Compressed files and QuickBMS reimport problem (to anyone!)
Posted: Mon Jan 23, 2012 11:04 pm
I post this thread here because often some people reported problems in reimporting compressed files via quickbms -r -w even if the original size of the modified file is the same (for example you modified one or no bytes).
as already explained the problem is not caused by quickbms but it's just a limitation of the various zlib/deflate open source libraries that are not much optimized for the compression ratio.
for example zlib has almost no optimization (worst results) while the one used in 7zip (advancecomp, default in quickbms) is better but still not enough.
note that I talk about the files compressed with zlib/deflate and NOT the other files, so I refer just to the modified files that have the same size of the original ones. ok?
as far as I know kzip http://advsys.net/ken/utils.htm is one of the best and slowest tools for the deflate compression so from QuickBMS 0.5.5 (released just now) it's used automatically if the compressed file is too big.
so instead of giving you the usual error message about the file size now QuickBMS will automatically try to compress it with kzip.
everything is completely transparent to you except for a message that notifies the usage of this alternative method (because it's a work-around that consumes many resources and so it must be visible when used).
if you have had this problem in the past and you want to try it feel free to report your results and what you think.
this is the last chance because there are no other solutions except writing an extreme deflate compression algorithm by myself... and no I have no intention and capabilities for doing it :)
QuickBMS:
http://quickbms.aluigi.org
as already explained the problem is not caused by quickbms but it's just a limitation of the various zlib/deflate open source libraries that are not much optimized for the compression ratio.
for example zlib has almost no optimization (worst results) while the one used in 7zip (advancecomp, default in quickbms) is better but still not enough.
note that I talk about the files compressed with zlib/deflate and NOT the other files, so I refer just to the modified files that have the same size of the original ones. ok?
as far as I know kzip http://advsys.net/ken/utils.htm is one of the best and slowest tools for the deflate compression so from QuickBMS 0.5.5 (released just now) it's used automatically if the compressed file is too big.
so instead of giving you the usual error message about the file size now QuickBMS will automatically try to compress it with kzip.
everything is completely transparent to you except for a message that notifies the usage of this alternative method (because it's a work-around that consumes many resources and so it must be visible when used).
if you have had this problem in the past and you want to try it feel free to report your results and what you think.
this is the last chance because there are no other solutions except writing an extreme deflate compression algorithm by myself... and no I have no intention and capabilities for doing it :)
QuickBMS:
http://quickbms.aluigi.org

