Re: LZ: New idea for compression program!


[Prev][Next][Index][Thread]

Re: LZ: New idea for compression program!



At 08:33 PM 9/11/96 -0500, you wrote:
>On Wed, 11 Sep 1996, Scott J. Rein wrote:
>


>You have a good point.  It'll be better for the users if
>you do it this way.  But how much code will you have to
>put in the compressed program to make this work?  I would
>get good, solid estimates of that before deciding.  You
>might be right but if you're not you need to know it.


That is why I have been talking and not working as some people
in this group sent me mail about :(.  It is simply a question
of whether it is worth the extra amount of code in each
comrpessed string vs. one shell that will take up more space,
but include the decompressor.


>If you write a shell, you'll have to make PROGRAM_ADDR point
>to the beginning of the program when you execute it.  If you
>don't write a shell you'll need to make PROGRAM_ADDR point to
>the beginning of the compressor before you run it, and then
>set it back to the new address of the decompressed program.
>You'll need to find out from the zshell people if you have to
>reset it or not.


So I would have to set it twice.  I plan on having the
decompressor return to the compressed string and then have the
compressed string run the decompressed one.  does this solve
the problem?


>You might not need to create variables if you can just get
>access to some ram that you know is unused to decompress the
>program into and then execute it from.  There might be pointers
>in the system to the bottom or top of available memory.  There
>probably are.  I don't have any idea if they're known or not.
>If not, you'll have to find out about using a variable.


I didn't think about just placing it in the memory.  Yah, if
there are pointers to the free space then I don't have to worry
about actually creating a string, just an address.


--	Scott Rein
	srein@rain.org


Follow-Ups: References: