Re: [PATCH: gdb/mi + doco] -var-update

[Nick Roberts <nickrob at snap dot net dot nz>]

 > This is basically OK.  A couple of things:
 ...

 > Stray colon there.
 ...

 > Both here and in mi-cmd-var.c, please list these in the changelog:
 > 	* mi/mi-cmds.h (mi_no_values, mi_with_values, mi_simple_values)
 > 	(mi_all_values): New declarations.
 > 	* mi/mi-cmds.c (mi_no_values, mi_with_values, mi_simple_values)
 > 	(mi_all_values): New string constants.
 ...

OK. I've got that.

 > You've replaced "--all-values" in the source with "--with-values" here. 
 > Surely that's a bug?
 >
 > I don't remember the entire outcome of your discussion with Eli, but I
 > find the idea of having --with-values sometimes and --all-values other
 > times a bit confusing.  I went trying to figure out which meant what
 > and that's when I noticed this problem.

I don't mind that much which way its done.  These aren't commands that get
typed in by the user.  If you and Eli decide then I'll implement it, but
please see the comment below about "0"/"1" behaviour,

 > > !   if (argc == 1)
 > > !     name = argv[0];
 > > !   else
 > > !     name = (argv[1]);
 > 
 > Stray parentheses.

I don't follow.

 > > !   if (argc == 2)
 > > !     {
 > > !       if (strcmp (argv[0], "0") == 0
 > > ! 	  || strcmp (argv[0], mi_no_values) == 0)
 > > ! 	print_values = PRINT_NO_VALUES;
 > > !       else if (strcmp (argv[0], "1") == 0
 > > ! 	       || strcmp (argv[0], mi_with_values) == 0)
 > > ! 	print_values = PRINT_ALL_VALUES;
 > > !       else
 > > ! 	error (_("Unknown value for PRINT_VALUES: \
 > > ! must be: 0 or \"%s\", 1 or \"%s\""), mi_no_values, mi_with_values);
 > > !     }
 > > !   else
 > > !     print_values = PRINT_NO_VALUES;
 > >   
 > >     /* Check if the parameter is a "*" which means that we want
 > >        to update all variables */
 > 
 > IIRC, you added the "0"/"1" compatibility to -var-list-children to make
 > life easier for Apple.  Is that right?  If so, do they need it here
 > also, or can we get away with just --all-values?  I've no real
 > objection to the 0/1, but they're a bit ugly.

 I think I originally copied the "0"/"1" arguments for -var-list-children
from existing behaviour for -stack-list-locals.  I also think that Apple
had already done something similar but different (looking through the e-mails
their arguments had reverse the order: SHOW-VALUE VAROBJ-HANDLE).  If these
are removed then I need to keep "-all-values" for -var-list-children for
backward compatiblity (GDB 6.1 to 6.3?).

 > This paragraph must be in the wrong place.  It's mostly duplicating the
 > paragraph above it.
 ...

 Whoops.  I've diffed the wrong file.  I'll re-submit it with the patch for
 the source when I know what syntax you want.

Nick