From avodonosov at yandex.ru  Wed Dec 28 20:53:38 2011
From: avodonosov at yandex.ru (Anton Vodonosov)
Date: Thu, 29 Dec 2011 00:53:38 +0400
Subject: [anaphora-devel] Test ANAPHORA-TEST::SCOND.3 fails on Windows.
Message-ID: <553101325105618@web72.yandex.ru>

Hello.

I decided to setup a  test environment where CL libraries may be tested on different Lisps.
Anaphora is already included, because it is one of the most often downloaded Quicklisp projects.

Results maybe found here 
http://common-lisp.net/project/cl-test-grid/pivot_ql-lib_lisp.html (the ok/fail links refer to 
the test logs).

As you can see, on Windows implementations one of the tests fails:

  Test ANAPHORA-TEST::SCOND.3 failed
   Form: (ANAPHORA:SCOND
            ((SYMBOL-VALUE '*DEFAULT-PATHNAME-DEFAULTS*)
             (LET ((ANAPHORA-TEST::TMP ANAPHORA:IT))
               (UNWIND-PROTECT (PROGN (SETF ANAPHORA:IT (TRUENAME "/tmp/")) 
                                                        (NAMESTRING *DEFAULT-PATHNAME-DEFAULTS*)) 
                   (SETF ANAPHORA:IT ANAPHORA-TEST::TMP)))))
Expected value: "/tmp/"
Actual value: "c:/tmp/".
 
If you like the idea of continuous testing, you may want to fix this, so that this failure does 
not obscure real regressions which may happen in future.

Best regards,
- Anton



From martin at lispworks.com  Fri Dec 30 10:25:40 2011
From: martin at lispworks.com (Martin Simmons)
Date: Fri, 30 Dec 2011 10:25:40 GMT
Subject: [anaphora-devel] Test ANAPHORA-TEST::SCOND.3 fails on Windows.
In-Reply-To: <553101325105618@web72.yandex.ru> (message from Anton Vodonosov
	on Thu, 29 Dec 2011 00:53:38 +0400)
References: <553101325105618@web72.yandex.ru>
Message-ID: <201112301025.pBUAPe9S018255@higson.cam.lispworks.com>

>>>>> On Thu, 29 Dec 2011 00:53:38 +0400, Anton Vodonosov said:
> 
> Hello.
> 
> I decided to setup a  test environment where CL libraries may be tested on different Lisps.
> Anaphora is already included, because it is one of the most often downloaded Quicklisp projects.
> 
> Results maybe found here 
> http://common-lisp.net/project/cl-test-grid/pivot_ql-lib_lisp.html (the ok/fail links refer to 
> the test logs).
> 
> As you can see, on Windows implementations one of the tests fails:
> 
>   Test ANAPHORA-TEST::SCOND.3 failed
>    Form: (ANAPHORA:SCOND
>             ((SYMBOL-VALUE '*DEFAULT-PATHNAME-DEFAULTS*)
>              (LET ((ANAPHORA-TEST::TMP ANAPHORA:IT))
>                (UNWIND-PROTECT (PROGN (SETF ANAPHORA:IT (TRUENAME "/tmp/")) 
>                                                         (NAMESTRING *DEFAULT-PATHNAME-DEFAULTS*)) 
>                    (SETF ANAPHORA:IT ANAPHORA-TEST::TMP)))))
> Expected value: "/tmp/"
> Actual value: "c:/tmp/".
>  
> If you like the idea of continuous testing, you may want to fix this, so that this failure does 
> not obscure real regressions which may happen in future.

Moreover, (TRUENAME "/tmp/") gives an error on stock Windows because the
directory doesn't exist.

I think the test should use its own variable, not *DEFAULT-PATHNAME-DEFAULTS*.
This is also safer if (SETF ANAPHORA:IT ANAPHORA-TEST::TMP) fails for some
reason.

__Martin