EMULAB Forum

Please login or register.

Login with username, password and session length
Advanced search  

News:

The new forum is online, hope you enjoy it!

Pages: 1 2 3 4 [5] 6 7 8 9 10
 41 
 on: 30 March 2024, 01:29 
Started by Floob - Last post by Floob
I'm sure I'm missing something obvious here, but struggling to use the rebuilder tool as I keep getting "no match" in the trace logs.

I have produced an xml file using this file:
https://github.com/mamedev/mame/releases/download/mame0139/mame0139b_64bit.exe
with the command mame64.exe -listxml >139b.xml

This seems to generate a valid xml
I run this command where the input folder has 0.138 verified romsets (approx 8,650 zip files - 24GB)

rebuilder.exe -x 139b.xml -i D:\input -o D:\test -l trace

Code: [Select]
[2024-03-30 01:00:09.496] [info] loading: 139b.xml
[2024-03-30 01:00:09.584] [info] building split view, rebuilder map and destination patterns
[2024-03-30 01:00:09.586] [info] rebuild start
[2024-03-30 01:00:09.586] [info] analyzing rebuilder source: D:\input, please be patient
......
[2024-03-30 01:00:09.587] [trace] D:\input\galaxian.zip no match
......
[2024-03-30 01:00:09.590] [info] analyzing rebuilder destination: D:\test, please be patient
[2024-03-30 01:00:09.591] [info] optimizing
[2024-03-30 01:00:09.591] [info] 0 unique matches found
[2024-03-30 01:00:09.591] [info] 0 archives with 0 matches can be fully copied
[2024-03-30 01:00:09.592] [info] 0 roms, 0 chds will be rebuilt
[2024-03-30 01:00:09.592] [info] 0 files already exist; 0 files will be replaced
[2024-03-30 01:00:09.593] [info] finalizing and cleaning up
[2024-03-30 01:00:09.593] [info] rebuild end, took 0.007286 seconds

And it does that for all the romsets leaving the destination directory empty. Surely it should find some matches.
I can use the rebuilder in clrmamepro without issue, but I'm missing something obvious here?

 42 
 on: 11 March 2024, 19:39 
Started by john iv - Last post by john iv
Thanks for the notes.  I just let it run over the rest of the CHDs normally and only found one decompression error on "MAME\ROMs\kof98um\kof98um_v1_00.chd".  Replaced it and things were fine.  Cheers -

 43 
 on: 11 March 2024, 09:01 
Started by john iv - Last post by Roman
well...yes, it could be that running multiple chdman instances gives a slight boost....but in old cmpro I don't think I will add it.
The new scanner runs multiple threads for scanning sets anyhow but currently it only reads chd headers and does not perform a deep chdman based verify check.
Didn't have much time in January/February to look at the new scanner though....but it's alive and kicking....

 44 
 on: 10 March 2024, 22:38 
Started by john iv - Last post by john iv
Ah, too bad.  I thought it might have the ability to kick off X separate versions of the chdman.exe verify based on cores and collect the results.

 45 
 on: 10 March 2024, 20:54 
Started by john iv - Last post by Roman
Cmpro simply runs chdman and does nothing by itself…..

 46 
 on: 10 March 2024, 20:05 
Started by john iv - Last post by john iv
Hi Roman, is there a way to have cmpro use as many CPU cores as possible to run through its deep -verify check on a CHD scan?  I'd still like to finish off checking my collection (made it to the guitar freaks last time) but it seems to do one at a time and take a good amount of time.  I thought I saw an option for cores somewhere in my wanderings of your settings.  Thanks -

 47 
 on: 27 February 2024, 17:21 
Started by john iv - Last post by john iv
Thanks for the info; I like this storage technique much more, w/ dedicated subdirectories.  Very tidy. :)

 48 
 on: 27 February 2024, 07:01 
Started by john iv - Last post by Roman
The only difference is that (when hash collision name is turned on) all rom files (when using full merged sets) are stored in archive subfolders (named after the clones).
The parent files are on root level, the clones separated in subfolders.
So your puckman.zip holds the puckman rom files in its root...plus a folder for each clone (which then contain the belonging clone files).

If turned off, this will automatically only be used in cases where real hash collisions (identical name/different hashes within a parent/clone relationship) appear. So overwriting is prevented.
Yep, the naming of this option might be a bit misleading. You can define the storage method (subfolder etc) in Profiler Options by the way....

By the way, this storing mode does not work for decompressed sets, since MAME's loader, which looks for hashes, doesn't walk into subfolders.

The new rebuilder and new scanner use this storing method by default (actually there is no option to switch) for merge and standalone mode.

 49 
 on: 27 February 2024, 00:19 
Started by john iv - Last post by john iv
Hi Roman, troubleshooting all the CHD delta stuff and unavailables after audit recently, I switched from my long standing 'Normal Mode' to 'Hash Collision Name' for the Full Merge Mode to see if that would help on some rebuilds.  What's the difference and is it rare to use that mode vs 'Normal' or should I?

 50 
 on: 23 February 2024, 17:57 
Started by john iv - Last post by john iv
ok...all baddump chds:

acorn/bbc.cpp:Torch CH240[torchh]: torch_utilities
amiga/cubo.cpp:Candy Puzzle (v1.0)[cndypuzl]: cndypuzl
..

Thanks Roman, Windyfairy got back to me about this after I mentioned it in a private message at Bannister's.  It's an issue they and VAS are now aware of.
Here's the pull request: https://github.com/mamedev/mame/pull/12057

Pages: 1 2 3 4 [5] 6 7 8 9 10

Page created in 0.245 seconds with 17 queries.

anything
anything