RFA: mi: correct loop printing args or locals

RFA: mi: correct loop printing args or locals

[Jim Blandy]

Without this patch, on platforms using STABS, MI's
-stack-list-arguments command describes frames for `callee1' in
`gdb/testsuite/gdb.mi/basics.c' like this:

frame={level="3",args=[name="intarg",name="strarg",name="fltarg",name="fltarg"]}
Note that `fltarg' appears twice.  This is because there is both a
PSYM (parameter symbol) stab and an LSYM (local symbol) stab for
`fltarg', which I'm pretty sure is the standard (and correct) STABS
way to describe such functions.  Anyway, the loop is clearly incorrect
for blocks that contain a mix of arguments and non-arguments.

2002-02-03  Jim Blandy  <jimb@redhat.com>

	* mi-cmd-stack.c (list_args_or_locals): Move declaration of
	print_me inside the loop body, so it gets re-initialized every
	iteration.  The cases for the different symbol kinds leave
	print_me unchanged if they don't want the symbol printed.

Index: gdb/mi/mi-cmd-stack.c
===================================================================
RCS file: /cvs/cvsfiles/devo/gdb/mi/mi-cmd-stack.c,v
retrieving revision 1.11
diff -c -r1.11 mi-cmd-stack.c
*** gdb/mi/mi-cmd-stack.c	2001/10/15 22:15:56	1.11
--- gdb/mi/mi-cmd-stack.c	2002/02/04 01:43:35
***************
*** 216,222 ****
    struct block *block;
    struct symbol *sym;
    int i, nsyms;
-   int print_me = 0;
    static struct ui_stream *stb = NULL;
  
    stb = ui_out_stream_new (uiout);
--- 216,221 ----
***************
*** 229,234 ****
--- 228,235 ----
      {
        ALL_BLOCK_SYMBOLS (block, i, sym)
  	{
+           int print_me = 0;
+ 
  	  switch (SYMBOL_CLASS (sym))
  	    {
  	    default:

Re: RFA: mi: correct loop printing args or locals

[Andrew Cagney]

> frame={level="3",args=[name="intarg",name="strarg",name="fltarg",name="fltarg"]}
> Note that `fltarg' appears twice.  This is because there is both a
> PSYM (parameter symbol) stab and an LSYM (local symbol) stab for
> `fltarg', which I'm pretty sure is the standard (and correct) STABS
> way to describe such functions.


I don't know about correct.  However it is standard.  Think of the 
number of times you've received a debug-info sorting patch that 
accidently re-ordered the stab info :-)

 > Anyway, the loop is clearly incorrect

> for blocks that contain a mix of arguments and non-arguments.


Yes, fine.  I take it you're having fun with the MI testsuite.
Feel free to just commit fixes to the other bugs you find.

Andrew