test for excess errors

[naveen yadav]

Hi Ian,

I have check the testcode.


      1 /* This is from PR c/25892.  See Wpointer-sign.c for more details.  */
      2
      3 /* { dg-options "-Wall" } */
      4
      5 void foo(unsigned long* ulp); /* { dg-message "note: expected
'\[^\n'\]*' but argument is of type '\[^\n'\]*'" "note: expected" {
target *-*-* } 5 } */
      6
      7
      8 void bar(long* lp) {
      9   foo(lp); /* { dg-warning "differ in signedness" } */
     10 }


so why this fail

On Tue, Aug 17, 2010 at 7:35 PM, Ian Lance Taylor <iant@google.com> wrote:
> paladin tripathi <paladin.tripathi@gmail.com> writes:
>
>> I am trying to test gcc testsuite on target in some test cases i am
>> getting the following logs
>>
>> Executing on host: arm-none-linux-gnueabi-gcc
>> ./gcc.dg/Wpointer-sign-Wall.c   -Wall -S  -o Wpointer-sign-Wall.s
>> (timeout = 300)
>> pid is 15130 -15130
>> ./gcc.dg/Wpointer-sign-Wall.c: In function 'bar':
>> ./gcc.dg/Wpointer-sign-Wall.c:9: warning: pointer targets in passing
>> argument 1 of 'foo' differ in signedness
>> ./gcc.dg/Wpointer-sign-Wall.c:5: note: expected 'long unsigned int *'
>> but argument is of type 'long int *'
>> output is ./gcc.dg/Wpointer-sign-Wall.c: In function 'bar':
>> ./gcc.dg/Wpointer-sign-Wall.c:9: warning: pointer targets in passing
>> argument 1 of 'foo' differ in signedness
>> ./gcc.dg/Wpointer-sign-Wall.c:5: note: expected 'long unsigned int *'
>> but argument is of type 'long int *'
>>
>>  output is:
>> ./gcc.dg/Wpointer-sign-Wall.c: In function 'bar':
>> ./gcc.dg/Wpointer-sign-Wall.c:9: warning: pointer targets in passing
>> argument 1 of 'foo' differ in signedness
>> ./gcc.dg/Wpointer-sign-Wall.c:5: note: expected 'long unsigned int *'
>> but argument is of type 'long int *'
>>
>> PASS: gcc.dg/Wpointer-sign-Wall.c  (test for warnings, line 9)
>> FAIL: gcc.dg/Wpointer-sign-Wall.c (test for excess errors)
>>
>>
>> Could some one explain me what is this "test for excess errors"
>> is this run time error or compile time error
>
> DejaGNU, which is what the gcc testsuite uses, is verifying that there
> are no unexpected errors.  Failing the "test for excess errors" means
> that the test generated an unexpected error.
>
> In this case the problem appears to be that the test case does not
> expect the message at line 5.  When I look at the test case in current
> gcc sources, I see that the message on line 5 is expected.  So I would
> guess that you are using a newer gcc with an older testsuite, or some
> such mismatch.
>
> Ian
>