what does 9013 have that 9005 lacks

what does 9013 have that 9005 lacks

Post by Geoff Colem » Wed, 09 May 1990 02:28:32



        Could someone out there please summarize the differences between
the 9013 and 9005 loads? I'm wondering wether I should spend more time
working around "features" of AIX or if they are changed (fixed) in the new
load. One thing I have hit a roadblock on is getting a bidirectional
uucp line to work. If anyone else out there has managed this could you
please drop me a line as to how to do it. As well changing the speed line
in /etc/ports to 19200 seems to have no effect. Is there something else
that has to be changed to change the speed for a getty.

        Overall I'm quite pleased with the 6000. I've moved over cnews,
and vn with no problem (except that cron setup is brain dead). I
moved over smail 3.1 with only a few problems (awk doesn't like "if ( !dummy )
and if a string is defined as char *dummy="string" it segment violates if you
try to change a charter in dummy). We have our PROGRESS applications moved over
and they run great.

Geoff Coleman
Unexsys Systems


 
 
 

what does 9013 have that 9005 lacks

Post by Steve DeJarne » Wed, 09 May 1990 10:06:11



>    Overall I'm quite pleased with the 6000. I've moved over cnews,
>and vn with no problem (except that cron setup is brain dead). I
>moved over smail 3.1 with only a few problems (awk doesn't like "if ( !dummy )
>and if a string is defined as char *dummy="string" it segment violates if you
>try to change a charter in dummy).

        The problem where modifying dummy causes a Segmentation violation is
due to ANSI C, I suspect.  ANSI C says that you can't change any part of a
string constant (or words to that effect).  If you want to get around this,
change the definition/declaration from char *dummy="string" to
char dummy[]="string".  Not much difference, but just enough for ANSI.

Quote:>Geoff Coleman



(415) 855-3510       These opinions are my own.  I doubt IBM wants them.......

 
 
 

what does 9013 have that 9005 lacks

Post by Bjorn Engs » Wed, 09 May 1990 17:07:20



|
|and if a string is defined as char *dummy="string" it segment violates if you
|try to change a charter in dummy)
Yeah, that's perfectly legal, it's documented, and xlc can put "string" in
the data segment if you add -qnoro option.  
--

                Path:           uunet!mcsun!orcenl!bengsig
 
 
 

what does 9013 have that 9005 lacks

Post by Bjorn Engs » Thu, 10 May 1990 17:37:36



  [as a followup to why char *dummy="string"; dummy[0]='S'; segm violates]
|If you want to get around this,
|change the definition/declaration from char *dummy="string" to
|char dummy[]="string".  Not much difference, but just enough for ANSI.
As I pointed out in another followup, constant strings should not be
modified according to ANSI.  The precise wording in K&R2 (A2.6) is that the
behaviour of modifying a string literal is implementation defined.

char *dummy1  = "string1";
char dummy2[] = "string2";

dummy1[0] = 'S';        /* violates ANSI by trying to change 's' in "string1" */
dummy2[1] = 'T';        /* violates as well trying to chage 't' in "string2"  */

dummy1 = "new1";      /* is OK, since it changes the char pointer dummy1, but
                           after doing it, the string "string1" cannot be
                           accessed any longer using dummy1 */

strcpy(dummy2,"new2");        /* violates, since you try to overwrite "string2" */

The xlc compiler for AIX 3.1 can be given the option -qnoro to put string
literals into the data-segment, and then they can be modified.

Please continue this discussion in comp.lang.c if necessary.
--

                Path:           uunet!mcsun!orcenl!bengsig

 
 
 

what does 9013 have that 9005 lacks

Post by Dick Du » Fri, 11 May 1990 05:23:32




>   [as a followup to why char *dummy="string"; dummy[0]='S'; segm violates]
> |If you want to get around this,
> |change the definition/declaration from char *dummy="string" to
> |char dummy[]="string".  Not much difference, but just enough for ANSI.

No.  DeJarnett is correct that this will make the problem go away (although
it's incorrect to say that there's "not much difference").  The first
declaration is for a pointer; the second is for an array.  Read on...

Engsig continues:

Quote:> char *dummy1  = "string1";
> char dummy2[] = "string2";

> dummy1[0] = 'S';   /* violates ANSI by trying to change 's' in "string1" */
> dummy2[1] = 'T';   /* violates as well trying to chage 't' in "string2"  */

No; the first violates ANSI but the second does not.  The first assignment
is incorrect because dummy1 is a pointer to a constant string; the assign-
ment attempts to modify a constant.  But the second assignment is correct
because dummy2 is a variable; it's an array of characters.  The constant
"string2" is used to provide the initial value of the array (and implicit-
ly the length), but dummy2 is a variable which is an array of eight char-
acters; assignments to its elements are allowed.

(Assignments *to* dummy1 are allowed also, of course; it's a pointer.  But
you can't assign *through* it until it's no longer pointing at a constant.
That is, in context of the above, you could say
        dummy1 = dummy2;
which makes it point at the variable array; then it would be OK to say
        dummy1[0] = 'S';
which would have the same effect as
        dummy2[0] = 'S';
--

   ...If you plant ice, you're gonna harvest wind.

 
 
 

what does 9013 have that 9005 lacks

Post by Geoff Colem » Sat, 12 May 1990 06:21:59




-> |
-> |and if a string is defined as char *dummy="string" it segment violates if you
-> |try to change a charter in dummy)
-> Yeah, that's perfectly legal, it's documented, and xlc can put "string" in
-> the data segment if you add -qnoro option.  
-> --

->           Path:           uunet!mcsun!orcenl!bengsig

        I figured that it was probably an ANSI C thing but being someone
who always thought one of the advantages of C was that it allowed you to
do what you wanted (or at least told it to do) I don't keep up as much
with the standards work and such as I should.
        The other problem is at this point the documentaion available
from IBM in a hardcopy format is close to zero.
        An example of this is that I finally figured out how to get our
tape to quit acting as a tape retensioner. If you go into smit and
through devices you can change the tape characteristics. The last of the
characteristics  deals with retensioning. Now if only IBM would default this to no.

 
 
 

what does 9013 have that 9005 lacks

Post by Mark Bro » Sat, 12 May 1990 22:01:17



Quote:Coleman) writes:
>    An example of this is that I finally figured out how to get our
> tape to quit acting as a tape retensioner. If you go into smit and
> through devices you can change the tape characteristics. The last of the
> characteristics  deals with retensioning. Now if only IBM would

default this to no.

We default to "retension" because this is the more safe, conservative
approach. And you know where the machine was made....

Disclaimer: I don't speak for IBM. They don't make me wear ties.
Mark Brown   IBM AWD / OSF  | Facts are simple and facts are straight

The Bad     uunet!osf!mbrown| Facts all come with points of view
The Ugly     (617) 621-8981 | Facts don't do what I want them to

 
 
 

what does 9013 have that 9005 lacks

Post by Bjorn Engs » Sat, 12 May 1990 16:38:30



|char dummy2[] = "string2";
|dummy2[1] = 'T';       /* violates as well trying to chage 't' in "string2"  */
This couldn't be more wrong of course.  dummy2 is an array of 8 characters
initialized with "string2" (and a 0), and it is perfectly legal to overwrite
this array.

Even if you think you know C, it seems you can blow it - at least I can :-(
--

                Path:           uunet!mcsun!orcenl!bengsig

 
 
 

what does 9013 have that 9005 lacks

Post by Mark Whetz » Sat, 12 May 1990 22:39:36


   [ c discussion deleted ]
Quote:>    The other problem is at this point the documentaion available
> from IBM in a hardcopy format is close to zero.
>    An example of this is that I finally figured out how to get our
> tape to quit acting as a tape retensioner. If you go into smit and
> through devices you can change the tape characteristics. The last of the
> characteristics  deals with retensioning. Now if only IBM would default this to no.

Not on 9005O build it wont.   Smit brings up the change/show menu
and it had the retension setting protected and says Field cannot be changed.

About documentation, you are mostly correct, but the INFO command can
be fairly helpful, but the hypertext docs do still need MUCH work especially
on the O build.
--
Mark Whetzel     My comments are my own, not my company's.
Western Geophysical - A division of Western Atlas International,

                                   UUNET address:  uunet!airgun!markw

 
 
 

what does 9013 have that 9005 lacks

Post by Mark Bro » Sun, 13 May 1990 03:05:54



Whetzel) writes:

Coleman) writes:
> Not on 9005O build it wont.   Smit brings up the change/show menu
> and it had the retension setting protected and says Field cannot be changed.

> About documentation, you are mostly correct, but the INFO command can
> be fairly helpful, but the hypertext docs do still need MUCH work especially
> on the O build.

Agreed, the 9005 build is/was shaky. Please remember that even 9013 is
PRE-release...

I think I *can* speak for IBM, in this one case, in saying that we are
glad for your bug reports/comments.

DISCLAIMER: I don't speak for IBM. They don't make me wear ties.
Mark Brown   IBM AWD / OSF  | Facts are simple and facts are straight

The Bad     uunet!osf!mbrown| Facts all come with points of view
The Ugly     (617) 621-8981 | Facts don't do what I want them to

 
 
 

what does 9013 have that 9005 lacks

Post by jim fro » Thu, 17 May 1990 11:53:55



>    The other problem is at this point the documentaion available
>from IBM in a hardcopy format is close to zero.

We have wonderful documentation for all of the compilers, although OS
documentation is nonexistent.  Unfortunately the compilers are the
things we least need documentation for (come on now, who wants VS
COBOL documentation? :-)

Quote:>    An example of this is that I finally figured out how to get our
>tape to quit acting as a tape retensioner. If you go into smit and
>through devices you can change the tape characteristics. The last of the
>characteristics  deals with retensioning. Now if only IBM would default this to no.

PLEASE point me towards this fix!

jim frost
saber software

 
 
 

1. followup to differences between 9013 and 9005



->>>
->>>   Could someone out there please summarize the differences between
->>>the 9013 and 9005 loads? I'm wondering wether I should spend more time
->>>working around "features" of AIX or if they are changed (fixed) in the new
->>>load.
->>>      [ deleted stuff... ]
->>
->> The big things fixed for me was csh (yes, csh scripts now run), smit and some
->> X stuff. In addition, the PC simulator is now running.
->>
->> There is really no comparison between the two builds. It's well worth the time
->> to upgrade.

        We would gladly if two things would happen. First of all if the local
IBM office would get a load newer than 9005K1. THe second is that someone
at IBM told me that the Object File Format has changed from 9005 and since
we are dependent on having PROGRESS run and since PROGRESS have told us that
they are tired of hitting a moving target we have to wait for a GA operating
system before PROGRESS will port their code again.
        In the mean time I'm waiting for a second 6000 box but I want to know
if I should bother with trying to port more code like KERMIT or should just
wait for a new compiler which doesn't fail horribly on it (fatal error in
xlcentry).

Geoff Coleman
Unexsys Systems

->>
->>
->>
->>

2. Arrrgh! rsync "chroot failed" error message!

3. differences between 9013 and 9005 (plus some impressions)

4. How to Kill Message Once Sent ..?

5. Linuxant and Broadcom Wireless on an HP NX-9005

6. Sun Packages

7. AIX 9013 & pcsim

8. Possible to trick RADEON with Xfree86 SVGA 3.x?

9. Problem with re-partitioning HD for DOS after having Linux on it

10. Can this be done without having to spoof IP's ??

11. Having Problems Installing From DOS partition.

12. Newbie having trouble doing the "Samba"

13. HELP: Connectivity between DOS/DOS and DOS/Linux