Mame2003-plus RDB/XML

This might be a stupid question as I’m not 100% sure how everything fits together yet, but hope someone can clear this out for me…

The RDB is missing a lot of entries that are in the mame2003-plus.xml file and as a result a few roms are not picked up when importing on retroarch. That’s only about ~15 roms out of my 300, and I can add them manually to the playlist with a text editor and they do work, so no big problem, but I would just like to learn what’s going on really.

Just as an example of the roms that are not imported, (the main rom for Enduro Racer) is not in the RDB but 2 bootleg roms are. Am I supposed to use one of the bootleg roms instead or should the RDB file match the entries in the XML file?

Yeah, looks like some parent roms do not work all that well in mame2003+ and the bootleg versions are the way to go. A couple seem to run just fine but maybe something will go boom later, lol. So I should trust the .rdb more than the .xml I guess.

It is not stupid, it is very repetitive. :confused:

All Arcade Playlists are created with Manual Scanning and a DAT/XML.
The dats only organize the names. RDBs organize everything and should take care of the names, but… :man_shrugging:

The dats can be downloaded from it or can be generated with the menu of the same emulator, by pressing the tab key.

Captura de pantalla de 2023-11-28 15-44-30

And the dat(xml) generated by the emulator is stored in… The “saves” folder. :man_shrugging:

1 Like

RDBs have to be updated manually by a contributor from a full set in order to get the resulting zip checksums.

I’m not sure about mame2003-plus specifically, but for example for FBNeo this update hasn’t happened for a while.

This is why manual scan is the preferred method since it is more up-to-date and doesn’t depend on having torrentzipped files.

1 Like

I would like to learn how to do this, I have asked for help on github with no success, maybe it is something too easy to explain.

It’s been a while since they have been updated. It’s a mess because the fullsets are combined, he has to take a reference file inside the zip/7z. Which file does he take? Because if it takes them all, the games will appear scattered in all the cores.
And if MAME takes a century with Manual Scan, I don’t even want to imagine with the automatic one.

Well, not mine, it’s now organized with an auto scan, ha. :slight_smile:

Yeah I knew about the xml, that’s why my Q was more about why they differ from the RDB, especially with a core like this that I don’t think is updated that often. I did not remember that a manual scan is still recommended, I don’t do this every year, more like once every 5, I’m getting old.

In this case it was a bunch of roms that actually don’t run that well or had some corrupted graphics when I run them manually, while their bootleg version runs properly, so I’m kinda glad I switched roms and ended up with both a few more properly working roms + matching the rdb/the benefits of an auto scan. Maybe someone tried to include working roms in the RDB? Not sure.

I’m also setting up FBNeo in parallel with its own set and only 1 rom out of 293 does not match, so I just have to deal with that one rom, don’t see a big problem with auto scan, although obv. I go for a modest number of roms.

1 Like

You mean it’s not up to date for the latest stable version of the core ( isn’t it?), or are you running nightlies and it’s just not updated for those?

With the stable version I only have 1 rom not properly imported with an auto scan (but I’m keeping my collection relatively small).

The normal distribution channels will always provide the latest version from git. In that regard, there isn’t really a “stable” release of the core.

The latest RDB update is from April 2022 (link), so whatever changes to the romset from that date won’t be considered.

So, for the auto scan to work, you will have to rebuild that one rom for the core at the date the RDB was created, in case there are any changes, and then torrentzip the resulting file.

Fortunately the database can be found here: libretro-database/metadat/fbneo-split at master · libretro/libretro-database (

The first dat file is used for compiling the RDB, the second one is for building the set.

1 Like

Manual Scan has always been recommended for arcade and it is the best option to keep your playlists up to date.

The DATS/XML is generated by the creators of the sets of the original dumpeos, you have basic information and checksum to check and manage the ROMs.

The RDB ‘RetroArch Data Base’ (i guess), are generated from the dats/xml, games are not included because they appear in the RDB, it is the opposite.

mame2003+ is updated 1 or 2 times a month and fbneo 4 or 6 times a month. If the last RDB is from April 2022, then there are as 24 current versions of FBNeo. You can always make a backup and freeze the Core.

Scan Directory does not work, I did the test with Mortal Kombat 2 in each Core, no list is generated, and the MAME games appear in FBNeo, although, this has the names of the different ROMs.

Please do the lists as you want. But perhaps it is better to use the method recommended by developers of emulators/cores. I already passed the official FBNeo link, it is the same process for all the arcades.

1 Like