Jump to content
LaunchBox Community Forums

GreatStone

Members
  • Posts

    37
  • Joined

  • Last visited

Everything posted by GreatStone

  1. @JoeViking245 Challenge or not, ikaruga is now part of hi2txt
  2. @BothEyesShut As the author of hi2txt, how to not appreciate such kind words ? And having @JoeViking245 tagging me here sounds like a challenge to support ikaruga: so, I am able to launch the game with mame, it is very very (very) slow, but I managed to make a hiscore and the nvram/sram content generated seems to be easy to decode! Is it playable for somebody on a powerful machine? Does it worth to support it in hi2txt with such mame status?
  3. @filarenas3 @JoeViking245 @Jason Carr After a lot of trials, it appears that recent mame versions have modified the dumped content related to "special items counters", for whatever reason. I have slightly modified the 'bublbobl.xml' file (describing how to decode the game hiscores .hi file) to fix it, as it is sharing the same XML than 'bublboblr' It will be part of the next release. I also attached the new file here in case you want to have a working version right now (LaunchBox should have a hi2txt.zip file somewhere, where all xml files are stored). Thanks, @filarenas3, for reporting the issue bublbobl.xml
  4. @JoeViking245 I will have a look at this bubkboblr" specifically: thanks for the issue report In the last versions of hi2txt, I have added some checks to try to detect when nvram are full of scrap (and it is happening sometimes with mame) to avoid sending irrealistic score to the globbal ladder.
  5. @Snakeystew Do you know which exact LaunchBox version have been used? I am curious because the last beta is embedding the last hi2txt release that should forbid this behavior.
  6. @Jason Carr A generic algorithm remains possible: when you parse hi2txt output, if SCORE+NAME columns are detected (and user configured one or more preferred NAME), take NAME into account. If only SCORE is detected, upload SCORE as it is implemented now, not taking into account any user configuration.
  7. @Jason Carr I let you decide, obviously. Personally, I am more frustrated by a garbage score coming in the community table, than having to play again to beat my score: I find it more "fair" and "enjoyable" for the global community (better to miss something than corrupt the scores), as we see that emulation or hiscore plugin are not 100% accurate and will never be.
  8. @Underoath13 Bubble Booble: Sometimes, some mame games doesn't record anything into the nvram until a hiscores is done So, launching hi2txt before the session gives you nothing or all scores with value 0, then after the session, a hiscores table full of scores != 0 ("default" one after 1 play) + the new hiscore of the player in the middle => nothing I can do here and surely, not a big issue as default scores are rarely very high @Ateb Blazing star: The log shows no hiscores saved by mame It happens: mame emulation and/or hiscore.dat saving system are not 100% accurate. So, no new score hi2txt can decode here, only default one. => mame issue for this game, except if you have desactivated nvram saving in some mame ini files (ex: mame.ini => nvram_save 1, nvram_directory nvram) => nothing I can do @Elhora @Ateb Andro Dunos, Metal Slug, Gradius, Blazing star, airass: no updates of score sometimes? It happens: mame emulation and/or hiscore.dat saving system are not 100% accurate. => if LB log files shows decoded scores; but not your last hiscores inside, mame was not able to save it. => Nothing hi2txt can read here, if the nvram/hi files are not updated by mame or hiscore plugin. @Elhora altair: no updates of score sometimes? Seeing the log... TRACE: data taken from file: G:\LAUNCHBOX\Emulators\Mameui64\nvram\altair\nvram TRACE: structure simple elt: UNKNOWN (226 bytes read on 256) ERROR: Wrong SCORE value (0xF212) encoding detected: it can be due to a temporary corrupted .hi or nvram file 2020-05-05 07:45:33 PM MameHighScores.Parse: No new scores found to upload! ...the new "corrupted data detection system" is working and has been able to skip a corrupted nvram/hi files. It happens: mame emulation and/or hiscore.dat saving system are not 100% accurate and sometimes record garbage. => except detecting garbage and preventing upload of them, nothing hi2txt can do here. => And that's a good news to avoid uploading garbage scores @BlackKnight1965 any mame derivative (like GroovyMame) is supported by hi2txt as soon as the hiscores files are saved in hi/ and nvram/ subdirectories. I will let Jason comments on LB side. @Elhora fireshrk Seeing the log... 2020-05-06 09:44:45 PM RomPath after M3U build: G:\LAUNCHBOX\Emulators\Mameui64\Roms\fireshrk.zip 2020-05-06 09:48:08 PM Aborting MameHighScores.Parse: No high score or nvram files found. ...shows no hiscores saved by mame It happens: mame emulation and/or hiscore.dat saving system are not 100% accurate. So, no new score hi2txt can decode here, only default one. => mame or hiscore plugin issue for this game => nothing hi2txt can do here @Jason Carr LB Improvement request (important one): log also hi2txt traces when you call it before the play session, so that we can better understand the potential root cause (mame, hi2txt, LB) LB Improvement request (less important, but for robustness and less database maintenance): let the user specify the mame NAME (3 letters) he will enter (or more than one, why not) everytimes he made a hiscores and LB should only deal with lines having such NAME. it will fix a lot of strange situations. Ouf... What a debugging marathon this night
  9. @sundogak sinistar: at the star: "sinimmortals" hiscores (=alltime hiscores): all set to 0 by the game itself "today's" hiscores: not 0, but not saved by mame then, when player is making a hiscores: "sinimmortals" hiscores: updated correctly, decoded correctly on my side "today's" hiscores: not saved by mame => I do not see any glitch here, difficult to say more without seeing a log with trace ON as well as the nvram file @Cronos74 samsho: unsupported game, never decoded the nvram file for it or its clones. => expected behavior to not find any "XML description" (which means no XML file provided by hi2txt explaining how to decode the nvram file saved by mame) @Cronos74 afighter: score is stored in nvram AND .hi files, on 4 bytes each nibble must be a number [0-9], so max score is 99 999 999. last release of hi2txt (1.12) is detecting, in this case, if any nibble is different than [0-9] (ex: corrupted nvram with some 0xFF, leading to a very very big score), and do not display anything so, theorically, it cannot be 2 704 151 200, in any case. => are you sure that the last hi2txt release is used? => difficult to say more without seeing a log with trace ON as well as the nvram+.hi files @Jason Carr Just for you information.
  10. @Jason Carr When do you need the new hi2txt build (fixing issues detected in the past days by users and preventing uploading wrecked scores) for the integration into LaunchBox? The more time I have, the more fix I put inside, but it is already better than the current public one.
  11. @Jason Carr I recommend to use "-trace -r ..." instead of "-rd ...": debug mode is adding traces AND additional fields found in the .hi file that can potentially break your parsing. -trace is only activating the traces. My bad...
  12. @Jason Carr Is there a way for @nicr4wks to modify the way hi2txt is launched to call it with "-rd" argument instead of "-r"? It will display a lot of debug messages allowing us to understand what is going on. Also, how can we see in the log how hi2txt is called, to be sure?
  13. @Blitz Mary I am nearly happy to see this A good testcase with all needed info, and a way to test the coming hi2txt release: when using my new hi2txt local build with this .hi file, I am able to catch the crap content and reject the file, preventing any upload to LaunchBox server. > hi2txt.exe -r C:\Downloads\pacland.hi ERROR: Wrong SCORE value encoding detected: it can be due to a temporary corrupted .hi or nvram file @Jason Carr Golden Axe itself separate the 3 scores: even if you do multiple very high scores with Tyris, and one low score with Ax, the game will keep 1 for Ax, 1 for Tyris (and 1 for Gilius). There is no sense to combine them in one table. Perhaps one character is easier to play with, some games have tables per difficulties, etc. It make sense to fight for hiscores on each separate table, IMHO
  14. @Elhora inside LaunchBox installation, there is a hi2txt.zip file: overwrite the existing xml with this one
  15. @Elhora Golden Axe is special: currently, the "score" is displayed as a "strength". consequently, launchbox cannot detect this as a score. I have modified the xml to display also Tyris and Gilius strength, and renamed the column SCORE. As soon as LaunchBox will be able to handle more than one hiscores table per game (@Jason Carr?) (3 tables for Golden Axe, one per character), it will be possible to be loaded. @andyco47 rallyx and nrallyx xml have been improved so that a "SCORE" header is now available for LaunchBox to read goldnaxe.xml rallyx.xml
  16. @Blitz Mary deleting .hi file yes. Other users seem to have been able to save their scores, so it should fix. deleting hiscore.dat, no, as it will also prevent the hiscore plugin to save any score at all... for some rare situation, hiscore.dat should be perhaps fixed for the fshark entry by the hiscore.dat maintainer, I don't know.
  17. @Blitz Mary OK, so it is clear now that the issue is coming from mame itself or most probably hiscore.dat entry for fshark: sometimes, the entry in hiscore.dat cannot accurately describe when and what to store in the .hi file, in rare situation. Deleting .hi file is the only way to solve it. Perhaps hiscore.dat maintainer can do something, but it is outside of my own scope. Anyway, thanks a lot for your reactivity to help me understand what was going on. @Jason Carr How can we prevent to send such garbage to global community ladder? Do you have something in mind I can do? The only potential solution I see currently is trying to add a new feature to give max value for score or area here (inside the hi2txt xml describing how to decode the .hi file), so that if one decoded value > max, then we decide that the .hi file is corrupted and do not display anything?
  18. @Blitz Mary Thanks for the .hi file. Now, what I am seeing here is a completely corrupted .hi file content, explaining why hi2txt is extracting that... When you start again mame with this game, is the ingame hiscores table is showing the right scores on screen or the same garbage also?
  19. @Blitz Mary Having the log greatly help! May I request also the .hi file (located in <mame_install_folder>\hi\fshark.hi) ? I am very curious to see its content...
  20. @Blitz Mary I tried today, 2 times, with mame 0.219 and 0.220, reaching 108 000 and hiscores are decoded well for me. Can you provide more info, like LaunchBox debug log & exact mame version used?
  21. Ok, so I finally found the source of the "151515..." scores. For games relying on nvram, depending on when you go out of mame, the content of the file can be full of 0xFF (nvram not yet written by the board?), or partially plagued with such 0xFF (nvram under writing?): centiped has given me a lot of different nvram contents, even without achieving a hiscore. I am implementing a generic code (whatever is the game) to detect this situation as much as possible. I am still thinking about what to do after: replace such score by "0" instead? do not display anything at all if one score is like that? the 2nd solution could be also a way to deal with the more problematic case of NAMEs as 0xFF can be a possible encoding sometimes for a letter. Perhaps safer to skip the whole file even if one of your valid hiscores will be missed: at least the central DB will be cleaner. What do you think? @Jason Carr
  22. 3stooges is fixed: 22 entries were read instead of only 21, leading to a score of FFFF, converted into this strange number 3stooges.xml
  23. Duplicating here my reply about "1515151515151515" score from the original big thread: @metalr About the games reporting "1515151515151515" as a top score, can anyone give me a logfile from LaunchBox showing that it is what is displayed by hi2txt itself? LaunchBox is logging hi2txt output and I would like to see one of them with this score... to be sure it is coming form here. Currently, my best guess is: program finds a nvam/hi file, existing but locked or empty, so that a read() will return "-1", converted into 0xFFFFFFFF, converted again into decimal as "1515151515151515", letters by letters. So, a read file error. I love such mystery
  24. @Jason Carr @flytechguy @Underoath13 About the games reporting "1515151515151515" as a top score, can anyone give me a logfile from LaunchBox showing that it is what is displayed by hi2txt itself? LaunchBox is logging hi2txt output and I would like to see one of them with this score... to be sure it is coming form here. Currently, my best guess is: program finds a nvam/hi file, existing but locked or empty, so that a read() will return "-1", converted into 0xFFFFFFFF, converted again into decimal as "1515151515151515", letters by letters. So, a read file error. I love such mystery
  25. @MustardCowboy Zaxxon scores are fixed with the attached xml file to put inside the hi2txt.zip file burried into LaunchBox. zaxxon.xml
×
×
  • Create New...