wololo.net/talk

Didn't work on 6.35 or 5.00 m33 either. PSPLink (on 5.00) gives me an error at 0x088...something, which would indicate that the jump to 0x09000000 didn't happen, but I heavily modified your files for debug purpose so my crash is probably not reliable.
I unfortunately don't have the time I'd like to spend debugging this.

JJS, there's still a high chance that my build environment is completely busted for some reason, so forgive me for asking, but did you manage to run that signed version of HBL on your PSP?

It certainly works for me . Seriously, I didn't mean to waste anyones times with this .

The patched HBL runs for me on 6.20 TN and OFW 6.20 and 6.37.

Did you try to rebuild the exploit SDK after your build of the signed launcher? I am using MinPSPW 0.11 now and I suspect that other versions will produce slightly different code so that the stubs end up at a different memory address. Of course h.bin will in this situation try to run functions from random addresses and goes down in flames.

Anyway, I attached my build of the launcher (unencrypted for not having to upload 5 MB).

JJS wrote: Did you try to rebuild the exploit SDK after your build of the signed launcher? I am using MinPSPW 0.11 now and I suspect that other versions will produce slightly different code

Ah, I might still be using 0.10, will try that, and your Eboot too.

Edit: yup, your eboot seems to work, so that's my stubs which are incorrect...

I'll add all this to the SVN if you don't mind

submitted your changes to the SVN

I know it's not my place to be posting here, but couldn't HBL be made much more efficiently if it was rebuilt. Instead of having it hack itself you could clean up the code so it's a homebrew that loads homebrews. Wouldn't doing that increase homebrew compatability, aswell as enable prx's to load? (like daedalus)

First, it's not hacking itself. The EBOOT loads h.bin and the rest comes on. And second, this is not a petition forum. You have the source code, do it yourself.

jaja2u : that's a perfectly fine suggestion, but the good point of doing it the way JJS did is that there is close to no modification to HBL itself, and JJS "only" had to create the launcher and a few config files for HBL.
On top of that it guarantees we have a compatibility similar to Patapon without the need to do extensive testing, since the code of HBL hasn't changed since R112.

Yes, if we built HBL again from the ground up, we could optimize many things.
It would also take 6 months, and I'm definitely not sure it would increase compatibility (at least not at first) or enable plugins (at least not any better that what we could do with the current version).

The current solution is the best for now, in terms of maintenance. If one day it becomes clear that signed homebrews will stay forever on the PSP, it might be a good idea to optimize HBL for this, but we have no short-term plan for this, I think.

Also, I kind of disagree with moskito on one point, I think this verson of HBL is hacking itself, since it loads HBL which then unloads the homebrew in a somewhat dangerous way (like all versions of HBL always unload the game the rely on)

I have nothing against jaj2u suggestion, but it's quite more time expensive to code what he has said, that's why I'm saying he can try to do it himself.

Wololo wrote:I think this verson of HBL is hacking itself, since it loads HBL which then unloads the homebrew in a somewhat dangerous way (like all versions of HBL always unload the game the rely on)

Well still stand on my point. Being "dangerous" (as far as loading a plain binary is "dangerous" ) does not mean it's hacking itself. There's no hack whatsoever, it loads a file into memory and jumps to it. I sincerely cannot see where the hack is. Anyway, this discussion is pretty much pointless.

I just submitted R114. Turns out that some files were missing from R113 that prevented building of the launcher and the signed HBL. I introduced some defines to remove the unnecessary hooks, the creation of the kernel memory dump and the p5 stub parsing. Also the HBL loading address is now considerably higher. The launcher hasn't changed.

Edit: Submitted R115 because the last commit didn't include the config folder for the signed HBL. I swear that TortoiseSVN showed me those files as "added", grumblegrumble .

JJS wrote:I swear that TortoiseSVN showed me those files as "added"

You cannot trust a Microsoft's OS...