RFC: fix `gdb -write' case

[Jan Kratochvil]

On Tue, 10 May 2011 04:02:47 +0200, Yao Qi wrote:
> Is this patch to fix bugs against gdb 6.8?

No.  In FSF gdb-6.8 it worked.  In FSF GDB HEAD it is broken.  It is
a regression.


> I am still unable to reproduce this problem on gdb 20110505-cvs.
> 
> $ ./gdb/gdb -nx -write ~/Work/gdb-write
> (gdb) set var s="b"
> evaluation of this expression requires the target program to be active

This worked with gdb-6.8, therefore you have reproduced the problem.


> (gdb) b main
> Breakpoint 1 at 0x80483b7: file gdb-write.c, line 3.
> (gdb) run
> Starting program: /home/yao/Work/gdb-write
> BFD: reopening /home/yao/Work/gdb-write: Text file busy
> 
> [This message above repeats]
> Breakpoint 1, main () at gdb-write.c:3
> 3	main(){return s[0]=='a';}
> (gdb) set var s="b"
> BFD: reopening /home/yao/Work/gdb-write: Text file busy
> ....
> Invalid cast.
> 
> Is it a separate issue?

This breaks for me a similar (but exactly the same) way with gdb-6.8.
I find this part as offtopic for this Bug.


Thanks,
Jan