Re: [rfa/testsuite] advance.exp, until.exp: accommodate different return lines

[Elena Zannoni <ezannoni at redhat dot com>]

Michael Elizabeth Chastain writes:
 > This patch fixes a bug in gdb.base/advance.exp and gdb.base/until.exp.
 > I analyzed this bug here:
 > 
 >   http://sources.redhat.com/ml/gdb-patches/2003-01/msg00621.html
 > 

Wow, somehow I missed your reply. Makes sense, I didn't see it because
I have used the old gcc.

 > Briefly:
 > 
 >   the program under test calls 'func(c);'.
 >   the user types in the 'advance' command
 >   gdb runs to the return breakpoint in the calling frame
 > 
 > The source line of the return breakpoint might be on the same line as
 > 'func(c)', or it might be on the next line.  This depends on whether
 > the machine code has any instructions after the 'CALL' instruction,
 > such as stack cleanup instructions.  This is ABI and platform dependent.
 > 
 > So I enhanced the tests to recognize both responses from gdb.
 > 
 > Testing: my usual testbed: native i686-pc-linux-gnu; gcc 2.95.3 and gcc
 > 3.2.1 (and some more gcc's); dwarf-2 and stabs+.  This fixes 5 FAIL's in
 > all the gcc 3.X configurations.  All tests now PASS in both advance.exp
 > and until.exp.
 > 
 > Elena Z wrote these tests, so I am hoping that she can approve
 > this patch (or recommend it to Fernando N for approval).
 > 
 > OK to apply?
 > 

Sure, thanks

elena

 > Michael C
 > 
 > 2003-02-01  Michael Chastain  <mec@shout.net>
 > 
 > 	* gdb.base/advance.c (marker1): New marker function.
 > 	* gdb.base/advance.exp: When the 'advance' command lands on the
 > 	return breakpoint, it can legitimately stop on either the
 > 	current line or the next line.  Accommodate both outcomes.
 > 	* gdb.base/until.exp: Likewise.
 > 
 > Index: advance.c
 > ===================================================================
 > RCS file: /cvs/src/src/gdb/testsuite/gdb.base/advance.c,v
 > retrieving revision 1.1
 > diff -u -r1.1 advance.c
 > --- advance.c	15 Jan 2003 14:25:11 -0000	1.1
 > +++ advance.c	3 Feb 2003 05:41:42 -0000
 > @@ -29,6 +29,10 @@
 >    x = 4;
 >  }
 >  
 > +void marker1 ()
 > +{
 > +}
 > +
 >  int
 >  main ()
 >  {
 > @@ -38,6 +42,7 @@
 >    b = 3;    /* advance this location */
 >      
 >    func (c); /* stop here after leaving current frame */
 > +  marker1 (); /* stop here after leaving current frame */
 >    func3 (); /* break here */
 >    result = bar (b + foo (c));
 >    return 0; /* advance malformed */
 > Index: advance.exp
 > ===================================================================
 > RCS file: /cvs/src/src/gdb/testsuite/gdb.base/advance.exp,v
 > retrieving revision 1.2
 > diff -u -r1.2 advance.exp
 > --- advance.exp	20 Jan 2003 15:40:07 -0000	1.2
 > +++ advance.exp	3 Feb 2003 05:41:42 -0000
 > @@ -63,8 +63,12 @@
 >  # Verify that "advance <funcname>" when funcname is NOT called by the current
 >  # frame, stops at the end of the current frame.
 >  #
 > +# gdb can legitimately stop on either the current line or the next line,
 > +# depending on whether the machine instruction for 'call' on the current
 > +# line has more instructions after it or not.
 > +#
 >  gdb_test "advance func3" \
 > -	"in main.*func \\(c\\).*stop here after leaving current frame..."\
 > +	"(in main|).*(func \\(c\\)|marker1 \\(\\)).*stop here after leaving current frame..."\
 >  	"advance function not called by current frame"
 >  
 >  # break at main again
 > Index: until.exp
 > ===================================================================
 > RCS file: /cvs/src/src/gdb/testsuite/gdb.base/until.exp,v
 > retrieving revision 1.2
 > diff -u -r1.2 until.exp
 > --- until.exp	20 Jan 2003 15:40:07 -0000	1.2
 > +++ until.exp	3 Feb 2003 05:41:43 -0000
 > @@ -76,6 +76,6 @@
 >  # stop at main, the caller, where we put the 'guard' breakpoint.
 >  #
 >  gdb_test "until marker3" \
 > -	"$hex in main.*argc.*argv.*envp.*at.*${srcfile}:82.*marker2 \\(43\\)." \
 > +	"($hex in |)main.*argc.*argv.*envp.*at.*${srcfile}:(82.*marker2 \\(43\\)|83.*marker3 \\(.stack., .trace.\\))." \
 >  	"until func, not called by current frame"
 >