Re: [resubmit] gdb.base, r*.exp thru w*.exp

[Joel Brobecker <brobecker@adacore.com>]

On Thu, May 20, 2010 at 01:29:47PM -0700, Michael Snyder wrote:
> Incorporating suggestions and changes, including Joel's new function.
> 

> 2010-05-20  Michael Snyder  <msnyder@vmware.com>
> 
> 	* gdb.base/relational.exp: Replace send_gdb with gdb_test.
> 	* gdb.base/relocate.exp: Replace send_gdb with gdb_test.
> 	* gdb.base/return.exp: Replace send_gdb with gdb_test.
> 	* gdb.base/return2.exp: Replace send_gdb with gdb_test.
> 	* gdb.base/sepdebug.exp: Replace send_gdb with gdb_test.
> 	* gdb.base/setvar.exp: Replace send_gdb with gdb_test.
> 	* gdb.base/shlib-call.exp: Replace send_gdb with gdb_test.
> 	* gdb.base/shreloc.exp: Replace send_gdb with gdb_test.
> 	* gdb.base/sigall.exp: Replace send_gdb with gdb_test.
> 	* gdb.base/sigbpt.exp: Replace send_gdb with gdb_test.
> 	* gdb.base/signull.exp: Replace send_gdb with gdb_test.
> 	* gdb.base/so-impl-ld.exp:Replace send_gdb with gdb_test.
> 	* gdb.base/varargs.exp: Replace send_gdb with gdb_test.
> 	* gdb.base/volatile.exp: Replace send_gdb with gdb_test.
> 	* gdb.base/watch_thread_num.exp: Replace send_gdb with gdb_test.
> 	* gdb.base/watchpoint.exp: Replace send_gdb with gdb_test.
> 	* gdb.base/whatis.exp: Replace send_gdb with gdb_test.

Looks good overall. It's a long patch, so I only skimmed it.
Just a couple of comments...


> -set name "set catch vfork, never expected to trigger"
> -send_gdb "catch vfork\n"
> -
>  # If we are on HP-UX 10.20, we expect an error message to be
>  # printed if we type "catch vfork" at the gdb gdb_prompt.  This is
>  # because on HP-UX 10.20, we cannot catch vfork events.
>  
> -if [istarget "hppa*-hp-hpux10.20"] then {
> -    gdb_expect {
> -	-re "Catch of vfork events not supported on HP-UX 10.20..*$gdb_prompt $"
> -		{pass $name}
> -	-re "$gdb_prompt $"
> -		{fail $name}
> -	timeout {fail "(timeout) $name"}
> +set name "set catch vfork, never expected to trigger"
> +gdb_test_multiple "catch vfork" $name {
> +    -re "Catch of vfork events not supported on HP-UX 10.20.*" {
> +	pass $name
>      }
> -} else {
> -    gdb_expect {
> -	-re "Catchpoint \[0-9\]* .vfork..*$gdb_prompt $"
> -		{pass $name}
> -	-re "Catch of vfork not yet implemented.*$gdb_prompt $"
> -		{pass $name}
> -	-re "$gdb_prompt $"
> -		{fail $name}
> -	timeout {fail "(timeout) $name"}
> +    -re "Catchpoint \[0-9\]* .vfork.*$gdb_prompt $" {
> +	pass $name
> +    }
> +    -re "Catch of vfork not yet implemented.*$gdb_prompt $" {
> +	pass $name
>      }
>  }

You slightly changed the implementation, but I think that in practice
this is going to be the same, so I'm OK with that.

> +gdb_test "break 999" "No line 999 in file .*" \
> +    pass "break on non-existent source line"
       ^^^^ ???
> +gdb_test_multiple "break $bp_location1" \
> +    "set to-be-silent break bp_location1" {
> +	-re "Breakpoint (\[0-9\]*) at .*, line $bp_location1.*$gdb_prompt $" {
> +	    pass "set to-be-silent break bp_location1"
> +	}

How about using test_gdb in this case?


-- 
Joel