Jonas Brømsø Nielsen wrote:
On 03/09/2008, at 16.41, Gabor Szabo wrote:
On Wed, Sep 3, 2008 at 5:08 PM, Sisyphus <[EMAIL PROTECTED]>
wrote:
I tried to find the original post where Graham expressed the view quoted
above ... but failed to do so.
Seems to me that Graham's argument also applies to situations where
'perl
Makefile.PL' went fine, but the 'make' process failed. That is, if
the make
process fails "then you cannot know if the tests fail or not, hence an
UNKNOWN situation".
Maybe there should be more possible report values?
FAIL and PASS can be kept for when the actual tests fail or pass
but there should be also
MAKE_FAIL fr when the make phase failes
MAKEPL_FAIL fails when Makefile.PL exits with error code
and I leave it open to the discussion what to do when
Makefile.PL exits with 0 but there is no Makefile ?
Returning 0 without having accomplished the primary goal of the
Makefile.PL, to generate a make file, seems like a serious error/bug.
I expect that scenario to be a possibility, but hopefully not a common one.
>
>> Gabor
>
> jonasbn
>
>
> --
> pauseid: JONASBN
> email: [EMAIL PROTECTED]
> blog: http://use.perl.org/~jonasbn/journal/
I've been sitting on the side lines watching this thread wondering when
you were going to come up with another change the module authors have to
make to get the right results into cpan test reports.
I sincerely hope I am misunderstanding the comment "Returning 0 without
having accomplished the primary goal of the Makefile.PL, to generate a
make file, seems like a serious error/bug" because I only changed my
Makefile.PL to do exactly that (exit with 0 without generating a
Makefile) at this lists advice. I also included build_requires,
configure_requires etc etc etc in the hope of getting successful
reports. Please tell me I've misunderstood the above comment because I
don't think it is fair on module authors to continually expect them to
make changes that don't stand the test of time.
Martin
