PAL releases of .hack games are intermittently unstable. about open-ps2-loader HOT 28 CLOSED

Actually, I think it ended up being better that you gave me those files. Thank you.
They seemed to have changed compiler settings or something, and two of the later games had a rather different layout. So the patch would not have worked.

In the end, I just hardcoded the addresses of the required functions for each release. It's easier and faster than telling the EE to scan 1MB of code lol.

from open-ps2-loader.

Please wait for pull request #139 to be merged in. Some silly mistake happened.

from open-ps2-loader.

EDIT: Nevermind. I realize this is going to be more trouble than its worth (need to locate functions too).
I think I'll just make a generic patch.

As for the IGR fault, we can refer to your past comments. But I don't know how fixable it is.

from open-ps2-loader.

@aphirst if you are using the auto-bot to get the ifcaro builds, it is broken. It's not been providing accurate recent updates. This is the auto-bot I am referring to: http://www.ps2-home.com/forum/viewtopic.php?f=83&t=4301

When I released OPL 1286_DB in that same post, I also provided the most current version of the ifcaro build: http://www.ps2-home.com/forum/viewtopic.php?p=30745#p30745

So I would suggest to test that version instead and not be grabbing them from the auto-bot since it's broken for now.

Not sure what exact changes you are referring to, but they have all been added to the Daily Builds except for the most recent changes ( #133 and #134 ) that were approved in the past 24 hrs. -- I am moments away from pushing a new release that addressed those new changes into OPL, so wait for that.

So in the future, please provide a link to the commit versions, so I know what you are referring to. This would avoid confusion.

Now then, I also must let you know that there are some bugs still with VMC and also using USB HDDs with OPL.

And I think those issues are no longer relevant to this .hack games fix. So I would suggest to start a new ticket issue here at GIT about this (but if the bug is only present in the ifcaro repo -- DB bugs should not be reported under the ifcaro repo) or start a new topic in a forum (of your choosing), since right now when we are posting here -- all devs get a message -- so we are kind of forcing them (keeping them as hostages) to read every little thing we write in here -- not cool.

And once we narrow down this bug in a forum based setting, then we can perhaps see what commit caused what and then bring it to the developer's attention. That should be the way to go about this -- at least that is the protocol we've been following for several years when it comes to OPL tests. -- so if you see that I no longer reply to you here in this pull request ticket (even though you mention me), is that I am following that protocol. Meaning -- waiting until you present the new bug you are experiencing in a forum and also learn all the info we need to see what you are working with and etc.

from open-ps2-loader.

Alright.

1. Neither "1286 DB" nor "1162 ifcaro" (http://www.ps2-home.com/forum/viewtopic.php?p=30745#p30745) are able to load (to the Bandai logo, after the language selector) any of the PAL dot Hack games now.
2. Neither "1292 DB" nor "1166 ifcaro" (http://www.ps2-home.com/forum/viewtopic.php?p=30745#p30815) are able either.
3. The old SendSpace-hosted builds, originally uploaded by @sp193 , which showed promise in fixing the issue, are now dead links.

Whatever changes @sp193 was working on simply must not have been merged properly, and are currently unavailable for further testing.

from open-ps2-loader.

I am working on a possible patch for these games on this branch: https://github.com/sp193/Open-PS2-Loader/tree/dothack-vmc2
The patch changes the call to LoadExecPS2(), to call scePadEnd() before calling LoadExecPS2(). This should prevent any problems related to not stopping PADMAN from spamming DMA transfers from the IOP.

As of now, the patch only targets SLES-52237. The patch should not affect other games. And so at this point, we could say that the other 3 games might be working... through sheer luck.
I just looked at .hack//Mutation (SLES-52467) and it also has the same problem.

I shall update the patch.

from open-ps2-loader.

Yeah, it's strange how, despite only checking for the executable for Part 1 in your patch, things started working (booted past the language selector) in the others too.

One thing I'd considered to share, but which now I'm not sure you needed, were file lists from each of the ISOs.

Part 1: https://ptpb.pw/kmL5.txt
Part 2: https://ptpb.pw/HOGF.txt
Part 3: https://ptpb.pw/mR0w.txt
Part 4: https://ptpb.pw/WbWU.txt

Edit: Which executable is the language selector, if the SLES_*. file is the game's main ELF?

Also, after we've cleared this first issue, is it worth a few more comments on the 2nd issue I'd discovered (the "log-in screen" bug)? If so, I can do my best to summarise it, since it's still present in the build that results from your patches here.

from open-ps2-loader.

I sent you the executables. I've just noticed you update your comment, but at any rate, you have them now.

Well in that case, I'll just repeat a very brief summary, in the interest of "information locality":

Ever since ps2-home's "r797" (which apparently corresponds to this commit), VMC and IGR both being enabled causes the following bug:

• Boot the game and load a save file / start a new game
• Bring up the "The World" log-in screen (you don't need to actually log in)
• Return to the AltimitOS menu (the "game menu" while logged-out)

The trap is set: doing either of the following two actions will blackscreen the game:

• Bring that log-in screen up again, or
• Press START, and attempt to return to the Title Screen

This is particularly crippling, since the game involves you logging in and out of "The World" pretty frequently.

Disabling either IGR (Mode 6) or VMC seems to avert the issue, though it's important to note that the IGR and VMC themselves both work fine (at least, until the game is frozen).

from open-ps2-loader.

If you have a build you'd like me to test, just let me know. I can do so on a 30k and a 50k, and from 2 different HDDs via the official adapter. I can also test with a GameStar adapter later in the week, which might be interesting.

I'll bet these games had a really interesting development history, especially in the context of the now-defunct online "Frägment". They have a lot of quirks. For example, being really picky about the point in 0-255 where the analog stick "rest" position is - which causes many 3rd-party controllers, and even PCSX2's linux input plugin to make the game mostly unplayable.

from open-ps2-loader.

Here you go: https://www.sendspace.com/file/bxl7uo

from open-ps2-loader.

Here you go: https://www.sendspace.com/file/bxl7uo

These freezes are earlier than before. Before, the freeze was after the fade-out, just before the Bandai logo would have showed up.
Enabling mode 6 does not prevent these freezes.
Disabling VMC does not prevent these freezes.

from open-ps2-loader.

The updated fix was merged into master as commit ec29800.
I couldn't replicate the other glitch though. But it is possible that I did not take the exact same steps as you have described. Unfortunately, time is up for me.

from open-ps2-loader.

@sp193 Hang on a moment, did you change the patch since the version last week which only worked for parts 1 and 2 ?

from open-ps2-loader.

Yeah. I forgot that the 2 later games had a slightly different layout for one of the patch locations, which was why it did not work.
It should be fine now.

from open-ps2-loader.

Ah, I see. Cool.

@Jay-Jay-OPL: Is this change already incorporated in one of the PS2-Home daily builds?

I'll test this as soon as I can, and hopefully we can lay this principal issue finally to rest. :)

from open-ps2-loader.

@aphirst, no, I'm going to wait, because this fix caused another issue with another module (IGS). So I am going to wait until that gets fixed or a better solution is offered. #113

At the same time, there is another update to OPL to fix another game, and that one also fails with using USB HDD: #112 (comment)

So to prevent more bugs being presented in the forum because of these recent changes, I will wait until the coast is a bit clear. :)

But what I can do is provide you a test build for you to test... Just let me know here...

from open-ps2-loader.

@aphirst: you can also get a development build from @AKuHAK's bot: https://akuhak.github.io/test_build/

from open-ps2-loader.

Hmm, when I try the +VMC build from the 27th of July (45c9a78), OPL crashes while trying to load the games, i.e. while still in OPL. I'll try to fiddle around a bit.

When in the Games list, the freeze occurs at "Loading Config". If running "test" from inside a game's settings, it just freezes there.

Oddly this only seems to affect the .hack games. Wrath of Cortex, Valkyrie Profile, and We Love Kaamari can all be launched by OPL just fine, but the .hack games (PAL and NTSC) fail to be launched by OPL even past OPL itself.

Edit: The previous build from 22nd July (45c9a78 - why is this the same as the newer one?) is able to launch .hack, but blackscreens after the language screen, regardless of which part.

from open-ps2-loader.

@sp193 Are you able to provide a build which incorporates the (updated) changes you made, so that we can at least be sure whether the issue with @AKuHAK's build is a regression, or what?

from open-ps2-loader.

It's because I made a mistake while revamping system.c for the IOP-side patches. Only VMC doesn't work.

Here's the related pull request: #118

from open-ps2-loader.

Sorry about my absence. Let's see if I can get back on the ball.

First thing I did was try out "v1286" from http://www.ps2-home.com/forum/viewtopic.php?p=30745#p30745 on a fresh PS2 setup (formatted HDD, FMCB/FHDB, OPL, and re-transfered dot Hack Part 1). Then I enabled VMC by setting up one virtual memory card.

Straight away, I'm getting the black-screen-after-language-select issue again. Is this change not incorporated yet into PS2-Home's builds? Do you have a build (or builds) which I can now further test, so that this can finally be closed?

from open-ps2-loader.

Aha, I decided to also test against the most recent Akuhak build (Aug 10). This loaded the game fine! I'll test Parts 2-4 imminently. (edit: I used the -VMC build, of course, and similarly re-enabled a virtual memory card before loading the game)

@Jay-Jay-OPL: It looks like the relevant changes haven't been upstreamed into the PS2-Home builds. Once I confirm the fixes working for Parts 2-4 too, what can be done? :)

from open-ps2-loader.

Well, the Akuhak build works for the first 3 parts. For the 4th part, the game complains that I haven't got a Dualshock 2 plugged in (when I do), yet the soft-reset back to the PS2 Browser still works. (I'm testing these first without disabling IGR).

Oddly enough, the Language selector doesn't complain.

I mean... I think this controller is a Dualshock 2... I did at one point wonder whether it was actually a PS1 Dualshock. I'll work out how to tell, or, I'll find another controller to test.

Edit: Yep, it was a PS1 Dualshock. Plugging in a PS2 Dualshock 2 gets the 4th part of Dot Hack fully loaded with Akuhak's builds. Not with the PS2-Home builds, of course...

from open-ps2-loader.

@aphirst: So, is it a DS or DS2?

@Jay-Jay-OPL: The VMC-Thing is it's own (kind of) issue. ;)

from open-ps2-loader.

@TnA-Plastic It was a DS1 in the end. As I added to my previous post:

Edit: Yep, it was a PS1 Dualshock. Plugging in a PS2 Dualshock 2 gets the 4th part of Dot Hack fully loaded with Akuhak's builds. Not with the PS2-Home builds, of course...

from open-ps2-loader.

So this issue can be closed? All games are fixed? ;)

from open-ps2-loader.

@TnA-Plastic I can't confirm that until I have the chance to disseminate my findings to the PS2-Home forum too, nor until I can confirm that I'm actually testing with builds which are actually representative of the state of OPL. I'll get round to the former as soon as I can, and hopefully the latter will naturally develop as part of that process.

from open-ps2-loader.

If there is nothing further to add, we can consider this as solved.

from open-ps2-loader.