On Thursday, May 23, 2002, at 07:20 PM, Andreas J. Koenig wrote:
>> On Thu, 23 May 2002, Ken Williams wrote:
>>> Schwern said:
>>>> I'm not quite sure I get this one. What's writing the YAML?
>>>
>>> It would be written by YAML as part of the 'make dist' ('Build dist',
>>> actually), from options given in the Build.PL file.
>
> Not the Makefile.PL?
It could be written by either the Makefile.PL or the Build.PL,
depending on whether EU::MM or Module::Build is running the
build/dist process. I probably won't contribute the necessary
EU::MM code myself, but it would be fairly simple for someone
with enough understanding of the MM guts and enough time to do
it. I've got Module::Build code for it already.
>> Oh I remember, you want something which can be read quickly
>> and without
>> needing an eval().
>
>> But - if the Build.PL is there as human-readable source code, then the
>> argument about the generated metadata file being human-readable goes
>> away. The generated file (YAML or whatever) would be read only by
>> machine, a human would surely go straight to the Build.PL. Or have I
>> got this the wrong way round?
>
> Depends on which human you mean. The author of the XXX.PL file will
> prefer to edit hist XXX.PL file. The user of module Foo will prefer to
> read either the canonical format or the output of some tool which in
> turn will read the canonical format. The author might prefer to not
> even distribute his XXX.PL some day.
Right. And if non-perl people are going to be dealing with perl
module distributions, I'm sure they'd find a simple meta-data
file much more readable than a XXX.PL file.
-Ken
