Jim Meyering wrote:
Matthew Woehlke <[EMAIL PROTECTED]> wrote:
Jim Meyering wrote:
I see that both freebsd6 and netbsd1.6 disallow the usage in
your test case:
netbsd$ echo x | /bin/dd bs=3 ibs=10 obs=10 conv=sync | wc -c
dd: ibs: illegal argument combination or already set
0
Solaris 10 does what you propose.
Can anyone check other implementations?
I can test oodles of systems (about six Unix flavors across almost a
dozen OS/HW combinations), but I'm not clear what you are looking for.
Should I just send the output from each system?
Yep, just run this (assuming your vendor-provided dd is in /bin/dd):
echo x | /bin/dd bs=3 ibs=10 obs=10 conv=sync | wc -c
Ok:
==== Set 1
0+1 records in
1+0 records out
3
==== Set 2
0+1 records in
1+0 records out
10
==== PPC/Darwin 7.9.0
dd: ibs: illegal argument combination or already set
0
==== x86/Darwin 8.8.1
dd: bs supersedes ibs and obs
0+1 records in
1+0 records out
3 bytes transferred in 0.000039 secs (76725 bytes/sec)
3
====
Set 1 was:
IRIS/IRIX 6.5
Alpha/OSF 4.0F, 4.0G, 5.1
SPARC/Solaris 5.6 - 5.10
x86/Solaris 5.10
PA-RISC/HP-UX 11.00, 11.11
IA64/HP-UX 11.22
PPC/AIX 5.3
Set 2 was various versions of coreutils on x86/Linux, IA64/Linux, and
the infamous ;-) Tandem-NSK/OSS. Some example versions are 4.1, 4.5.3,
5.97 and 6.6; coreutils has been doing it "wrong" since it was known as
fileutils :-).
Some in Set 1 also had the '3 bytes transfered...' line, but I don't
recall seeing any of the warnings ala x86/Darwin 8.8.1.
--
Matthew
We interrupt this e-mail to bring you a lame signature attempting to be
witty.
_______________________________________________
Bug-coreutils mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/bug-coreutils
