`interpreter mi' with -noasync

`interpreter mi' with -noasync

[Nick Roberts]

The `interpreter mi' command works fine as follows:

nick:{nick}63: gdb -n -ann=2 ~/myprog
GNU gdb 5.3.90_2003-08-22-cvs
Copyright 2003 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for details.
This GDB was configured as "i586-pc-linux-gnu"...
breakpoints-invalid

^Z^Zframes-invalid


^Z^Zpre-prompt
(gdb) 
^Z^Zprompt
interp mi "sho hei"

^Z^Zpost-prompt
&"sho hei\n"
^done,value="23"
(gdb) 

^Z^Zpre-prompt
(gdb) 
^Z^Zprompt


But when used with -noasync it seems to get a bit confused:


nick:{nick}61: gdb -q -n -ann=2 -noasync ~/myprog

^Z^Zbreakpoints-invalid

^Z^Zframes-invalid

^Z^Zpre-prompt
(gdb) 
^Z^Zprompt
interp mi "sho hei"

^Z^Zpost-prompt
&"sho hei\n"
^done,value="23"
(gdb) 
~"\n\032\032pre-"
~"prompt\n"
(gdb) 
^Z^Zprompt


Is that appropriate behaviour or shall I file a bug report?

Nick

Re: `interpreter mi' with -noasync

[Andrew Cagney]

> But when used with -noasync it seems to get a bit confused:

That's not suprizing.  There has never been any intent for MI to work 
with anything but async.

Andrew


> nick:{nick}61: gdb -q -n -ann=2 -noasync ~/myprog
> 
> ^Z^Zbreakpoints-invalid
> 
> ^Z^Zframes-invalid
> 
> ^Z^Zpre-prompt
> (gdb) 
> ^Z^Zprompt
> interp mi "sho hei"
> 
> ^Z^Zpost-prompt
> &"sho hei\n"
> ^done,value="23"
> (gdb) 
> ~"\n\032\032pre-"
> ~"prompt\n"
> (gdb) 
> ^Z^Zprompt
> 
> 
> Is that appropriate behaviour or shall I file a bug report?
> 
> Nick
> 

Re: `interpreter mi' with -noasync

[Nick Roberts]

 > > But when used with -noasync it seems to get a bit confused:
 > 
 > That's not suprizing.  There has never been any intent for MI to work 
 > with anything but async.

I use -noasync with Fortran programs and on a Windows platform (NT, at least)
because, otherwise, Emacs receives the annotations from GDB in the wrong
order. -async doesn't seem to be asynchronous from the users point of view i.e
a second GDB command does not start execution until the first one has
completed.  Although I don't understand what -noasync does, I use it because
seems to do the `right thing'.

Is this use surprising or necessary? Could I get by without -noasync? (I just
want to receive output from GDB in its proper order).

Nick

Re: `interpreter mi' with -noasync

[Andrew Cagney]

>  > > But when used with -noasync it seems to get a bit confused:
>  > 
>  > That's not suprizing.  There has never been any intent for MI to work 
>  > with anything but async.
> 
> I use -noasync with Fortran programs and on a Windows platform (NT, at least)
> because, otherwise, Emacs receives the annotations from GDB in the wrong
> order. -async doesn't seem to be asynchronous from the users point of view i.e
> a second GDB command does not start execution until the first one has
> completed.  Although I don't understand what -noasync does, I use it because
> seems to do the `right thing'.
> 
> Is this use surprising or necessary? Could I get by without -noasync? (I just
> want to receive output from GDB in its proper order).

There's two things:
- event loop
- async interaction
Async is only fully implemented with the remote target (target 
remote-async).  For the user, asynchronous behavior also only occures 
when explicitly specified vis:
	(gdb) run &
After that development is somewhat stalled - native targets need to be 
switched to the event loop.

What exactly goes wrong with the output?

Andrew