Hello,

I was wondering if there was a plan to implement rebuilding from a single source into multiple DAT's at the same time. This would save a lot of time if you could just load multiple dat's into memory, point the rebuilder at a directory, and it organized the files according to the ROM directories specified in each of the loaded DAT's. I know that TIM does this, and it makes reorganizing a directory of tosec files very easy. I'd just like to know if this kind of thing is possible with clrmame or if it may be implemented in the future.

Thanks

It won't be included. Use the batch run options instead. It allows you to auto-rebuild, auto-scan, auto-create profiles and rompaths for billions of tosec et al stuff. You can reach the 'Batcher' by selecting more than one profile to load.


Roman Scherzer

I could be wrong, but this will rescan the directory over and over again for each dat right? If I had say 20,000 files, it would be a lot easier to scan it once rather than it being scanned over and over again. If I had 100 dat's setup in the profiler, it could possibly take 100 scans of the same directory to pick up a file. This other way would scan the directory once and determine which DAT file the file was associated with.

-sl0ppy

> It won't be included. Use the batch run options instead. It allows you to
> auto-rebuild, auto-scan, auto-create profiles and rompaths for billions of tosec
> et al stuff. You can reach the 'Batcher' by selecting more than one profile to
> load.
>
>
> Roman Scherzer
>

Well yes, the source will be scanned per profile but scanning time is usually not a thing to care about if the files are packed. Either do it that way or merge dats.


Roman Scherzer

Just something to consider I suppose. I just found it really easy to point TIM to a directory and have it automatically sort files and clear dupes on just one scan. I use clrmame for everything else outside of that. If I could do that with Good, non-good, no-intro, MAME, Artwork, Video or any other non-tosec dat with clrmame it would be awesome. I know that you can't cater to all of our needs and requests though :)


> Well yes, the source will be scanned per profile but scanning time is usually
> not a thing to care about if the files are packed. Either do it that way or
> merge dats.
>
>
> Roman Scherzer
>

well...the idea is nice but it still needs major core changes (since you don't have one dat in memory but multiple). Theoretically an internal datfile merge would do what you want ...or just a list of databases...not as evil as I first thought but you still need additional work in thinking about per-database paths or other handling....and I simply don't have the time at the moment.

Not to mention that holding 1247141 tosec datfiles in memory will be awful ;)


Roman Scherzer

Yeah I'm sure it wouldn't exactly be too friendly to your system memory, but these days I see firefox taking up 500+ megs of memory. Although, I really have no idea how much memory would actually be sucked up if you loaded 200 dats at the same time. Would it be possible to compress the merged dats in memory?

> Would it be possible to compress the merged dats in memory?

well...on my list is a reduce-database-memory item for a long time....since I didn't care that much about resources (and MAME is still rather small) I didn't touch it. Of course there are several possible methods to reduce the memusage...(e.g. by not keeping the full db in memory...or reducing / compressing the data) but of course this has an impact on scanning/rebuilding speed.



Roman Scherzer