Tue, 29 Apr 1997 23:54:30 +0200 (MET DST)
> The CMUCL bytecode file format is endian-dependent,
> there's .lbytef for little-endian and .bytef.
Is there more to it than just swapping bytes as the files are read,
perhaps skipping strings/symbols?
If not, perhaps a minor change in bytecode file format can make it possible?
After all, swapping bytes can be considered a free
operation wrt the time needed to fetch the files from disk/network,
all the more when some people are considering postprocessing the
bytecode with JITs and the like...
Even a complete newbie like I could jump in and add byteswap support
to the linux ufs driver, so that a Linux PC could equally read either-endian
(i386-FreeBSD and Sparc-Sun) partitions; that's *trivial*!
Just make it easy to quickly identify the endianness of your code...
Anyway, ===> FOLLOWUP-TO: email@example.com <===
== Fare' -- firstname.lastname@example.org -- Franc,ois-Rene' Rideau -- DDa(.ng-Vu~ Ba^n ==
Join the TUNES project for a computing system based on computing freedom !
TUNES is a Useful, Not Expedient System