Re: Guile 2.0.6 - some tests fail

[Jan Synacek <jsynacek@redhat.com>]

On 10/06/2012 12:55 PM, Mark H Weaver wrote:
> Jan Synacek <jsynacek@redhat.com> writes:
>> when I run 'make check', some guile tests fail and I'm not sure why [1, 2].
>>
>> [1] http://jsynacek.fedorapeople.org/guile/check-guile.log
>> [2] http://jsynacek.fedorapeople.org/guile/check-guile-fails.log (only FAILs)
> 
> Quoting from [2] (but rearranging the order somewhat):
> 
>> XFAIL: chars.test: basic char handling: comparisons: char=? #\A #\a
>> XFAIL: chars.test: basic char handling: comparisons: char=? #\A #\B
>> XFAIL: chars.test: basic char handling: comparisons: char=? #\B #\A
>> XFAIL: chars.test: basic char handling: comparisons: char<? #\A #\A
>> XFAIL: chars.test: basic char handling: comparisons: char<? #\B #\A
>> XFAIL: chars.test: basic char handling: comparisons: char<=? #\B #\A
>> XFAIL: chars.test: basic char handling: comparisons: char>? #\A #\A
>> XFAIL: chars.test: basic char handling: comparisons: char>? #\A #\a
>> XFAIL: chars.test: basic char handling: comparisons: char>? #\A #\B
>> XFAIL: chars.test: basic char handling: comparisons: char>=? #\A #\a
>> XFAIL: chars.test: basic char handling: comparisons: char>=? #\A #\B
>> XFAIL: chars.test: basic char handling: comparisons: char-ci=? #\A #\B
>> XFAIL: chars.test: basic char handling: comparisons: char-ci=? #\B #\A
>> XFAIL: chars.test: basic char handling: comparisons: char-ci<? #\A #\A
>> XFAIL: chars.test: basic char handling: comparisons: char-ci<? #\A #\a
>> XFAIL: chars.test: basic char handling: comparisons: char-ci<? #\B #\A
>> XFAIL: chars.test: basic char handling: comparisons: char-ci<=? #\B #\A
>> XFAIL: chars.test: basic char handling: comparisons: char-ci>? #\A #\A
>> XFAIL: chars.test: basic char handling: comparisons: char-ci>? #\A #\a
>> XFAIL: chars.test: basic char handling: comparisons: char-ci>? #\A #\B
>> XFAIL: chars.test: basic char handling: comparisons: char-ci>=? #\A #\B
> 
> These are not failures in Guile itself, but rather inappropriate uses of
> 'expect-fail' in test cases where #f is the expected result.  You
> needn't worry about them, and I have fixed these test cases on the
> stable-2.0 branch.
> 
>> XFAIL: eval.test: map: documented?
>> XFAIL: eval.test: define set procedure-name: procedure-with-setter
>> XFAIL: goops.test: defining classes: define-class: bad init-thunk
>> XFAIL: list.test: append!: wrong argument: improper list and empty list
>> XFAIL: list.test: append!: wrong argument: improper list and list
>> XFAIL: list.test: append!: wrong argument: list, improper list and list
>> XFAIL: list.test: append!: wrong argument: circular list and empty list
>> XFAIL: list.test: append!: wrong argument: circular list and list
>> XFAIL: list.test: append!: wrong argument: list, circular list and list
> 
> These are expected failures, i.e. reminders to Guile developers to
> improve the implementation.  I just fixed the first three 'append!'
> cases (checking for improper lists) on the stable-2.0 branch, and will
> look into fixing the others at some point, but you needn't worry about
> them.
> 
>> FAIL: ftw.test: file-system-fold: test-suite (never enter)
>> FAIL: ftw.test: file-system-fold: test-suite/lib.scm (flat file)
>> FAIL: ftw.test: scandir: no select
>> FAIL: ports.test: %file-port-name-canonicalization: absolute canonicalization from ice-9
> 
> These are genuine problems that most users haven't run into AFAIK, and
> we should try to find the underlying cause.  Obviously they are all
> filesystem related, so the first question that comes to mind is: What
> filesystem type is your Guile source directory stored on?  Can you think
> of anything else unusual about your system that might affect filesystem
> behavior?
> 
> Thanks very much for bringing these problems to our attention.
> 
>      Mark
> 

Thank you for your answer.

I'm running ext4 on an LVM partition. No acls, no funky file permissions. I've
also tried turning selinux off, but that didn't help. I've gathered some
additional info about my filesystem [3]. Other than that, I can't think of
anything else that could be wrong/special in any way. If there's more info that
you need, just let me know.

[3] http://jsynacek.fedorapeople.org/guile/fs.info

Thank you,
-- 
Jan Synacek
Software Engineer, BaseOS team Brno, Red Hat