[RFC] Set default value of detach_fork to off

[RFC] Set default value of detach_fork to off

[Hui Zhu]

Hi guys,

Looks multi-inferior is more better than before.  I suggest set
default value of detach_fork to off.
Then when the user debug the multi-process program, they can debug the
inferior like multi-thread inferior.

What do you think about it?

Thanks,
Hui

2010-05-08  Hui Zhu  <teawater@gmail.com>

	* infrun.c (detach_fork): Change default value to 0.


---
 infrun.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

--- a/infrun.c
+++ b/infrun.c
@@ -109,7 +109,7 @@ int sync_execution = 0;
 static ptid_t previous_inferior_ptid;

 /* Default behavior is to detach newly forked processes (legacy).  */
-int detach_fork = 1;
+int detach_fork = 0;

 int debug_displaced = 0;
 static void

Re: [RFC] Set default value of detach_fork to off

[Mark Kettenis]

> From: Hui Zhu <teawater@gmail.com>
> Date: Sat, 8 May 2010 20:19:29 +0800
> 
> Hi guys,
> 
> Looks multi-inferior is more better than before.  I suggest set
> default value of detach_fork to off.
> Then when the user debug the multi-process program, they can debug the
> inferior like multi-thread inferior.
> 
> What do you think about it?

I think we should keep the current default.  While the ability to
debug both sides of a fork is a nice feature to have, I think most
people are only interested in debugging the process they started the
debugger on.  It is not unusual for programs to use system(3) to
execute system binaries, and I doubt people are interested in
debugging those.  Typically those binaries will be stripped anyway.
Setting detach_fork to off will only interfere.
> 
> 2010-05-08  Hui Zhu  <teawater@gmail.com>
> 
> 	* infrun.c (detach_fork): Change default value to 0.

Re: [RFC] Set default value of detach_fork to off

[Joel Brobecker]

> I think we should keep the current default.  While the ability to
> debug both sides of a fork is a nice feature to have, I think most
> people are only interested in debugging the process they started the
> debugger on.  It is not unusual for programs to use system(3) to
> execute system binaries, and I doubt people are interested in
> debugging those.  Typically those binaries will be stripped anyway.
> Setting detach_fork to off will only interfere.

Agreed.

-- 
Joel