Argh, I forgot to count the 2^(n-1) bytes for the table of compressed values...
This makes the compression gain negative in all cases, ouch...
I understand that changing the location of zero page variables to addresses having the same value than the most common values can lead to better compression ratios, but this is impossible for my program because I have a table of 80 pointers in page zero that must be kept in consecutive locations...
Ok, back to manual compression, then...
Minigames 2006 : Go !
Ah, you're not using the latest version...
It now has easier controls, on http://oric.free.fr/TAPES/proton.zip
(and source code too... quite funny to see that for a 1K binary, a very sparsely commented source is 16K, or 10 pages long...
It now has easier controls, on http://oric.free.fr/TAPES/proton.zip
(and source code too... quite funny to see that for a 1K binary, a very sparsely commented source is 16K, or 10 pages long...