EMULAB Forum
clrmamepro [English] => clrmame Discussion => Topic started by: remax on 28 August 2012, 14:13
-
For a few of my Mame and Mess CHD, i have some strange reactions of CMP :
The CHD are good (good SHA1 when compressed and good Data-Sha1 when uncompressed, -verify is ok), but CMP keep asking me
Set: NFL Blitz (boot ROM 1.2)
Name: blitz
File: E:\UME\roms\blitz\blitz.chd
Do you want to add the missing CHD?
(blitz.chd is already in this directory), and weither i answer yes or no, i always have the same reaction :
NFL Blitz (boot ROM 1.2) [folder: blitz - size: 544kb]
missing but fixable (-> blitz) chd: blitz.chd [chd-sha1: 9131c7888e89b3c172780156ed3fe1fe46f78b0a]
I don't really understand what is happening. I tried to recompress the CHD (and the others that has this problem) with freshly compiled CHDman, but without success.
-
Clean the profiler cache and do a new scan. Turn off virusscanners for rompaths in question...and be sure to use an official mame with direct data import in the profiler (i.e. no 3rd party datfile).
-
Clean the profiler cache and do a new scan. Turn off virusscanners for rompaths in question...and be sure to use an official mame with direct data import in the profiler (i.e. no 3rd party datfile).
I forgot to say that it is the case since at least three months. Since, i even tried a fresh install of CMP without any success. I use my own freshly compiled SVN releases from the SVN server (you can hardly use a more official build), or official "uX" releases but it don't change a thing.
-
You can send me your cmpro.ini file and the belonging *.cmp file from cmpro's settings folder. But first try an official MAME binary and turn off 3rd party software which might interfear disk access and run a new scan.
-
OK
-
I'm going crazy lol...
First of all, to be clear, before posting here i had 2 problems : the one described on this topic and another : samples are randomly marked as unneeded. When i say randomly it is as this : Everytime i update the binary (or create another clean profile), it choose one or two set of samples and decide one for all that it is unneeded (so ask for erasing). Then with another binary (official or compiled by myself), it will choose others samples and perhaps the firsts ones will be now accepted without problem.
It's unlikely that these samples are really unneeded as they are marked then as missing with exactly the same names inside the archive, and as far as i know, samples are only checked by their names, not their CRC.
Back to my initial problem. I rebuilded my whole MAME set. First with the 0.146 binary on the Mame site then with the 0.146u5 found on Mameworld (as the official site don't release ux binaries, there is hardly a more official binary). At the end, no more message like the one i quoted above (about CHD)
Then i tryed with my compiled latest SVN binary and the other problem came back : some of the samples are randomly erased.
I thought that was the fault of my binary, so triied to go back to the last binary i hadn't problem with but the problem stayed, even with the official MAME binary.
Of course, when i tryed to come back to the older binaries, i emptied the profile cache, i manually erased the content of the cache folder (certainly redundant but i wanted to be sure), i erased the cmp and dat file, i created a new profile (and my antivirus is disabled and i don't have any software that work on this (or other) folder).
I didn't tried to rebuild my set again, it's very long (~24h to do actual set->0.146->0.146u5) and i'm sure there must be a way to solve the problem without having to.
In a way, i'm not suprised the first problem (with CHD) disappeared as it was appearing very randomly too (on a file that had no reason to), even if it was staying forever since it appeared.
That's probably something on my side, but i can't figure what...
-
Random file access errors sound a little bit like some installed program (like virusscanners, firewalls, zip plugins etc, sata drivers.....scanning over network drives via samba) do interfear somehow.... Check if you got something of them running in the background and for a test, turn them off.
Rebuilding is fast when you turn off recompression.
Yes, samples are only checked for name/case/exist....There are a lot of nodump samples in the official MAME binary (most likely handled as unneeded) so you will get differerent effects if you scan e.g. with the Progretto Samples datfile and with MAME.
-
Well i rebuilded again, test against the new set then came back to my old set, and now all errors are gone.
That doesn't explain how my SVN build make this errors happens to a point that you need to rebuild a set to clean the errors.
It's not about installed programs, as i keep them to the minimum on my PC and try to keep control. Moreover, the errors are not so randoms : they are randoms at first, but then stay consistent for months (it's always the same files reported when these errors start to happen).
Guess i'll have to do the same with my MESS set (some PSX chd are reported the same way if i do a full scan with a list imported from MESS, whereas nothing get reported if i use software list to scan)...
A bit annoying to not really understand what happens here, as i will be harmless if these come back...
Thank you anyway ;)
-
you might want to send me "your" mame binary which caused the issues and the "official" one which works fine....
...or you diff the -listxml output of both....
-
Yay... problem is back again, updating with the (totally official) binary from the Mame site...
Some of the samples are erased once again, and one chd as the "missing but fixable" bug.
I guess that rebuilding again will fix it, but i am not going to do that everytime i update my set :(
What files can i send you?
-
you don't need to rebuild every time....and besides, I had no probs with .147. (there are no chd changes compared to 146u5 anyway and samples related, there is one new set iirc.
There's something on your system which is interfearing the process...so that the first attemp to read the chd and checking the existance fails, while a later check for fixable missing ones succeeds....as mentioned before, you should exclude any mame roms/samples related folder from your virusscanner software (or use a better one :)) since such random effects are usually caused by such software
-
I scan with AV disabled...
Just a last thing and i stop bothering you with this problem...
Another strange behavior which might be related : when i do "new scan" with samples unticked (but with the samples folder set) it still ask me to delete the samples files it want to delete when i do a scan with everything ticked.
By the way, congrats for your second child ;)
-
Well, be aware of the correct error message...regarding the samples, it most likely asks you about removing unneeded sets (not samples). So an unticked samples checkbox doesn't affect this.
I will double check the chd exists/missing but fixable checks...maybe there is some weird ordering effect which causes this.
Regarding the samples (or maybe sets) messages, can you give me a detailed example with filenames please?
-
Yeah, of course it see them as unneeded files, but as they are in the sample folder and not in the rom folder, it shouldn't even see them.
Here is two snapshots with the mentionned problems :
(http://imageshack.us/a/img543/5772/unneeded.png)
(http://imageshack.us/a/img717/6742/unneeded2.png)
my config :
(http://imageshack.us/a/img191/9413/configzk.png)
If you need anything else, just ask ;)
-
well, sure it should see them since the "unneeded sets" check checks rom and sample paths for unneeded entries....but actually, amora is needed. I see that you're using 7z files, can you send me that amora.7z file please?
By the way...you don't have all sets enabled. I can see that in the title bar of the scanner. Seems like you got some sets disabled and actually that can cause the sample issues...e.g. if you disabled the armora etc sets.
Go to "set information" (Scanner tree window button bottom left) and let me know what settings you got there....and maybe you should enable all sets there...
-
well, sure it should see them since the "unneeded sets" check checks rom and sample paths for unneeded entries...
Sounds a bit weird for me... If sample is unticked, i can't see why it should have a look into the sample folder. But you must have your reasons and it is not the problem here ;)
I see that you're using 7z files, can you send me that amora.7z file please?
I hope posting samples here isn't a problem. It's not as problematic as posting roms but feel free to erase it as soon as you wish
http://www.sendspace.com/file/ua9461 (http://www.sendspace.com/file/ua9461)
N.B : Just notice that these files are not a problem when i check samples using the AntoPISA dat, so i doubt it's anything compression related.
By the way...you don't have all sets enabled. I can see that in the title bar of the scanner. Seems like you got some sets disabled and actually that can cause the sample issues...e.g. if you disabled the armora etc sets.
Go to "set information" (Scanner tree window button bottom left) and let me know what settings you got there....and maybe you should enable all sets there...
Well i have the "CD-I CD-Rom" List set disabled in "Systems...", but as far as i can tell, everything is checked in "set information". I used "select all" and rescan : same problem. I checked CDI set and rescan, same problem (and no more message about sets in the task bar, that means that all sets are selected i think).
-
A set is a collection of roms and/or chds and/or samples...so the unneeded sets check checks both. The "samples" checkbox in combination with "name" or "case" will then check the files within the sample sets.
Thanks for the files, I will check them later.
-
thanks....I can repeat the unneeded sample issue...funnily it seems to be related to software lists...
and actually I fixed it now :)
Now I will have a look at the missing but fixable chd one....
also fixed :)
-
Great!!! Thank You very much. At least, that wasn't me being mad :D
How can we DL the update? Via auto update or do we need to compile a SVN? (i have a Mingw environment setup for Mame and QMC2).
-
look out for a new release around this weekend on the main page....the build in update function can be used to update your release then....
-
Fine ;)
have a good day.
-
Well. I confirm that totally fixed it... in Mame.
I still have the same pb with CHD when scanning MESS roms. Don't know if you have specific code for each "driver" but, if it's the case, you'll probably have to do the same fix there ;)
(And yes, i cleared profile cache before scanning).
-
The problem is rompath ordering related, not anything driver related...For which chd do you get the problem?
-
The CHD from madden96, rockman8 and traidp
All 3 from [SOFT] Sony Playstation CD-Rom List
Note that these CHD have no problems when only scanning the softlist using the psx.xml from the hash folder.
I only have problem with these 3 CHD when scanning using the mess executable
-
(http://imageshack.us/a/img10/8808/chdf.png)
Note that, unlike MAME, i use a non standard, but correct i think, configuration.
I store the bios files in the bios directory
I store the software list in a folder called software
Hence, clrmame detects sets in two places and reports it, but it doesn't bother me.
And i can't check the unneeded tick as if i do that, it will try to move everything from software folder into bios folder and would want to erase every softlist that exist as .xml and not in the MESS exe.
NB i unchecked from "Sytems..." some of the heavier softlists (Megacd, pcecd, etc...) that i don't have.
-
Regarding the "set exists in various rompaths" message, it seems that you e.g. keep the set "c64_cs" in bios as 7z and as zip..guess one can be removed. The other ones seem a bit odd to me...you go e.g. apple2gs.7z as bios (which seems to be fine) and the set as folder in e:\ume\software\apple2gs.
Seems like you rompath setup is a bit wrong...I'd assume that the apple2gs software list got an own rompath assigned (in settings and in systems)...
A clean softwarelist/bios splitting via sysdefpaths should also prevent you from your unneeded problem where sets are moved to a place you don't want them to be moved.
Regarding the chd missing but fixable ...I will have another look soon.
-
Wow wow wow... this one is totally for me!
When i first tested this configuration, i tryed to put E:\ume\bios and E:\ume\software in the roms folder list. This was obviously incorrect, but i forgot it and was lost in between the long list of e:\ume\software\nameofthelist that were automatically added when they are set in "Systems".
And i was so convinced that the behaviour i described in my earlier post was the natural consequence of my non standard setting that i didn't even tried to double check my list.
Now that i removed this entry, it's really better. There are still 5 or 6 conflicts, like sufami for snes, at the same time a bios and a member of the softlist, but i don't think clrmame can do anything about it.
Of course, i didn't fixed the CHD folder ;)
(http://imageshack.us/a/img24/3908/chdu.png)
-
hmm...ok I will look at the "is bios and also part of the software list" thingie...can you send me some example files for that?
...and of course I will also check the chd one....
-
So the good news, as i said in my Mp is we can say that i have fixed the duplication problem.
The bad news is : i DL the cdi softlist and activated both in mame and mess and gave me an orgy of the CHD bug
http://imageshack.us/a/img853/7021/sanstitre6iu.png (http://imageshack.us/a/img853/7021/sanstitre6iu.png)
I think it's all the same thing a problem with CHD priority in softlists.
(I didn't expanded item to see the whole list, but believe me, each of these bugs are the same missing but fixable bug).
-
Try this
http://mamedev.emulab.it/clrmamepro/binaries/cmpro20120928.rar (http://mamedev.emulab.it/clrmamepro/binaries/cmpro20120928.rar)
-
We're nearly there... All the "missing but fixable" bugs have disappeared from every place.
... But there is a new small problem that has appeared : A few CHDs from the softlist in MAME are tagged as unneeded whereas they are totally valid. I scan with MAME 0.147 official build.
This problem doesn't happen in a MESS scan though (with the same CHDs), and in a direct CDI SL scan
Here is a snap that show this :
(http://imageshack.us/a/img580/9297/sanstitre2ss.png)
Well perhaps i'll wait till the next MAME release to see if the problem is still there and i'll report the result to you ;)
(or it's again something on my side, anyway i'll do some tests to be sure. So, no need to hurry)
NB : You can't imagine how i would have liked to come here and just says "yeah, it totally worked". I'm sorry for that :(
-
Well perhaps a quicker way to fix it : according to the devs, there should be no cdi and cd32 SL in mame.
-
A chd is listed as unneeded if the sha1 (for older chds, md5), information in the chd header is not matching at all to the database....
so either the sha1 is not part of the -listxml / -listsoftware part or - typical for converting baddump chds to chd version v5 - the chd is bad.
Can you please send me one (or let me know where I can download one) which causes the problem for you?
-
I didn't put all the info. When the chd is erased, it ask for a chd with a specific sha1 which match the sha1 i obtain using chdman -verify.
Anyway, no problem, i'll send you a PM with one of the CHD
-
It seems that my question is going to make the SL forbiden in mame :-[
So, as i said, don't put too much of coding into it as it's perhaps going to be deprecated.
-
hmm...can't repeat this here...the chd is found and matched correctly...not handled as unneeded. maybe your cmpro.ini and belonging .cmp file from the settings folder can help....
-
http://www.sendspace.com/file/gbpbki (http://www.sendspace.com/file/gbpbki)
-
Try this
http://mamedev.emulab.it/clrmamepro/binaries/cmpro20121002.rar (http://mamedev.emulab.it/clrmamepro/binaries/cmpro20121002.rar)
-
Try this
http://mamedev.emulab.it/clrmamepro/binaries/cmpro20121002.rar (http://mamedev.emulab.it/clrmamepro/binaries/cmpro20121002.rar)
And it gave...
A TOTALLY CLEAN SCAN!
Thank you for your great work! :)
I sent you a little compensation for your hard work. Have a nice day ;)
-
thanks a lot for the donation! highly appreciated!
regarding the fix, all three issues which you reported are actually related and the current fixes are more or less workarounds. I will fi it properly next week.