Text 411, 51 rader
Skriven 2006-03-03 17:00:27 av mark lewis (1:3634/12.0)
    Kommentar till text 408 av Peter Knapper (3:772/1.10)
Ärende: z.dll
=============
 PK> Hi Mark,

 ml> i seem to have a conflict with z.dll... i run webalizer 
 ml> which has its own z.dll... and neither seems to work 
 ml> with the other... why? i dunno... i've tried to change 
 ml> the name of z.dll for binkd to zlib.dll but it fails to 
 ml> load... however, i was able to get webalizer to stop 
 ml> putting entries in the popuplog by not loading the 
 ml> binkd z.dll...

 PK> It sounds like there may be a .DLL version mismatch to me. 

well, based on this, i went digging with my hex tools... i see that both dlls
contain 1.1.3 version numbers with in them... however, the first thing i noted
was that the webalizer one also has a "dos stub"(?) at the top that says that
it cannot be run in DOS mode... the one i got with binkd does not...

additionally, the one with webalizer contains this (hex and printables)...

0008F0  07 00 A8 00 01 BC 35 08 45 4D 58 4C 49 42 43 4D        5 EMXLIBCM
000900  03 45 4D 58 00 00 00 00 00 00 00 00 00 00 00 00   EMX

while the one with binkd contains this...

000C70  07 00 F4 00 01 BD 7A 08 45 4D 58 4C 49 42 43 53        z EMXLIBCS
000C80  03 45 4D 58 00 34 00 13 8D 03 03 00 00 02 00 1C   EMX 4


i'd almost hazard a guess that the top listing is compiled with the medium
memory model while the bottom one is compiled with the small memory model...
but that's a pure wildassed guess... other than the differences between them
with and without that stub, i don't have any more of a clue than a 2x4 O:)

oh! i also noticed that one has the internal functions listed at the beginning
of the file whereas the other has them listed at the end... physically
speaking, that is... i think i noted some effects of some sort of compression
on the binkd version, too... effects like seeing 1.1.3 listed only once and
some sequences of letters missing in additional places... ie: foobar showing up
and later on foxxar showing up where xx represents gobbledegook instead of
"ob"... it is a classic indicator of compression but may also be something done
by the compiler, too...

 PK> I don't run the Dynamically linked version of BinkD, I use the 
 PK> Static version so I don't see any such problem.

hummm... i hadn't thought about doing that... it is something i might look
into??

)\/(ark

 * Origin:  (1:3634/12)