This is the mail archive of the gcc-help@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: help building cross compilation ia32 to ia64


Tim Prince wrote:
Although your 'ia64-linux-gnu' Linux distro would be old, with 'glibc-2.2.2' or something,
Why this propaganda? If you sincerely believe OP is confused between x86-64 and ia64, say so or ask for clarification.
Ok, I looked at that (HP) site and there were both 'amd64' and 'ia64' toolchains... Why I wrote that
"propaganda", maybe came from the fact that I myself first saw that 'ia64' being 'x86_64' and didn't
care to edit everything away... Or the current flu mixed my brains... Sorry for this!
I don't see much relevance of your opinion on "mainline distros" apparently excluding the SuSE, RH, and Debian which are in widespread use.
Yes, Debian and the RHL-ES seem still to support Itanium... No evidence about SuSE, nothing seen
about 'ia64', not even about 'x86_64' on their "SuSE Linux 10" pages, a 'x86' server product seemed to be
available...


OP presumably has a target in mind.
But couldn't tell what on earth that is... Maybe it is "WinNT4.0 for ia64" !
I'm somewhat amazed by the contrast in your advice to use a current release or newer version of binutils, but not to use anything else from the last 3 years. RH 4.3 or newer definitely have advantages over earlier RH, aside from needing one of the binutils upgrades you mentioned.
GIGO - "Garbage In - Garbage Out" ! People get bullshit if they ask using bullshit !

Ok, too many questions like "I have a problem, can you help me?" are seen and when seeing that one either
doesn't answer at all or then switches the "lecture mode" on :-( The least what could be expected is that there
were at least some clues for what on earth the problem is!


Whether someone who cannot see any difference between 'Linux/ia64', 'AIX/ia64', 'HP-UX/ia64 and
'FreeBSD/ia64' plus that 'WinNT4.0/ia64' (if it existed, maybe 'Win2k/ia64' too) but see them all as 'ia64',
could succeed in any builds without continous hand-holding, is quite unclear.... :-)


Anyway one serious problem is that the downloaded tools were for the Linux/ia64 host, not for the 'ia32'
host ! I checked the 'tco-gcc-20061029.inst.tar.gz' package, its 'gcc' told :


$ /data1/opt/toolchain/tco-20061029/bin # file gcc
gcc: ELF 64-bit LSB executable, IA-64 (Intel 64 bit architecture), version 1 (SYSV), for GNU/Linux 2.4.1, dynamically linked (uses shared libs), for GNU/Linux 2.4.1, not stripped


The best approach in cases like this could be at least to tell :

1. What is the host system ? Fedora 5 executables running OK on Red Hat 7.3 doesn't sound that clear...
If one downloads prebuilt binaries, their makers usually have some 'bleeding edge' system :-(


2. What is the target system? Producing a toolchain with 'gcc-4.1.1-glibc-2.5' and then trying to run the
produced apps on a "Red Hat Linux 7.2/ia64" doesn't either sound very sane....



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]