Ken Williams wrote:
> I noticed the subclass just wraps this method in an eval() and returns
> false when the eval() fails. Has that been a problem?
I believe that is the only way to prevent the build from dying if there is no C
compiler (i.e. I think this is what M::B should be doing internally).
> On my OS X machine with perl 5.8.6 and M::B 0.27_03 I get some funny
> output from running 'perl Build.PL':
>
> ---------------------------
> Creating custom builder _build/lib/version/Builder.pm in _build/lib/
> version
> Checking whether your kit is complete...
> Looks good
>
> Checking prerequisites...
> Looks good
>
> Checking whether your kit is complete...
> Looks good
>
> Checking prerequisites...
> Looks good
>
> Creating new 'Build' script for 'version' version '0.56_03'
> ---------------------------
>
> Looks like that's because you're calling new() twice. That's to work
> around some problem though, right?
Yeah, I'm not sure the best way around that. I need to call new() once to get
the commandline argument and then build up the real build arguments based on
whether the commandline "--perl_only" was passed in or if there is no C compiler
available.
I asked at one point about being able to take an existing M::B object and update
all of the configuration based on information gathered during the initial object
creation. I believe the reponse was something along the line of "that would be
nice but for the moment new() is run-once" (I'm paraphrasing).
> I also get tons of verbose output from 'Build test', do you see that too?
Yeah, I finally figured out to look at Test::Harness for how to determine
whether the tests are being run with verbose enabled (duh). Those test section
headings will vanish in the next release.
John
--
John Peacock
Director of Information Research and Technology
Rowman & Littlefield Publishing Group
4720 Boston Way
Lanham, MD 20706
301-459-3366 x.5010
fax 301-429-5747
|
