gdb hang

gdb hang

[Sami Wagiaalla]

Hi Jan,

This should fix the gdb hang problem with open office:

	108568024684babfeb3876f76064ff9cfced827a

I was not able to get the reproducer to work the commit
includes a test case which causes the hang.

I still use find_pc_line, although more responsibly now :)

Is there a better way to find out the current line of execution ?
I have tried get_current_source_symtab_and_line but it seems to
always be off by 5 lines.

Sami

Re: gdb hang

[Jan Kratochvil]

On Thu, 03 Sep 2009 20:56:39 +0200, Sami Wagiaalla wrote:
> This should fix the gdb hang problem with open office:
>
> 	108568024684babfeb3876f76064ff9cfced827a

Great, thanks, going to push it.


> I was not able to get the reproducer to work the commit
> includes a test case which causes the hang.

Was the user-submitted reproducer reproducible for you and got it fixed?


> Is there a better way to find out the current line of execution ?
> I have tried get_current_source_symtab_and_line but it seems to
> always be off by 5 lines.

get_current_source_symtab_and_line is not right, it does change on `list main'.


+  int current_line = find_pc_line (get_frame_pc (get_current_frame ()), 0).line;

I see there you should use get_selected_frame() instead, shouldn't you?
After
(gdb) up
(gdb) up
get_current_frame() returns frame #0 while get_selected_frame() returns #2.

If you do not agree I will try to find a countercase whether I am right or not.



Thanks,
Jan

Re: gdb hang

[Sami Wagiaalla]

>> I was not able to get the reproducer to work the commit
>> includes a test case which causes the hang.
> 
> Was the user-submitted reproducer reproducible for you and got it fixed?
> 

No, I could not get it to work. I am still messing around with it.

> +  int current_line = find_pc_line (get_frame_pc (get_current_frame ()), 0).line;
> 
> I see there you should use get_selected_frame() instead, shouldn't you?
> 

Yes, makes sense. I will try it and push an update.

Sami

Re: gdb hang

[Jan Kratochvil]

On Thu, 03 Sep 2009 21:25:05 +0200, Sami Wagiaalla wrote:
>> Was the user-submitted reproducer reproducible for you and got it fixed?
>
> No, I could not get it to work. I am still messing around with it.

FYI reported by David Tardon as fixed (gdb-6.8.50.20090818-8.fc12.x86_64).


Thanks,
Jan