[RFC] xfail a break.exp test for remote targets

[RFC] xfail a break.exp test for remote targets

[Fred Fish]

I just posted a fairly long analysis to the gdb-bug list about how
remote targets can miss breakpoints internal to a line or fail to
print the breakpoint info for a breakpoint at the end of a stepping
range, when single stepping over a source line.  This patch will fix
the testsuite failure until the remote/native difference in resuming
execution in the single step case can be resolved.

Comments?

-Fred


2002-03-02  Fred Fish  <fnf@redhat.com>

	* gdb.base/break.exp: Setup xfail for remote targets for the
	"step onto breakpoint" due to the way remote targets handle
	breakpoints at the end of a stepping range.

Index: gdb.base/break.exp
===================================================================
RCS file: /cvs/cvsfiles/devo/gdb/testsuite/gdb.base/break.exp,v
retrieving revision 1.65
diff -c -p -r1.65 break.exp
*** break.exp	2002/01/13 00:32:27	1.65
--- break.exp	2002/03/02 18:50:04
*************** gdb_test "break +1" \
*** 222,227 ****
--- 222,237 ----
  
  # Check to see if breakpoint is hit when stepped onto
  
+ # This will always fail for remote targets that step over the entire
+ # source line in one operation local to the target because gdb won't
+ # have breakpoints inserted for the first single step and thus won't
+ # print the breakpoint at the location it stops at after stepping.
+ # Note that removing the current stopped-at breakpoint (number 5) would
+ # cause this test to succeed.  Perhaps we should try it both ways.
+ if [is_remote target] {
+     setup_xfail "*-*-*"
+ }
+ 
  gdb_test "step" \
      ".*Breakpoint \[0-9\]+, main \\(argc=.*, argv=.*, envp=.*\\) at .*$srcfile:86.*86\[\t \]+return argc;" \
      "step onto breakpoint"

Re: [RFC] xfail a break.exp test for remote targets

[Michael Snyder]

Fred Fish wrote:
> 
> I just posted a fairly long analysis to the gdb-bug list about how
> remote targets can miss breakpoints internal to a line or fail to
> print the breakpoint info for a breakpoint at the end of a stepping
> range, when single stepping over a source line.  This patch will fix
> the testsuite failure until the remote/native difference in resuming
> execution in the single step case can be resolved.
> 
> Comments?

No, just disable the 'e' packet -- it's broken.

> 
> -Fred
> 
> 2002-03-02  Fred Fish  <fnf@redhat.com>
> 
>         * gdb.base/break.exp: Setup xfail for remote targets for the
>         "step onto breakpoint" due to the way remote targets handle
>         breakpoints at the end of a stepping range.
> 
> Index: gdb.base/break.exp
> ===================================================================
> RCS file: /cvs/cvsfiles/devo/gdb/testsuite/gdb.base/break.exp,v
> retrieving revision 1.65
> diff -c -p -r1.65 break.exp
> *** break.exp   2002/01/13 00:32:27     1.65
> --- break.exp   2002/03/02 18:50:04
> *************** gdb_test "break +1" \
> *** 222,227 ****
> --- 222,237 ----
> 
>   # Check to see if breakpoint is hit when stepped onto
> 
> + # This will always fail for remote targets that step over the entire
> + # source line in one operation local to the target because gdb won't
> + # have breakpoints inserted for the first single step and thus won't
> + # print the breakpoint at the location it stops at after stepping.
> + # Note that removing the current stopped-at breakpoint (number 5) would
> + # cause this test to succeed.  Perhaps we should try it both ways.
> + if [is_remote target] {
> +     setup_xfail "*-*-*"
> + }
> +
>   gdb_test "step" \
>       ".*Breakpoint \[0-9\]+, main \\(argc=.*, argv=.*, envp=.*\\) at .*$srcfile:86.*86\[\t \]+return argc;" \
>       "step onto breakpoint"