_NOT_WORKING or MACHINE_IS_SKELETON.

[ajacocks]

Yep! Quite a few work rather well.

Offhand:

• HP 9000/3xx (https://wiki.mamedev.org/index.php/Driver:HP300)
• HP 9000/4xx (https://wiki.mamedev.org/index.php/Driver:Apollo)
• SGI Indy (https://forums.bannister.org/ubbthreads.php?ubb=showflat&Number=114911#Post114911)
• SGI Indigo (https://forums.bannister.org/ubbthreads.php?ubb=showflat&Number=122528&Searchpage=1&Main=9487&Words=irix&Search=true#Post122528)
• Sun SPARC (https://forums.bannister.org/ubbthreads.php?ubb=showflat&Number=44666&page=515 and https://computernewb.com/wiki/Boot_SPARCStation_1_in_MAME)
• Intergraph (https://wiki.mamedev.org/index.php/Driver:InterPro)

I'd be happy to help, if there is something that I can do

[ksherlock]

Here is a version of Ample Lite with most of the machines above. You'll need to bring your own MAME.

[ajacocks]

Looks good! I tested with MAME 0.260 from https://sdlmame.lngn.net/

I used the following launch shell script:

#!/bin/bash

#
# Obtain path to script.
#
# Note that 'grealpath' provides reliable soft-link resolution, without requiring
# any Bash gymnastics.
#
script_file=$(/opt/local/bin/grealpath "${0}")
script_dir=$(dirname "${script_file}")

cmd="${script_dir}/mame"
bgfx_path="${script_dir}/bgfx"
languagepath="${script_dir}/language"
pluginspath="${script_dir}/plugins"
artpath="${script_dir}/artwork"
hashpath="${script_dir}/hash"
ctrlrpath="${script_dir}/ctrlr"

cmdline=()
cmdline+=( "${cmd}" )
cmdline+=( -bgfx_path "${bgfx_path}" )
cmdline+=( -languagepath "${languagepath}" )
cmdline+=( -pluginspath "${pluginspath}" )
cmdline+=( -artpath "${artpath}" )
cmdline+=( -hashpath "${hashpath}" )
cmdline+=( -ctrlrpath "${ctrlrpath}" )

# Note that Mame handles duplicate arguments, using the last one found.
# Net-Net, there's no need to check whether the user specified the same
# parameters, as theirs will take precedence.

"${cmdline[@]}" \
"${@}"

I used https://arcade.vastheman.com/minimaws/machine/ as a reference, and verified by running manually from the shell.

Testing the systems with default config, one by one, yielded these results:

1. DEC VT52 - works fine
2. DEC VT100 - works fine
3. DEC VT101 - works fine
4. DEC VT102 - works fine
5. DEC VT240 - works fine
6. DECstation 2100 - initializes correctly and hangs (as expected)
7. DECstation 3100 - initializes correctly and hangs (as expected)
8. DECstation 5000/133 - initializes correctly (with expected errors) and gets to SRM
9. PDP-11 (QBUS) - works fine, except that attaching scsi disks to the TTI QTS-1 fails due to command-line generation problems
   • Ample generates this for for a qts-1 scsi host in sbus slot 1:
     • mame.sh pdp11qb -skip_gameinfo -nosamples -window -nomax -video bgfx -rs232 null_modem -qbus:1 qts1 -:scsibus:0 aplcdsc
   • The correct command-line should be:
     • mame.sh pdp11qb -skip_gameinfo -nosamples -window -nomax -video bgfx -rs232 null_modem -qbus:1 qts1 -qbus:1:qts1:scsibus:0 aplcdsc
   • There should also be a BIOS version choice for the QTS-1 of 1.7 or 2.3, and an RS-232 interface
10. the pdp11ub and pdp11ub2 seem to be exact clones, and not actual unibus pdps, so we can probably mark these as non-working for now
11. HP 9000/310 - works fine
12. HP 9000/320 - works, but there seems to be an extra pair of ' characters in the CLI generator somewhere. For example:
    • mame.sh hp9k320 -skip_gameinfo -nosamples -window -nomax -video bgfx -sl1 98550 -sl2 '' -hard1 /Users/ajacocks/Documents/hp9k320-hdd.img -cdrm1 /Users/ajacocks/Downloads/hpux9_install.iso
13. HP 9000/330 - works fine
14. HP 9000/332 - works fine
15. HP 9000/340 - works fine, seems to be some quoting issues in the CLI generator, maybe this applies to all?
16. HP 9000/360 - works fine
17. HP 9000/370 - works fine, doesn't find a boot device, maybe MAME?
18. HP 9000/370 - works fine
19. HP 9000/382 - works fine
20. IBM PC RT 010 - can't select 5.25" disks for 5.25" drives. 3.5" drives seem to allow disk selection. Seems to suffer from similar CLI issues as the qbus pdp-11
    • for example, this:
      • mame.sh rtpc010 -skip_gameinfo -nosamples -window -nomax -video bgfx -:fdc:1 '' -flop1 '/Users/ajacocks/Downloads/IBM AIX 2.1.1 (5.25-1.2mb) [IBM RT]/maint/disk01.img'
    • should be:
      • mame.sh rtpc010 -skip_gameinfo -nosamples -window -nomax -video bgfx -isa6 fdc -isa6:fdc:fdc:0 35hd -isa6:fdc:fdc:1 525hd
    • I need to go through the floppy selection logic with a finer-toothed comb. 3.5" HD floppy drives don't seem to show up in the command-line, but 3.5" DD ones do
21. IBM PC RT 015 - same as 010
22. IBM PC RT 020 - same as 010
23. IBM PC RT 025 - same as 010
24. IBM PC RT A25 - same as 010
25. Intergraph machines power on, but will need more investigation, as I am not familar with CLIX
26. MIPS RC2030 - works fine
27. MIPS RS2030 - works fine
28. MIPS RC3230 - works fine
29. MIPS Magnum 3000 - works fine
30. SGI IRIS Indigo - works fine
31. SGI Indigo 2 - works fine
32. SGI IRIS Indigo R4000 - works fine
33. SGI IRIS Indigo R4400 - black screen, likely MAME
34. SGI Indy R4600 100 - works fine
35. SGI Indy R4600 133 - works fine
36. SGI Indy R5000 - works fine
37. SGI Personal IRIS 4D/20 - works fine
38. SGI Personal IRIS 4D/25 - works fine
39. SGI Personal IRIS 4D/30 - black screen, likely MAME
40. SGI Personal IRIS 4D/35 - works fine
41. Sony NWS-3260 - works fine
42. Sony NWS-3410 - flashing green cursor
43. Sony NWS-1580 - works fine
44. Sony NWS-5000X - works fine
45. Sun-1 - works fine
46. Sun 2/50 - works fine
47. Sun 2/120 - works fine
48. Sun 3/50 - works fine
49. Sun 3/60 - works fine
50. Sun 3/110 - works fine
51. Sun 3/75-180 - works fine
52. Sun 3/260 - works fine
53. Sun 3/E - black screen
54. Sun 3/80 - black screen
55. Sun IPC - works fine
56. Sun IPX - works fine
57. Sun SLC - works fine
58. Sun ELC - white screen
59. Sun 1+ - works fine

There are lots of card selections that aren't implemented for various systems, but I can go through those in detail in a different report. This one took a few hours to test out, already!

Thanks for the additions!

[ajacocks]

• Verified that commit dc6a215 fixed the 5.25" FDD issue with IBM RT.
• Verified that the SCSI device command-line selection seems to be fixed by commit 43a7f31
• The hp9000 slot selection cli generation is better, but you can still end up with blank slot entries by removing default cards
  • for example, removing the BASIC ROM card from slot 2:
    • mame.sh hp9k340 -skip_gameinfo -nosamples -window -nomax -video bgfx -sl2 '' -hard1 /Users/ajacocks/Documents/ip2000.chd -cdrm1 /Users/ajacocks/Downloads/hpux9_install.iso

[ksherlock]

Thanks for the testing and feedback. Ample tries to keep the previous selection for slots of the same name when the machine changes. For Apple II and Mac I think that makes sense, but I'll think about if there's a better way to handle that. You can right click and reset slots.

[ajacocks]

Glad to help!

I’d like to be able to assist with the configuration of existing machines and the addition of new ones, but I’m still working to understand the Python code.

The plists are pretty easy to understand, so I will probably start there.

[ajacocks]

I've finally gotten back to working on this, and I started with "indy_4610", since that works quite well in 0.288.

I noticed in Resources/indy_4610.plist that it has a slot device called "smartport". Obviously, the Indy doesn't have an Apple smartport, so is this just some kind of shorthand for where disk drives are connected? I couldn't see a way in the .plist to disable the literal flag -smartport, which is not supported by the indy driver.

Do you have a script that does .plist creation from the original .xml dump, since you mentioned that running it on launch is unreliable? or, do you do it by hand? I've been trying to use several other .plists as reference, but couldn't find one with scsi that didn't use "smartport".

Also interesting for this driver is that the "-ramsize" parameter no longer works. It now completely relies on slot-based RAM. I'm not sure how that will work in the context of Ample as a launcher that uses the CLI. There would likely have to be a .cfg generated with this content (assuming 256MB RAM):

<input>
    <port tag=":memctrl:SIMMS" type="CONFIG" mask="15"  defvalue="3" value="14" />
    <port tag=":memctrl:SIMMS" type="CONFIG" mask="240" defvalue="0" value="224" />
</input>

[ksherlock]

"smartport" is an internal, historical identifier. Real Ample doesn't use it in the generated command line (it generates -scsibus:2 harddisk, etc for indy_4610)

[ajacocks]

Interesting. I am testing on AmpleLinux, from the main branch. When launched, the default execution command for the "indy_4610" driver is:
mame indy_4610 -skip_gameinfo -bios "" -smartport "" -scsibus:1 harddisk -scsibus:2 "" -scsibus:3 "" -scsibus:4 "" -scsibus:5 "" -scsibus:6 cdrom -window -resolution 2560x2048 -video bgfx -nosamples -inipath /home/ajacocks/src/ample/AmpleLinux/mame -rompath /home/ajacocks/src/ample/AmpleLinux/mame/roms

If I click somewhere else in the window, and click back, it becomes:
mame indy_4610 -skip_gameinfo -scsibus:1 harddisk -scsibus:2 "" -scsibus:3 "" -scsibus:4 "" -scsibus:5 "" -scsibus:6 cdrom -window -resolution 2560x2048 -video bgfx -nosamples -inipath /home/ajacocks/src/ample/AmpleLinux/mame -rompath /home/ajacocks/src/ample/AmpleLinux/mame/roms

These are my only changes, so far:

 ajacocks  fw13  ~  src  ample  AmpleLinux   master  2✎  4+  $  git diff ../Ample/Resources/models.plist
diff --git a/Ample/Resources/models.plist b/Ample/Resources/models.plist
index 0a36eaf..3cedc3e 100644
--- a/Ample/Resources/models.plist
+++ b/Ample/Resources/models.plist
@@ -1129,5 +1129,18 @@
         </dict>
       </array>
     </dict>
+    <dict>
+      <key>description</key>
+      <string>Silicon Graphics</string>
+      <key>children</key>
+      <array>
+        <dict>
+          <key>description</key>
+          <string> Indy (R4600, 100MHz)</string>
+          <key>value</key>
+          <string>indy_4610</string>
+        </dict>
+      </array>
+    </dict>
   </array>
 </plist>
 ajacocks  fw13  ~  src  ample  AmpleLinux   master  2✎  4+  $  git diff ../Ample/Resources/roms.plist
diff --git a/Ample/Resources/roms.plist b/Ample/Resources/roms.plist
index 6aaecd1..b131255 100644
--- a/Ample/Resources/roms.plist
+++ b/Ample/Resources/roms.plist
@@ -1909,6 +1909,12 @@
       <string>pd3_mclr</string>
       <key>description</key>
       <string>Micron/XCEED Technology MacroColor 30</string>
+    </dict>
+     <dict>
+      <key>value</key>
+      <string>kb_ms_natural</string>
+      <key>description</key>
+      <string>Microsoft Natural Keyboard</string>
     </dict>
     <dict>
       <key>value</key>
@@ -2264,6 +2270,12 @@
       <key>description</key>
       <string>SDTANDY FDC</string>
     </dict>
+    <dict>
+      <key>value</key>
+      <string>indy_4610</string>
+      <key>description</key>
+      <string>SGI Indy (R4600, 100MHz)</string>
+    </dict>
     <dict>
       <key>value</key>
       <string>gpib_hardbox</string>

I have confirmed that this only happens if the indy_4610 driver is the last one used when you close Ample. When you re-open Ample, the launch command has -smartport "" in it until you click another driver, and click back on indy_4610.

NOTE Confirmed that this happens on all 4 SGI drivers that I have enabled.

[ajacocks]

Added PR #55 to enable working SGI drivers in the main branch