Gzip 1.2.4 hack (OpenWatcom compiles)

[Rugxulo]

I have seen Ilia M.'s Gzip 1.2.4 hack before (news://comp.compression), it's nice, but it never compiled for me. Now I notice that it's missing a prototype for deflate_max(). And yet still some compilers don't seem to like it (DJGPP, at least the output doesn't match, it's worse!). Nevertheless, I somehow stumbled upon how to build it with OpenWatcom 1.8, and that works fine, so now I have both Win32 (smaller!) and DOS builds.

http://rugxulo.googlepages.com/gzhakwat.zip (311k, includes everything!!)

P.S. I tried searching for a pre-existing topic about this, but apparently I'm a robot spammer because all my (10+) attempts to read the captcha failed !!!

[Rugxulo]

some compilers don't seem to like it (DJGPP, at least the output doesn't match, it's worse!).

Seems to work correctly with DJGPP (GCC) only if I define -DNO_ASM and remove any mention of match.S from the makefile. Just FYI in case anyone else gets frustrated !!

[Kaw]

[offtopic && if I'm correct]
Concrats with the 2^8th topic
[/offtopic]

[LovePimple]

Thanks Rugxulo!

[GerryB]

What is this hack? I poked around comp.compression and couldn't find any reference except this one.

[Bulat Ziganshin]

What is this hack? I poked around comp.compression and couldn't find any reference except this one.

try to search on the old forum

[encode]

What is this hack? I poked around comp.compression and couldn't find any reference except this one.

Are you sure? And how about this one?

[joerg]

"Deflate with better parsing
... leading higher compression
without breaking compatibility ..."

do you know pigz by Mark Adler ?

http://www.zlib.net/pigz/

- "parallel implementation of gzip"
- "is a fully functional replacement for gzip"
- "exploits multiple processors and multiple cores ..."

win32-binary by Eduardo Terol

the binary and the source-code
is downloadable from
http://www.leszer.net/index.php/my-software.html

but the output-file from pigz
seems not to be identically
with the output-file from gzip

it is a very interesting program
but i think
the program needs more testing
by several people

can you please do a test too?

[Black_Fox]

multi-processor support =/= better parsing

[joerg]

you are right of course

but such a multicore-/multiprocessor-implementation

is a test worth

isn?t it?

best regards