On Tue, Nov 28, 2000 at 02:06:41PM -0800, Stephan T. Lavavej wrote:
>The question is, if compression involves a one-time, five-minute
>cost on the part of the developer and saves everyone a few seconds of
>download time and a few K of HD space, then why not?
Perhaps since UPX requires a writable /tmp?
>No, the decompression cost is not paid again and again while the executable
>runs: there is no (significant) memory cost or compute cost. I don't claim
>to understand how UPX works, although I intend to learn some day, but the
>UPX homepage notes "Your executables suffer no memory overhead or other
>drawbacks".
UPX decompresses the program to /tmp, then runs it from there. Another
side effect is that the program (at least with my old version of UPX)
shows up as `3' instead of `mprime' in `top' and `ps', among others...
>The
>two things that don't work with UPX are executables that want to read data
>from themselves and things with "overlays".
This is for DOS only, and doesn't apply to Linux, since the Linux
version doesn't feature in-place decompression.
>UPX works for Linux too - isn't there a Linux port of Prime95? The entry
>for UPX's Linux capabilities in its readme does detail some of its inner
>workings. It is lengthy.
Oh, you're talking about the Windows version, not the Linux version. Oh
well, I'm not too sure it's a good idea to use in-place decompression,
when we already have enough problems with fragmented memory maps etc. --
as some seem to have noticed, there can be quite a hefty speed penalty
if the memory mapping has problems (cf. the ReCache discussion) :-)
/* Steinar */
--
Homepage: http://members.xoom.com/sneeze/
_________________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
