Oh, wait! CPAN doesn't do Build.PL IIRC. Can you replicate the error
with your svn checkout if you do
$ ./Build realclean
$ perl Makefile.PL
$ make
$ make test
I wrote all of this other stuff and then though of ^-- that, so try that
and then read the rest.
# from Jay Truesdale
# on Wednesday 02 August 2006 05:18 pm:
>All tests successful, 1 test and 12 subtests skipped.
>Files=21, Tests=713, 351 wallclock secs (180.87 cusr + 170.35 csys =
> 351.22 CPU)
Ok then. I'm drawing a blank on what generated the original "Argument
list too long" error. I'll have to wait until one of the wiser hackers
pokes at that, but I suspect CPAN. You might try
cpan> o conf
And see if you notice anything huge in that output. Also, if you
haven't tried it before:
cpan> clean Module::Build
cpan> look Module::Build
$ perl Build.PL
$ perl Build
$ ./Build test
$ exit
And even
cpan> clean Module::Build
cpan> test Module::Build
(where "cpan>" is the cpan shell prompt and "$" is a bash shell)
It might even be something of a "only happens on a fresh redhat box"
type of thing (we really need a good cross-platform testing
architecture, don't we.)
>Sometimes I see something like this: ok 57 - Makefile shouldn't exist
>Is this an error?
Nope. Just a success of something that shouldn't be there not being
there. ok means ok, but yes it would be okayer if it did not quite so
much read awkwardly.
Maybe "ok 57 - no stray Makefile"
$ ack "shouldn't exist"
t/compat.t
153: ok ! -e 'Makefile', "Makefile shouldn't exist";
185: ok ! -e 'Makefile', "Makefile shouldn't exist";
--Eric
--
The only thing that could save UNIX at this late date would be a new $30
shareware version that runs on an unexpanded Commodore 64.
--Don Lancaster (1991)
---------------------------------------------------
http://scratchcomputing.com
---------------------------------------------------
