cuneiform team mailing list archive
-
cuneiform team
-
Mailing list archive
-
Message #00785
Re: experiments with Cuneiform on ARM
On Fri, 2011-02-11 at 13:07 +0200, Aapo Rantalainen wrote:
> If yes, does it imply that ARM-version will work after warnings are fixed?
My experience with mc on MIPS tells me that if you fix all of the
pointer / type size warnings there is a high potential for it to start
working half-correctly at very least.
The reason behind those is basically that the code makes implied
assumptions on the type sizes which do not necessarily hold on various
architectures, which in turn leads to invalid computations and memory
corruption and as a result, although the program may compile, produces
garbage output.
Porting to MIPS / ARM is generally a very good way of cleaning up the
mess in your programs :-) maybe one day I will even get a ShivaPlug just
because of this...
--
Sincerely yours,
Yury V. Zaytsev
Follow ups