Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Seefwi]

Roman,

I was looking for an older version of clrmamepro (specifically the 64-bit version of 3.138a) so that I could run some comparisons between it and the latest version. So far, I have come up empty in my search. Are such older versions generally available? If so, could you tell me where I might find them? If not, would it be possible for me to get a copy of the aforementioned version from you directly?

Thanks for your time.

[Roman]

Generally I got all versions (well, from 3.111 onwards) in my SVN repository but I don't keep binaries.

What aspect do you want to compare anyway. It is recommended to use the latest (4.01) since it a) supports unicode and b) is faster in all aspects compared to 3.x.

[Seefwi]

What aspect do you want to compare anyway.

In short, someone created a torrent consisting of files scanned with clrmamepro 3.138a and run through torrentzip. For whatever reason, even after freshly installing all the tools (including clrmamepro 4.01) and following the instructions of the torrent creator exactly (in my mind), I can't get the torrent data to match up with the clrmamepro scan using the datfile the torrent creator provided. (The errors have to do with "not separated biosrom file" and "missing BIOS rom".)

This means that one of two things is going on:

• I've made a mistake somewhere. (Which is entirely possible and probably the more likely scenario; however, as I mentioned above, I followed the procedure given by the torrent author exactly and still ran into the problem.)
• There is some difference in the behaviour of the newest version of clrmamepro relative to version 3.138a.

Again, chances are I've made a mistake somewhere, but for my own sanity I'd like to ensure that is the case.

It is recommended to use the latest (4.01) since it a) supports unicode and b) is faster in all aspects compared to 3.x.

I am using the latest version and plan to continue to use the latest version going forward. On the off chance, however, that there is a discrepancy between 3.138a and 4.01, it might be useful for some people to know.

[Seefwi]

I should add that I wasn't the only person to run into this problem with this particular torrent, although I obviously can't speak to whether the others properly followed the procedures laid out by the torrent creator.

[Roman]

people shouldn't use torrentzip...easy as that
It does not support utf8 filenames and will actually kill utf8 stored names since it will reencode the names with local page encoding and in the past it used to have wrong size entries in the zipstructures which didn't match the real comment field length which was used to store the checksum etc...no idea if they ever fixed that...
...but besides that, it shouldn't be a problem to scan the provided data/files.
If you get not separated biosrom file /missing BIOS rom messages you should investigate the datfile for the mentioned name. Especially if the BIOS entries are well defined in the dat. There was a change in 4.x regarding the option to "keep bios files separated" but actually it was only an alignment of the rebuilder/scanner functions for this.
Default should be "separate bios sets" enabled in scanner advanced and rebuilder advanced.

If you find something spooky with more details, let me know

[Seefwi]

If you get not separated biosrom file /missing BIOS rom messages you should investigate the datfile for the mentioned name. Especially if the BIOS entries are well defined in the dat. There was a change in 4.x regarding the option to "keep bios files separated" but actually it was only an alignment of the rebuilder/scanner functions for this.
Default should be "separate bios sets" enabled in scanner advanced and rebuilder advanced.

I took a look at the datfile, and the problems stem from the only two bios entries where a bios is defined as using another bios, and games are using the derivative bios. They both look something like this:

<game name="foo" isbios="yes">
        <rom name="baz"/>
</game>

<game name="bar" isbios="yes" romof="foo">
        <rom name="baz" merge="baz"/>
</game>

<game name="game" romof="bar">
        <rom name="baz" merge="baz"/>
</game>

In these cases, clrmamepro 4.01 wants to keep the two bios copies of baz separate with one copy in each of the bios zips, and wants to leave baz out of any games defined as using bar as a bios (given a split-merged set).

On the other hand, the torrent files, which were supposedly successfully scanned with clrmamepro 3.138a, keep only a copy of baz in the foo bios, leaving the bar bios without baz. The games using the bar bios, however, include baz.

Again, I don't have a copy of 3.138a to verify that it does in fact do this, but seemingly knowledgeable people claim that the torrent files scan fine with that version.

All the rest of the games that simply refer to a single bios scan fine.

[Roman]

ah that...ok..that has nothing to do with torrent(zip).

Look at the history of 4.00b:
fixed: biossets with set romof are not handled as biossets (sys1 with soundboard)

So the current (>=4.00b) interpretation of that constellation is imo more correct. Since it's really a BIOS set, it will be handled as one (and not as a normal set (bar) with reference to a BIOS set (foo)).

[Seefwi]

Look at the history of 4.00b:
fixed: biossets with set romof are not handled as biossets (sys1 with soundboard)

Ah, I didn't think to check that history, sorry.

Well that puts my mind at ease. Thanks for taking the time to help solve this little mystery of mine, and for your work on clrmamepro.