This is the mail archive of the mailing list for the Archer 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: Problem with merge

Rick Moseley wrote:
Hi all,

I have had a bit of a problem with the merge, particularly with having it run the test suite after a particular merge. I thought I could point the finger at Phil's patch to breakpoint.c, but it seems he has wiggled off the hook as he has backed his patch out and the error still occurs. I thought I would give a few details here and if anyone sees anything familiar please provide some input.

The error that occurs is that the test suite barfs at the "break.exp" test by looping forever and finally consuming all of memory until the system locks up and has to have its power cycled. These tests are being run on an x86_64 running Fedora 10, although it has also failed on Fedora 8. Here is the output from running the test suite:

runtest break.exp --verbose --verbose break.exp

Looking at the test result output from gdb.out, I find the simplest symptomatic reproducer to be:

[pmuldoon@localhost rick_issue_obj]$ cat main.c

#include <stdio.h>

void print_hello_world ()
printf ("Hello world!\n");

int main ()
print_hello_world ();

[pmuldoon@localhost issue_obj]$ gcc -g3 -O0 main.c -o main_example

[pmuldoon@localhost issue_obj]$ ./gdb/gdb main_example

(gdb) b main
Reading in symbols for main.c...done.
Reading in symbols for main.c...done.
Breakpoint 1 at 0x4005c0: file main.c, line 10.

(gdb) r
Starting program: /home/pmuldoon/issue_obj/main_example

Breakpoint 1, main () at main.c:10
10      print_hello_world ();
Current language:  auto; currently c++

(gdb) p print_hello_world()
Hello world!

... and GDB never returns (well "never" being more than 1 hour in this test).

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