Open questions about integrating my Perl port
Aristotle Pagaltzis
I’ve dragged my feet on putting my port in the repo because of
a few issues on which I’m unsure how to best proceed.
Since I followed the CPAN distribution structure, what I have
created is more than a single file. So far, however, all of the
provided implementations are single files. Do I just stick the
directory in there in place of Stanis’ single-file port? How are
the tarballs on the project currently produced?
Also, as noted, I extracted the test data into JSON files, so
that the other languages can share it and only need to implement
a driver portion for each of the datasets. For the CPAN distro I
need this data in the tarball, so it must be included in some
form in the Perl subtree. How do we go about sharing the files?
Symlinks? (How) Does that work for Windows-based Subversion
users?
Regards,
--
Aristotle Pagaltzis // <http://plasmasturm.org/>
Steve Vinoski
> I've dragged my feet on putting my port in the repo because of
> a few issues on which I'm unsure how to best proceed.
Not to mention the holiday break!
> Since I followed the CPAN distribution structure, what I have
> created is more than a single file. So far, however, all of the
> provided implementations are single files. Do I just stick the
> directory in there in place of Stanis' single-file port? How are
> the tarballs on the project currently produced?
As far as I know, tarballs are produced by hand. At least that's how I
did the release that preceded the incorporation of the JavaScript
version -- that release touched all the languages we supported at the
time and introduced the Erlang version.
I don't see an issue with putting a whole directory in there for Perl.
Different languages have different needs.
> Also, as noted, I extracted the test data into JSON files, so
> that the other languages can share it and only need to implement
> a driver portion for each of the datasets. For the CPAN distro I
> need this data in the tarball, so it must be included in some
> form in the Perl subtree. How do we go about sharing the files?
> Symlinks? (How) Does that work for Windows-based Subversion
> users?
Could we put the test data at the top-level along with something as
mundane as a Makefile that could create the various tarballs, copy
test data into the Perl directory, etc.? I assume all active
developers on this project have access to something other than Windows
and so can deal with a simple Makefile.
IMO, of course, and open to other suggestions too of course.
--steve
Aristotle Pagaltzis
> Could we put the test data at the top-level along with
> something as mundane as a Makefile that could create the
> various tarballs, copy test data into the Perl directory,
> etc.?
That would work for me. I even enjoy writing Makefiles. :-)
Joe Gregorio
>
> On 1/6/09, Aristotle Pagaltzis <paga...@gmx.de> wrote:
>> I've dragged my feet on putting my port in the repo because of
>> a few issues on which I'm unsure how to best proceed.
>
> Not to mention the holiday break!
>
>> Since I followed the CPAN distribution structure, what I have
>> created is more than a single file. So far, however, all of the
>> provided implementations are single files. Do I just stick the
>> directory in there in place of Stanis' single-file port? How are
>> the tarballs on the project currently produced?
>
> As far as I know, tarballs are produced by hand. At least that's how I
> did the release that preceded the incorporation of the JavaScript
> version -- that release touched all the languages we supported at the
> time and introduced the Erlang version.
>
> I don't see an issue with putting a whole directory in there for Perl.
> Different languages have different needs.
>
>> Also, as noted, I extracted the test data into JSON files, so
>> that the other languages can share it and only need to implement
>> a driver portion for each of the datasets. For the CPAN distro I
>> need this data in the tarball, so it must be included in some
>> form in the Perl subtree. How do we go about sharing the files?
>> Symlinks? (How) Does that work for Windows-based Subversion
>> users?
>
> Could we put the test data at the top-level along with something as
> mundane as a Makefile that could create the various tarballs, copy
> test data into the Perl directory, etc.?
the languages would be good too.
-joe
> I assume all active
> developers on this project have access to something other than Windows
> and so can deal with a simple Makefile.
>
> IMO, of course, and open to other suggestions too of course.
>
> --steve
>
> >
>
Joe Gregorio http://bitworking.org