Anyone else feel that this version of HEN is more unstable then the previous? Like with the previous HFW and HEN it was 9 out of 10 times it would successfully start for me when I clicked on the icon. Now its 9 out of 10 times it fails and reboots the system.
If you are talking about the exploit initialization failures issue, it's not exactly a question of stability even though it may appear like that to you.
I only worked on PS3HEN for a short time during its development phase & I haven't checked the source used in successive releases but I doubt the code responsible for the initialization attempts success or failure has changed a iota ever since it was released in the first wave of ps3xploit tools releases, more than a year before HEN itself was released.
I wrote that code & there is a bug in it, a bug that I found & fixed a couple of years ago when I rewrote a number of functions I decided to include into my PS3 Toolset framework but because PS3HEN still uses an older version of the code, over 3 years old, it didn't benefit from that rewrite.
The HEN exploit initialization failures (and similar init failures in all other ps3xploit tools except the PS3 Toolset) happen when the target signatures the code searches for are not aligned in memory as the code expects.
Things like prior traffic before loading the exploit page or simple changes made to the exploit page, even a small addition or removal of text in the GUI, can potentially impact the layout of browser related data in memory, in other words, it modifies the likelihood of the target signatures to be located on certain offsets and that translates into a clear variation of the signature search success rates.
The XMB shortcuts to enable HEN for instance enjoy a better success rate because the minimalistic approach without GUI makes the target signature statistically very likely to be aligned in memory at offsets that the search algo won't miss. In the same way, it's better to load the exploit page in the browser as the homepage so there is no prior traffic, not even the **ps3xploit.com >Domain no Longer owned by team** (ps3xploit.me =new) main page.
Or alternatively add the exploit page as a favourite & set the homepage to blank so you can start the browser & use the favourite directly.
Currently, without modifying the source code, those are the only proven ways to improve the success rate.
It's more effective than clearing cookies, cache etc.. as usually advised. Cookies cannot impact memory layout, none are used by the exploit page so no cookie related data gets loaded into memory. As to cache clearing, I have not seen any statistical evidence of an impact on signature locations.
As it turns out, the whole signature search algorithm I used in that old exploit implementation is rubbish (lol), not only it fails to search half the possible offsets due to that bug but in terms of performance it is a terrible design as well, it does in a loop with thousands upon thousands of iterations of string cutting & string to number conversions, something that can be achieved in one line of js code extremely efficiently through string manipulations that the javascript language was specifically optimised to perform.
If the search algorithm in HEN was replaced by the new algo used in the Toolset, I am pretty sure the HEN exploit initialization stage could reach a 100% success rate & in a fraction of the time it currently takes.
