daveshields
commented
3 years ago Hib,
Sorry, I never had time to look into this.
I haven't worked on SPITBOL for X64 for a couple of years. Lately I've started work on porting SPITBOL to arm, with special emphasis on the Raspberry Pi.
If you are still interested in working on the X64 version, I have just invited you to join the project. I am more than willing to grant you the access so you can make the changes you deem fit.
I haven't touched OSINT except to rename a couple of files (they're no longer needed).
I have reorganized the main directory, mainly to rename many of the files, so I can use a common convention for x64 and arm implementations. These changes should be self-evident.
Please let me know if you are interested in working on this. Feel free to call me at 914-924-8781 if you have any questions (I live near New York City).
thanks, dave
On Thu, Sep 10, 2020 at 3:42 PM Hib Engler notifications@github.com wrote:
Closed #7 https://github.com/spitbol/x64/pull/7.
— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/spitbol/x64/pull/7#event-3752832164, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAARPPJCSH3BCQSUMYXAE4DSFETZTANCNFSM4EXGVWBQ .
hibengler
It took a while, as I am rusty at assembly, but I fixed the .spx file so it could be read without croaking. The main problem was that the stack size reported was 12 bytes long - 12? Turns out, there were some hard-coded 4's in there. That caused the redirect minimal code to go right on past the upper boundary of its checking and modify the offsets on much more than the stack.
Also, calling to assembly uses a new style to receive calls from standard C - different than pop/store of the past.
I learned alot about how the system works, by reading through this code, and 88source as a second source. And in doing so, I received dramatic speedups on one of my projects - an english parser. This parser knows 90,000 words. It took 11 minutes to generate the complex patterns to match that. On regular spitbol, the startup time was a much better 18 seconds. But with the .spx file, the speed went to 0.045 seconds to startup.
The compression code is still broken. That can be fixed at a later time. I am thinking of switching out that section for a gpg code block that includes digital signatures of the code, and encryption requiring certain keys. This would go a long way to abate the fear of spitbol, basically being self modifying code.