Topic: Status (Read 31844 times)

oddi · « **Reply #40 on:** 29 November 2014, 14:08 »

Roman, about rom names , i think that is not a problem - i retest all my mame sets with maually rename few roms name with random name and etc etc, reload cache database in QMC2 and used Romalayzer for checking roms and play rename game roms - all is fine, game playble, Romalayzer report - checksums fine . I see problem with names only when mame/mess collectors sync databases over internet - - trackers , ftp's and etc etc and used differents CMP versions.
About mame code - yesterday with Oxyandy we found few little bugs "maybe" , 2 or 3 duplicates chds , missing tags "merge=something" and etc.
For atm i used the last test build and think all is fine with my mame/mess roms/chds.
My opinion

oxyandy · « **Reply #41 on:** 29 November 2014, 17:10 »

Roman I have done a bit further testing...
I took the new Beta III binaries you posted,
dropped them into a fresh official 4.015 download, over writing the existing files..
Configured everything from scratch.
All seems to be working great

Nothing to report
Rest well

NOTE:
Others testing Roman's latest beta should do the same (as above) dropping into an existing pre-configured CMP folder is not advised.
I'm sure feedback is welcomed, how's does it work for you ?

Oddi, yes the DAT we were looking at, has some missing <merge tags> which could prevent duplicate chds being created.
Not cmp related, hopefully a future MAME release has those omissions corrected.

Roman · « **Reply #42 on:** 29 November 2014, 19:41 »

with all ok you mean no slowdown? havent checked the release binaries yet and wont do before tuesday.

regarding naming mame 156 got 4 or 5 sets with some naming issues where merge tags result in double romnames which cmpro also handles with some postfix checksum. I will check if there is a better solutiion. ..... next week

oxyandy · « **Reply #43 on:** 29 November 2014, 20:01 »

Quote

with all ok you mean no slowdown?

Correct, no slowdown

Roman · « **Reply #44 on:** 02 December 2014, 08:55 »

Interesting......

....and I guess I found now a proper solution to the double roms issue with lucky8a royalcdfr royalcrdd royalcrdg and that funky set....

oxyandy · « **Reply #45 on:** 04 December 2014, 05:08 »

I saw your post yesterday, then it's edit..
Is there anything you want tested / feedback with ?
Let us know.

Roman · « **Reply #46 on:** 04 December 2014, 06:18 »

well...I found a way how the double roms. of 156 are read in correctly so you dont need the crc poststring....currently I got two more things on my todo list which I need to get sorted before we come closer to a (test) release....I will shout if I got something....

Roman · « **Reply #47 on:** 04 December 2014, 22:50 »

http://mamedev.emulab.it/clrmamepro/binaries/cmp20141204.rar

well...only change is the new double-rom handling (lucky8/lucky8a, royalcdfr/royalcrd/royalcrdd/royalcrdg, funkyfig/funkyfiga in MAME .156)....the double roms (based on some weird merge tags) don't really need a rename (crc poststring) since in split mode, you won't be able to create overwrites and in full merge mode, the new hash collision naming will solve the problem...

oddi · « **Reply #48 on:** 05 December 2014, 06:15 »

Roman, u are Master !!!

, test this build - fix problem with "evil" funky and "royal" sets and etc etc.
Have one question :
"Normal mode" and "Hash collision names"
Normal mode - CMP add [set name] only if have collision.
Hash collision name - CMP add [set name ] to all merge sets with or without collision
thats right ?
and tnx

Roman · « **Reply #49 on:** 05 December 2014, 08:31 »

Well...normal mode means that you got the full merge mode as you know it...from the past...from the current version...with one exception: In case you got a clone set which got a hash collision with its parent/other clones, this set will have different rom names (the way the name is formed can be setup by yourself).
"hash collision mode" (well...I know 1 user who may need this mode ;-) ) means that you can arrange ALL sets in such a new-naming-way....so each clone can be named differently...Actually this may only be useful if we use a pattern which allows subfolder usage....."%f/%1"

oddi · « **Reply #50 on:** 05 December 2014, 09:48 »

Many tnx for answer

About variables %f [%1] - me think dont touch this parameres , good show

That is my opinion.
Btw -fast test my mame sets with "hash colision" enable - and CMP wanna rename million roms , heheh

, normal mode is good choice because CMP safe original roms names and rename only colisions.
Again very good job

Roman · « **Reply #51 on:** 05 December 2014, 13:50 »

I still wonder if it makes sense that - in case of a hashcollision - what gets renamed:

- in the wip version it's "the affected cloneset, but all rom files, no matter if only 1 had a hash collision"

you might also think of
- "only rename the affected romfile(s)" (pro: less renames, con: maybe visibility quality within a full merged set)
- "rename all roms in all clones of the current parent/clone relationship" pro: better visibility con: more renames

opinions please....

oddi · « **Reply #52 on:** 05 December 2014, 15:09 »

I thnik - rename only affected roms not sets:)
So Mame dumps will be as close to their original names, only few affected roms have differents names.

Roman · « **Reply #53 on:** 05 December 2014, 15:22 »

but it looks stupid if you use subfolders ;-)...then you got some rom files of clone A in a subfolder (the files with collision) while the rest of clone A stays in root....
...that's why currently all roms of A end in the subfolder...

...I only wonder if clone B, C ... should also get their own subfolder/naming....even if they don't have collisions...

oddi · « **Reply #54 on:** 05 December 2014, 16:22 »

No no Roman, idea with subfolders is closed! finito , fin , end

i say about rename affected roms - rom.bin [ set name ] , this idea i like it when cmp rename only affected roms.

Now CMP the last beta work perfect, CMP rename only affected roms and dont touch others

Roman · « **Reply #55 on:** 05 December 2014, 16:49 »

no...currently it renames ALL rom files of the cloneset which got at least one hash conflict...not only the romfile which got the conflict...

oddi · « **Reply #56 on:** 05 December 2014, 17:29 »

ok boss, i agree with u and wait final version

Roman · « **Reply #57 on:** 06 December 2014, 20:43 »

found a little hickup regarding the hash collision change with nodumps though. (nspirit)...fixed now....

...and oops..I was wrong...currently it already renames all roms in all clones of affected sets... (and not only the clone which got a collision)

...and added a check for "release" element users...(if you don't know what it is...don't worry...)

http://mamedev.emulab.it/clrmamepro/binaries/cmp20141206.rar

oddi · « **Reply #58 on:** 07 December 2014, 02:20 »

Fast testing my suffered full mame set , ~~i do not see anything for notes~~ , ~~whole is perfect~~

Bug - with every start and scan CMP wanna rename royal card set roms. attach set

Roman · « **Reply #59 on:** 07 December 2014, 09:04 »

hmm...yes...have to check that again....*sigh*

News:

Author Topic: Status (Read 31844 times)