herber@bgsuvax.UUCP (Steve Herber) (02/01/85)
Does anyone have a copy of AT&T's cpio tape i/o program ported over to 4.2BSD? It would be nice if I could read my System V tapes onto my 4.2BSD system WITHOUT having to boot up System V. Steve Herber osu-eddie!bgsuvax!herber Bowling Green State Univ.
david@ukma.UUCP (David Herron, NPR Lover) (02/12/85)
Yes, I have it on here. I just compiled it and it ran. It fails
in interesting ways however, which I don't feel like tracking down,
simply because I don't get that many cpio tapes.
I may have had to change *something*, don't remember if I did or not,
but if you are *really* wanting to know I can find out....
--
-:--:-
David Herron;
ARPA-> "ukma!david"@ANL-MCS or david%ukma.uucp@anl-mcs.arpa
UUCP-> {ucbvax,unmvax,boulder,research}!anlams!ukma!david
UUCP-> {mcvax!qtlon,vax135,mddc}!qusavx!ukma!david
UUCP-> {A-Large-Portion-of-The-World}!cbosgd!ukma!david
No stupid sayings (I can't think of one).
No stupid disclaimers (Nobody else would claim my statements anyway).smk@axiom.UUCP (Steven M. Kramer) (02/22/85)
It almost works on 4.1-2. There are 2 problems. First, u must
handle the different file types. Second, there is a variable
declared as short that is used to hold the mode word. When it
is checked vs. dirs on 4.2, sign extension occurs. To fix that,
either make it unsigned or make it a long. I forget its name. The
net effect of the bug is that the -d option doesn't work. The net
effect of the fix is that the -d option works. (How profound.)
--
--steve kramer
{allegra,genrad,ihnp4,utzoo,philabs,uw-beaver}!linus!axiom!smk (UUCP)
linus!axiom!smk@mitre-bedford (MIL)jsdy@hadron.UUCP (Joseph S. D. Yao) (02/22/85)
Minor problems: it does funny things with -l (made all mod times > 2^31)
and doesn't understand suymbolic links: it makes copies of files or
empty directories. May be other problems. We resolved to use 'tar'.
Joe Yao hadron!jsdy@seismo.{ARPA,UUCP}guy@rlgvax.UUCP (Guy Harris) (02/28/85)
> It almost works on 4.1-2. There are 2 problems.
Three, really: the guy who wrote it "knew" that the st_atime and
st_mtime fields of the "stat" structure were contiguous, so they
passed a pointer to the "st_atime" field to "utime". Well, guess what?
They're *not* contiguous in 4.2 (and, if you read the S3/S5 manual,
nowhere does it commit to them being contiguous; they went out of
their way not to commit to anything about the layout of the "stat"
structure). "pack"/"unpack" and "file" suffer from the same
deficiency.
Guy Harris
{seismo,ihnp4,allegra}!rlgvax!guy