News: 11 March 2016 - Forum Rules, Mobile Version
Current Moderators - DarkSol, KingMike, MathOnNapkins, Azkadellia

Author Topic: Princesstual Orgy / 1991 Du Ma Horse Racing CRC/MD5 Issues  (Read 345 times)

firedropdl

  • Jr. Member
  • **
  • Posts: 46
    • View Profile
Princesstual Orgy / 1991 Du Ma Horse Racing CRC/MD5 Issues
« on: August 17, 2017, 01:12:57 am »
Hello everyone  :)

I'm putting a list of CRC issues I've found on Translated NES and FDS games that I will eventually put up when I have figured all of it out.  Most of the issues have been because the CRC listed on the Patch pages are for headerless roms, and it looks as though a few of them might be because the CRC listed is the after-patch value.  This one is a different problem altogether.

http://www.romhacking.net/trans/202/

ROM / ISO Information:

    Tashiro Masashi no Princess ga Ippai (Japan).nes - NOINTRO
    CRC32: 1BB4E1E6
    MD5: 3E4C4B42D798135B1168CB51B822D80F
    SHA-1: 2AAA2CF1DCD998FC25D31CE00773D044EC0B7593
    SHA-256: AD9ACBD9960E5453379E9A76E79CD02249AA788FBE966EF5DB139D1C65DA84E0



All four values stated here for this game are not for this game, but they are for "Terao no Dosukoi Oozumou (J) [!].nes" in the GoodNES 3.23b. 

I was wondering if anyone from Suicidal Translations was still around, or if anybody else around here might know the true input and expected output CRC values for this translation?

The project webpage in the readme is defunct :(

Thanks!


EDITED TO ADD:

Can't figure out 1991 Du Ma Horse Racing either.  The only information on the patch page is the MD5

http://www.romhacking.net/translations/2227/

ROM / ISO Information:

    Country: China
    Rom Name: 1991 Du Ma Racing (As).nes
    File Size: 32.36 KB
    MD5 Checksum: 74967338fcfc1603c0758300445fe0d3

I've tried stripping the header as well as patching all currently known versions of this rom and I can't get that MD5 match.

However, there is a page that references that number for this game, but when you download it the MD5 is different and it's just one of the versions found in GoodNES 3.23b.  If you google the MD5 you'll get two results.  One of them is this site.  :)
« Last Edit: August 17, 2017, 02:21:54 am by firedropdl »

goldenband

  • Full Member
  • ***
  • Posts: 232
    • View Profile
Re: Princesstual Orgy / 1991 Du Ma Horse Racing CRC/MD5 Issues
« Reply #1 on: August 17, 2017, 10:41:37 am »
Yeah, I've been running into games where the CRC32 or MD5 seem to have surfaced out of thin air. In one case, the MD5 of a file matched what I had, but the CRC32 didn't -- which indicates that something may have gone very wrong when the CRC32 was calculated or input.

firedropdl

  • Jr. Member
  • **
  • Posts: 46
    • View Profile
Re: Princesstual Orgy / 1991 Du Ma Horse Racing CRC/MD5 Issues
« Reply #2 on: August 17, 2017, 12:03:38 pm »
Hey goldenband,

Yeah.  I've noticed mismatches between check values as well.  I'm documenting any errors that I find and I'll have a thread that breaks them all down.

I've actually figured out about 60-65% of the problems once I realized that some of these were giving you the value of the headerless rom.  There have also been 2 or 3 so far that were given the values of the rom after it's patched, and not before.

So far, along with the two I mentioned above, there are 3 or 4 others I just can't figure out at all including Sansuu 1 Nen - Keisan Game, Super Contra 7, and Youkai Club. 

Interestingly, the author of Super Contra 7 has a website with the patched rom that I tested that works, but any version of Super Contra 7 that I've found online doesn't match the MD5 value on his page pre-patch, post-patch, or without a header.    Also, the Sansuu 1 Nen game has a patched version I found on some really random website that is not in GoodNES or No-Intro, but it matches the check values of the author's page and when I tested it, it is translated.  (This means that this page also gives you the check values of the patched rom instead of the pre-patched rom).

So both of these I've found the final version, but I have no idea how they were patched originally.


I'll have a breakdown hopefully in a day or two with all of the ones that I figured out with the correct information as well as a section for the games that I just can't figure out.

goldenband

  • Full Member
  • ***
  • Posts: 232
    • View Profile
Re: Princesstual Orgy / 1991 Du Ma Horse Racing CRC/MD5 Issues
« Reply #3 on: August 17, 2017, 01:24:39 pm »
I'll have a breakdown hopefully in a day or two with all of the ones that I figured out with the correct information as well as a section for the games that I just can't figure out.

That'd be great -- and since I've been hoping to play the Youkai Club translation I'm particularly interested in how your research on that one comes out.

And yes, many of the checksums are for headerless ROMs -- which at least has the advantage that you can always construct a headerless ROM, while restoring an absent header isn't always possible. I remember going to patch a game, having a CRC32 mismatch, searching high and low for the correct ROM, finding it...and discovering that the difference between the two ROMs was all of one byte in the header.  :P

BTW the problem I mentioned is with Murder in the Mississippi:

http://www.romhacking.net/forum/index.php?topic=24653.msg341466#msg341466

CRC32 is bad, MD5 is valid, so how on earth was the incorrect CRC32 generated?

EDIT: Some interesting info about Youkai Club from the No-Intro site. I don't know if this is helpful, but:

Quote from: Powerpuff
HELP REQUEST : The "Impatchable" Youkai Club

I think the English Translation DAT Collection is complete with only one exception. There is one English translation I know that hasn't been already added.

It's Youkai Club, and its patch is here. The matter is that this patch seems impossible to apply!
I almost tried evrything. I applied the patch on a clean dump as said in the "Read me" file : wrong. I thought that maybe the ROM should be expanded before to be patched after. So I used KingMike's NFlate and expanded the ROM every way possible : I expanded the PRG, I expanded the CHR, I expanded both. Evrytime I applied the patch after : wrong.
:sdrop: Well, I don't know how to do.
Either I'm wrong or It's the patch itself that is wrong.

...If anyone knows how to patch this game, please tell me and it wil be added to the Dat.
Thanks.

EDIT : The "Impatchable" Youkai Club patched

The issue was far simplier than I thought.  :sdrop:
Sometimes accuracy doesn't get on well with ROM Hacking.
I used recent emulators to test the ROMs (for NES : Nestopia and FCEUX). Or another user simply applied the patch as I did fist and said it works well... with the old FCE Ultra. It only was an emulator issue. I used the old but still good VirtuaNES... and it works on it.

So obvious that I didn't think it, I already had the good result.

A new update to come soon.

FWIW I have a pre-patched ROM of the Youkai Club translation that doesn't work in Nestopia or Bizhawk (garbage screen, immediate character death). Is there an issue with the patch itself? I'd patch my own, but then we get back to the "no ROM with a matching CRC32/MD5" issue...

EDIT #2: OK, I've figured Youkai Club out. The BPS and IPS both work perfectly if applied to a headered ROM with the following hashes:

CRC32: 7fcefe0c
MD5: 0cd3eed8b490c9eede01b500580f9e23

Neither of those match the CRC32 and MD5 given on the project page for a ROM with header included.

However, the headerless version of the same ROM does match the CRC32 and MD5 on the project page -- but the patches won't work when applied to a headerless ROM, so it's kind of pointless. Still, the header itself can be restored to a headerless ROM by inserting this 16-byte hex string at the start:

4E 45 53 1A 08 04 21 40 00 00 00 00 00 00 00 00

So the question is, where the heck did those nonsense values (for a headered ROM) on the project page come from? The BPS wouldn't work, I don't think, if I didn't have exactly the right ROM to patch.  :huh:
« Last Edit: August 17, 2017, 03:35:29 pm by goldenband »

firedropdl

  • Jr. Member
  • **
  • Posts: 46
    • View Profile
Re: Princesstual Orgy / 1991 Du Ma Horse Racing CRC/MD5 Issues
« Reply #4 on: August 17, 2017, 02:34:44 pm »
That'd be great -- and since I've been hoping to play the Youkai Club translation I'm particularly interested in how your research on that one comes out.

You'll know whatever I find out.  That's one of the last few that I'm struggling with.  I've actually gone through every single NES translation now that is marked as "fully playable".  Only about 10 or so left that I still can't figure.  I'm hoping that pacnsacdave is still around here, because most of them are his translations :)  At least that makes it easy since I won't end up having to try to contact too many different people on this issue.  Even better is if I can't get a hold of him there are working versions of his games pre-patched available.  I would just like to be able to have the pre-patch info to put in this site as well is all.

Just an FYI on the off chance pacnsacdave actually reads this thread before I get all the final info together, here's the list of his games that I can't figure out.

1991 Horse Racing [NOT IN GOODNES 3.23b, 74967338fcfc1603c0758300445fe0d3 on [website withheld... google the MD5] as "1991 Du Ma Racing (As).nes", but the download is b41815bbb496c5505b3c113d5db51511.  Can't find this MD5 anywhere else.  Headerless and patched roms don't match either.]

Little Ninja Hattori [NOT IN GOODNES 3.23b, 9c00ae27de4ff78ae8675fdaa5fbf0d9 on [website withheld... google the MD5] as "Ninja Hattori Kun (J) [b1].nes", but the download is 5ba531e244ca12dc05a677b728a187c3.  Can't find this MD5 anywhere else.  Headerless and patched roms don't match either.]

Monkey King [NOT IN GOODNES 3.23b, the few versions I could find of this have MD5's of 99d210d6e6de3aa31b08a311632d86df "Monkey King (Unl).nes" and 71a4b2977f8bc8c15905749e10e38a95 "Monkey King (Ch) (Wxn).nes"]

Ninja Kid's Demon Castle Adventure [NOT IN GOODNES 3.23b.  Can't find 2F453F9834472EEDDA95C48964C3EF10 anywhere.  13994a3710af239dfea41e7abe478aeb on [website withheld... google the MD5] as "Ninja Kun - Ashura no Shou (J) [b2].nes", but the download is 4b52cd334fdcb4342625816f8bb60da5.  Can't find this MD5 anywhere else.  Headerless and patched roms don't match either.]

Q-Taro the Ghost [NOT IN GOODNES 3.23b, 9838e4ccb8406d487592545aa57700a4 on [website withheld... google the MD5] as "Obake no Q Tarou - Wanwan Panic (J).nes", but the download is d10a98c0aa8e67576f6c807ad3ff43b6.  Can't find this MD5 anywhere else.  Headerless and patched roms don't match either.]

Super Contra VII [NOT IN GOODNES 3.23b, 4b41d4e324a5bb9d7f51a6775272c6c6 on [website withheld... google the MD5] as "Super Contra 7 (Unl) [p1].nes", but the download is 4eabfa3b883817bfb609b5eb48b0e645.  Can't find this MD5 anywhere else.  Headerless and patched roms don't match either.]

I have one more thing to try on these games.  I just figured out the problem with the 1st grade and 3rd grade Keisan Games.  The given CRC32 and MD5 were for the patched rom with the header removed!!!  Not only that, but it wasn't the "[!]", it was for a Mapper 3 hack.

This might be the case with pacnsacdave's stuff here.  Quite a few of his other translations had no problems, but if this last idea doesn't work I don't know what else to do unless I can get in touch with him.


Quote
And yes, many of the checksums are for headerless ROMs -- which at least has the advantage that you can always construct a headerless ROM, while restoring an absent header isn't always possible. I remember going to patch a game, having a CRC32 mismatch, searching high and low for the correct ROM, finding it...and discovering that the difference between the two ROMs was all of one byte in the header.  :P

Yeah... not a ton of them, at least in the NES translations, but there were probably about 20-25 so maybe like 5-8% of the total ones.  Everybody makes mistakes :)

Quote
BTW the problem I mentioned is with Murder in the Mississippi:

http://www.romhacking.net/forum/index.php?topic=24653.msg341466#msg341466

CRC32 is bad, MD5 is valid, so how on earth was the incorrect CRC32 generated?

I'm not a pro at this stuff, but I think I can say with 99.99999% certainty that if you're getting the correct MD5 but the CRC32 is off that was just bad info that you were given.  I often wonder why the guys who translate and hack stuff here bother even listing anything other than the CRC32, but since there have been a few cases where they got it wrong but the MD5 and/or the SH1 were correct it was a checks and balances of sorts.

I just HATE when they don't provide the CRC32 value.  It's the easiest value to view on the fly since Windows and WinRAR easily show you it.  Viewing any of the others, at least as far as I know, requires you to unzip the file and then drag it into Hashmyflies.  That program has some buggy behavior itself which sometimes makes this process even more frustrating.


Anyway.... back to work.  Hopefully I can figure out Youkai Club for ya.  :)

August 17, 2017, 08:47:24 pm - (Auto Merged - Double Posts are not allowed before 7 days.)
So the question is, where the heck did those nonsense values (for a headered ROM) on the project page come from? The BPS wouldn't work, I don't think, if I didn't have exactly the right ROM to patch.  :huh:

I'm going to look into this more soon now that you found more info.  This was still one of the ones that I couldn't figure out.  I just went through the ones that were new since I started this work, so I had about a dozen more adds to do. 

I saw that you edited your reply hours ago but I'm in a provisional period where the moderators have to approve my post, so even though my reply just showed up recently it's been up here for hours now. 


I'm also going to look into the thread you posted about Murder in Mississippi since that was one of the handful of games that I'm still struggling with. 

After I look into those two resources you provided I'm going to put all the info together that I've been talking about.




EDIT:

Okay.  I see what you're talking about with Youkai Club now.  I didn't bother stripping the header and checking the value since the page had both headered and headerless values already. 

Either Stardust Crusaders put incorrect values for the headered version, or they had a different header on the rom when they patched it.  (The value of B292E1D5 is nowhere online... at least I exhausted all sources I'm aware of).

If you look at the page it does say the following which may explain this:  This was my first released translation. There was a minor header issue. Patch has been updated several years later with a few extra changes.


I'm not surprised that the patch won't work on the headerless rom after what Psyklax told me about them the other day.  I'm not sure why some of the translations here list the headerless values separately since most emulators today won't play them.  In this case I'm glad that they did though since we could figure out the right rom with that info.  I just hope that the different header causing the different CRC/MD5 values don't cause any issues.


EDIT 2:

I doubled checked the pre/post patches for Youkai Club and I'm 99.999999% there should be no issues at all. 

The patch rewrites bytes 06 and 07 to C1 and 80, no matter what was there before.  I assume that these were the only problem parts on the header and bytes 00 through 05 should be good since the patch doesn't do anything to them.  (I know this because I tried running the patch with all 00's in the header and the only two bytes that changed were 06 and 07). 

So assuming that 4E 45 53 1A 08 04 are good values for the first 6 bytes we should be fine.



EDIT:

You seem to have Murder on the Mississippi figured out on your own from what you said on the other thread.  Unless somebody chimes in and says otherwise, you can't possibly have a MD5 and SHA-1 match and a mismatch on the CRC32, so GAFF seems to have made a gaffe.  ;D

Actually, I double checked another very recent translation by GAFF Translations "Black Bass" that I had been having trouble with.  Same story there as well.

Both of these games state to use the verified dump of GoodNES 3.23b and the MD5/SHA-1 values match, but the CRC32 values do not. 

These will be included in the thread I'm going to make for fixes.


Cool.... 3 more figured out without having to bother somebody.  That leaves only 9 NES translations by 4 authors that I have to see if I can find help with. :)
« Last Edit: August 17, 2017, 10:45:37 pm by firedropdl »

KingMike

  • Forum Moderator
  • Hero Member
  • *****
  • Posts: 6067
  • *sigh* A changed avatar. Big deal.
    • View Profile
Re: Princesstual Orgy / 1991 Du Ma Horse Racing CRC/MD5 Issues
« Reply #5 on: August 17, 2017, 10:25:21 pm »
For the last one, the changes in bytes 06 and 07 equate to a mapper change from 66 to 140, I believe.
Two extremely similar mappers, the only difference is the mapper write address (140 places the mapper register into what is typically SRAM area).

(66 is GNROM, a standard Nintendo mapper used in such games as SMBDH. 140 seems to be called "JF-11", a Jaleco mapper used on three games, the two games in question (JF-11 Murder on the Mississippi and JF-12 Youkai Club) as well as one other (JF-14 Bio Senshi Dan).)
Quote
Sir Howard Stringer, chief executive of Sony, on Christmas sales of the PS3:
"It's a little fortuitous that the Wii is running out of hardware."

goldenband

  • Full Member
  • ***
  • Posts: 232
    • View Profile
Re: Princesstual Orgy / 1991 Du Ma Horse Racing CRC/MD5 Issues
« Reply #6 on: August 17, 2017, 10:55:37 pm »
@firedropdl: Great work -- thanks for all your efforts on this. It's truly maddening to want to try out a translation patch but be totally unable to find a compatible version of the ROM, so you're saving a lot of people a lot of headaches! :D

@KingMike: Whoa, that's very interesting that both games have the same mapper, and are 2 of only 3 to use it! That can hardly be a coincidence -- maybe the iNES headers were revised to reflect this issue? Or maybe these games were just tagged with the wrong mapper.

firedropdl

  • Jr. Member
  • **
  • Posts: 46
    • View Profile
Re: Princesstual Orgy / 1991 Du Ma Horse Racing CRC/MD5 Issues
« Reply #7 on: August 17, 2017, 11:00:33 pm »
For the last one, the changes in bytes 06 and 07 equate to a mapper change from 66 to 140, I believe.
Two extremely similar mappers, the only difference is the mapper write address (140 places the mapper register into what is typically SRAM area).

(66 is GNROM, a standard Nintendo mapper used in such games as SMBDH. 140 seems to be called "JF-11", a Jaleco mapper used on three games, the two games in question (JF-11 Murder on the Mississippi and JF-12 Youkai Club) as well as one other (JF-14 Bio Senshi Dan).)

Cool KingMike. Thanks for the info.  I'm pretty new to all of this, but I think I actually understand what you're saying there.  :)


I'm glad you replied for another reason too actually.  Your Light of Indra translation is one of the problem roms for me. 

You mention that before expanding the rom you need one with the following values:

CRC32: B2352DD2
MD5: 0E879C9A3C023132C504A010C0DEF083
SHA-1: 80DA4536C10294E83EC5EE8DD98C23F5BE7B0BB9

You then state what the values would be after expanding, but I haven't even gotten to that part because I've tried quite a few different rom versions and couldn't find one with those values. The current no-intro matches the verified dump of GoodNES.  Here's the 3 versions in GoodNES now, and their CRC values:

Indora no Hikari (J) [!]: ff03672e
Indora no Hikari (J) [hM02]: 11ec241a
Indora no Hikari (J) [hM02][o1]: ab8dedc7

Without headers, these are the values:

Indora no Hikari (J) [!]: 174f860a
Indora no Hikari (J) [hM02]: e8973b45
Indora no Hikari (J) [hM02][o1]: 15fce527


I was wondering if you could double check that they're correct values whenever you get a chance?

Thanks  ;D