Re: print long(s) within gdb

[DJBARROW at de dot ibm dot com]

Hi all,
Andreas Herrmann has found this bug it shows up on gdb-5.0 on intel too &
the repository from feb26.
It thinks a long long in certain situations is an int.

Have you got around to looking at my gdb patches yet ?, I mailed the
original ones over six months ago &
have spent over 200 hours making recommended modifications, redhat is even
shipping the 64 bit s390  beta
code for a while.

 Maintainers who procrastinate or need to check absolutely everything are
stopping progress aren't doing their job effectively.
 I think you should be accepting these unless you can find some real bugs
in the code which causes
failed builds on other platfroms if you haven't got a chance to look at the
code.


D.J. Barrow Gnu/Linux for S/390 kernel developer
eMail: djbarrow@de.ibm.com,barrow_dj@yahoo.com
Phone: +49-(0)7031-16-2583
IBM Germany Lab, Schönaicherstr. 220, 71032 Böblingen

---------------------- Forwarded by Denis Joseph Barrow/Germany/Contr/IBM
on 24.04.2001 19:17 ---------------------------

Andreas Herrmann
23.04.2001 15:14

To:   Denis Joseph Barrow/Germany/Contr/IBM@IBMDE
cc:
From: Andreas Herrmann/Germany/IBM@IBMDE
Subject:  Re: print long(s) within gdb  (Document link: Denis Joseph
      Barrow)


Hi Denis,

because of this f... Lotus Notes under Wine/Linux I couldn't detach
your version of gdb. But I've taken the gdb-5.0-0.s390.rpm from
ESA22_26 from tuxmaker to verify the problem on s390 with new(est) gdb
version.

First:
print sizeof(long) yields 4 on s390-gdb-version and 8 on s390x-gdb-version
(as it should)
print sizeof(long long) yields 8 for both gdb-versions

Second:
The problem still remains. If code was generated using "-gstabs", gdb is
not able to print
64 bit integers in the right way (only the first four bytes are used).

I attach a sample prog. Just compile it with "gcc -gstabs tst.c" and run it
with gdb.
If you perform "print c" or "print d" (after assignments of values) within
gdb on s390 or s390x, you receive the values 0.
But in fact they aren't. Using "x /g &c" and "x /g &d" show, c and d aren't
0.
Compiling the prog with "gcc -ggdb tst.c" on s390(x) works correct.
On Intel it works fine in both cases.


(See attached file: tst.c)



--
Linux for zSeries Development
Tel :  +49-7031-16-4640
Notes mail :  Andreas Herrmann/GERMANY/IBM@IBMDE
email :  aherrman@de.ibm.com


Denis Joseph Barrow
04/20/2001 06:29 PM

To:   Andreas Herrmann/Germany/IBM@IBMDE
cc:
From: Denis Joseph Barrow/Germany/Contr/IBM@IBMDE
Subject:  Re: print long(s) within gdb  (Document link: Andreas Herrmann)

Hi Andreas
Firstly I'd check how old the gdb you have is & tell me where you got it.
 If it is  prior to Jan 2001 it probably has serious problems, I missed a
few defines.


To  check print sizeof(long)

Here is one in  which the sizeof should be 8,


let me know if this fixes you up.



D.J. Barrow Gnu/Linux for S/390 kernel developer
eMail: djbarrow@de.ibm.com,barrow_dj@yahoo.com
Phone: +49-(0)7031-16-2583
IBM Germany Lab, Schönaicherstr. 220, 71032 Böblingen

tst.c