[GH-ISSUE #1914] PPSSPP crashes if a newer version of a third party's software is installed #643

Closed
opened 2026-03-17 14:54:05 +03:00 by kerem · 12 comments
Owner

Originally created by @unknownbrackets on GitHub (May 25, 2013).
Original GitHub issue: https://github.com/hrydgard/ppsspp/issues/1914

This is a bug which occurs on all recent git builds.

The new ATRAC3+ filter stuff on Windows makes it crash. Telling users to uninstall software they've installed and reinstall older (possibly insecure, unsupported, or not compatible with other software they have intentionally installed) versions of that software is not a solution.

-[Unknown]

Originally created by @unknownbrackets on GitHub (May 25, 2013). Original GitHub issue: https://github.com/hrydgard/ppsspp/issues/1914 This is a bug which occurs on all recent git builds. The new ATRAC3+ filter stuff on Windows makes it crash. Telling users to uninstall software they've installed and reinstall older (possibly insecure, unsupported, or not compatible with other software they have intentionally installed) versions of that software is not a solution. -[Unknown]
kerem closed this issue 2026-03-17 14:54:10 +03:00
Author
Owner

@solarmystic commented on GitHub (May 25, 2013):

@unknownbrackets

You're referring to OpenMG right? Can't afford to be vague in your problem description.

It's unfortunate but OMG in Sonic Stage is still a dependency at the moment. (It shouldn't have to be, in an ideal situation)

<!-- gh-comment-id:18441059 --> @solarmystic commented on GitHub (May 25, 2013): @unknownbrackets You're referring to OpenMG right? Can't afford to be vague in your problem description. It's unfortunate but OMG in Sonic Stage is still a dependency at the moment. (It shouldn't have to be, in an ideal situation)
Author
Owner

@unknownbrackets commented on GitHub (May 25, 2013):

It's not clear to me if that's the only software (SonicStage also?), but definitely #997 is rife with people being told they "need to downgrade." And the guilty code has already been merged into master.

I tried a 32 bit build and it's crashing for me to. I absolutely refuse to hunt for and downgrade some third party software (that's not some beta/alpha version), or install something if it crashes without it even installed - that's ridiculous. So I guess I will only be testing my code on 64 bit from now on. That rules out any changes to jit except 64-only ones, too.

-[Unknown]

<!-- gh-comment-id:18441168 --> @unknownbrackets commented on GitHub (May 25, 2013): It's not clear to me if that's the only software (SonicStage also?), but definitely #997 is rife with people being told they "need to downgrade." And the guilty code has already been merged into master. I tried a 32 bit build and it's crashing for me to. I absolutely refuse to hunt for and downgrade some third party software (that's not some beta/alpha version), or install something if it crashes without it even installed - that's ridiculous. So I guess I will only be testing my code on 64 bit from now on. That rules out any changes to jit except 64-only ones, too. -[Unknown]
Author
Owner

@solarmystic commented on GitHub (May 25, 2013):

@unknownbrackets

The crash shouldn't be affecting the normal orphis non A/V builds, right?

I mean it's not like you've merged all of the relevant oioitff's code from his media branch into the master?

It just occured to me that the presence of SS and OMG shouldn't change anything in the normal builds.

For example, I'm using the normal version of PPSSPP obtained from orphis.

I do not have SS and OMG currently installed.

I should not get the crash.

The ones who get it are the those using oioitff's build exclusively, right?

<!-- gh-comment-id:18441464 --> @solarmystic commented on GitHub (May 25, 2013): @unknownbrackets The crash shouldn't be affecting the normal orphis non A/V builds, right? I mean it's not like you've merged all of the relevant oioitff's code from his media branch into the master? It just occured to me that the presence of SS and OMG shouldn't change anything in the normal builds. For example, I'm using the normal version of PPSSPP obtained from orphis. I do not have SS and OMG currently installed. I should not get the crash. The ones who get it are the those using oioitff's build exclusively, right?
Author
Owner

@unknownbrackets commented on GitHub (May 25, 2013):

The crash is affecting normal master builds, such as those available from Orphis' build server. I haven't merged anything except rebasing on the current master.

It only crashes in some games, but a number of them, it makes testing/development impossible and unreliable.

I'm sure it started with v0.7.6-784-gf6acdab.

-[Unknown]

<!-- gh-comment-id:18441805 --> @unknownbrackets commented on GitHub (May 25, 2013): The crash is affecting normal master builds, such as those available from Orphis' build server. I haven't merged anything except rebasing on the current master. It only crashes in some games, but a number of them, it makes testing/development impossible and unreliable. I'm sure it started with v0.7.6-784-gf6acdab. -[Unknown]
Author
Owner

@hrydgard commented on GitHub (May 25, 2013):

I'll have a look soon and disable it by default. Sorry.

Oh and we will be able to get rid of all this SonicStage nonsense in the reasonably near future. @unknownbrackets , email me if you're interested in more details.

<!-- gh-comment-id:18442476 --> @hrydgard commented on GitHub (May 25, 2013): I'll have a look soon and disable it by default. Sorry. Oh and we will be able to get rid of all this SonicStage nonsense in the reasonably near future. @unknownbrackets , email me if you're interested in more details.
Author
Owner

@solarmystic commented on GitHub (May 25, 2013):

@unknownbrackets

I see what you mean now. It seems like that AsyncStreamflt.ax file bundled with the master builds could be the cause of the crashes indeed. The folder and files started bundling with the master builds since 0.7.6-799. You were right.

I think I'm unaffected because I don't even have SS or OMG present on my system at the moment.

To wit, the file AsyncStreamflt.ax bundled with 799 and beyond is to blame. 797 is the last one not to have the file bundled together with the master.

<!-- gh-comment-id:18444257 --> @solarmystic commented on GitHub (May 25, 2013): @unknownbrackets I see what you mean now. It seems like that AsyncStreamflt.ax file bundled with the master builds could be the cause of the crashes indeed. The folder and files started bundling with the master builds since 0.7.6-799. You were right. I think I'm unaffected because I don't even have SS or OMG present on my system at the moment. To wit, the file AsyncStreamflt.ax bundled with 799 and beyond is to blame. 797 is the last one not to have the file bundled together with the master.
Author
Owner

@dbz400 commented on GitHub (May 25, 2013):

I think when the time we discuss to merge the media branch , it is pretty stable and no crash at all .Unfornaturely when the time we start to merge it , filter stuff come and it come as a merit that work in the same way via ffmpeg except we find in later stage it crash somehow when it connect to the aysnc filter .. I think oioitiff already known the issue .Until the moment we got it fixed , we can disable it for now as @hrydgard mentioned.

<!-- gh-comment-id:18444546 --> @dbz400 commented on GitHub (May 25, 2013): I think when the time we discuss to merge the media branch , it is pretty stable and no crash at all .Unfornaturely when the time we start to merge it , filter stuff come and it come as a merit that work in the same way via ffmpeg except we find in later stage it crash somehow when it connect to the aysnc filter .. I think oioitiff already known the issue .Until the moment we got it fixed , we can disable it for now as @hrydgard mentioned.
Author
Owner

@hrydgard commented on GitHub (May 25, 2013):

As a temporary fix I think you can just remove the file filters/lib/AsyncStreamflt.ax.

<!-- gh-comment-id:18444548 --> @hrydgard commented on GitHub (May 25, 2013): As a temporary fix I think you can just remove the file filters/lib/AsyncStreamflt.ax.
Author
Owner

@unknownbrackets commented on GitHub (May 25, 2013):

Thanks. Deleting isn't really a solution either since it makes the tree dirty.

Just don't like to see the wrong solution pushed hard and seemingly accepted.

-[Unknown]

<!-- gh-comment-id:18451379 --> @unknownbrackets commented on GitHub (May 25, 2013): Thanks. Deleting isn't really a solution either since it makes the tree dirty. Just don't like to see the wrong solution pushed hard and seemingly accepted. -[Unknown]
Author
Owner

@hrydgard commented on GitHub (May 25, 2013):

It's the least wrong of the Sony-stuff hacks, attempting to tide people over until the upcoming real solution arrives.

But yes, I should not have merged it and enabled it by default without more testing.

<!-- gh-comment-id:18453301 --> @hrydgard commented on GitHub (May 25, 2013): It's the least wrong of the Sony-stuff hacks, attempting to tide people over until the upcoming real solution arrives. But yes, I should not have merged it and enabled it by default without more testing.
Author
Owner

@rpgadvocate commented on GitHub (May 25, 2013):

The solution to this is to revert the dshow filter. oiotoff's previous solution worked fine for everyone and was more stable and compatible. The filter broke quite a few games that had working sound.

@hrydgard I don't know about that "real solution." If it's the one I think you're talking about, it has been "almost finished" for months, so I think it might be a hoax.

<!-- gh-comment-id:18453490 --> @rpgadvocate commented on GitHub (May 25, 2013): The solution to this is to revert the dshow filter. oiotoff's previous solution worked fine for everyone and was more stable and compatible. The filter broke quite a few games that had working sound. @hrydgard I don't know about that "real solution." If it's the one I think you're talking about, it has been "almost finished" for months, so I think it might be a hoax.
Author
Owner

@hrydgard commented on GitHub (May 25, 2013):

@rpgadvocate , I have some newer information than you about it, I know with 100% certainty that it's real and it is closer than it was a few months ago at least :)

<!-- gh-comment-id:18453577 --> @hrydgard commented on GitHub (May 25, 2013): @rpgadvocate , I have some newer information than you about it, I know with 100% certainty that it's real and it is closer than it was a few months ago at least :)
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
starred/ppsspp#643
No description provided.