Beefy Boxes and Bandwidth Generously Provided by pair Networks
good chemistry is complicated,
and a little bit messy -LW

Re: Where to put the bits and pieces - Installing a complex CPAN dist

by adamk (Chaplain)
on Sep 23, 2008 at 12:52 UTC ( #713208=note: print w/replies, xml ) Need Help??

in reply to Where to put the bits and pieces - Installing a complex CPAN dist

See the behaviour of the current versions of Module::Install::Share and File::ShareDir for the "correct" pure-Perlish shared directory paths.
  • Comment on Re: Where to put the bits and pieces - Installing a complex CPAN dist

Replies are listed 'Best First'.
Re^2: Where to put the bits and pieces - Installing a complex CPAN dist
by Hercynium (Hermit) on Sep 23, 2008 at 16:57 UTC
    Thanks! I actually found those this past weekend, and while I don't think they quite fit here, they are *quite* handy! In a fit of Laziness, Impatience and Hubris I began crafting my own system-path-resolver-thingie and it's using File::ShareDir, along with File::HomeDir, et. al.

    I do plan on posting some of my notes here just as soon as I've finished figuring out which ones make me look smart and which ones betray my true level of ignorance. :)

    OK, a link to the notes has been posted in the update section of the OP!
      There's a few problems I see... Firstly, you've dumped straight under unix system share... what if there are non-Perl packages that share a name with a CPAN distribution? What about when there are multiple versions of Perl installed on the machine? Where is the equivalent path on Windows? How do they get packaged into Perl distributions. The reason we use the auto method is that it handles most of the above cases...
        *gah* I had written a response to this but lost it with an ill-placed mouse-click >:o

        Anyhow, the gist is that the concept floating around in my head is to *add* to the options a developer has when choosing paths, but to do so in a structured manner. Static data that is used only by your perl code should *certainly* be installed under the auto hierarchy, but sometimes an application has static data that should be in a place more accessible to the user and/or non-perl system components. Web apps especially seem to have this need.

        I would provide installation options for an author targeted not just for module distributions, but also for system applications and web applications. I explained this in more detail in the post that got lost, but I just want to add that when I get a few spare tuits, I think I will play around with implementing this as extensions/subclasses of your (and Ingy and Audrey's) Module::Install. I like the interface. I like the relative cleanliness. I like the flexibility for extension.

        I've added some more notes to my scratchpad which I *did* save before closing the wrong window *grumble*. I really appreciate your input and *greatly* admire your contributions to the Perl community.

Log In?

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://713208]
[Corion]: Hmmm. I feel a Meditation coming on. I wrote a module, DBIx::PivotQuery, which returns a table-like set of rows (AoA) but some columns are generated from column values, like in an (Excel) pivot table or a ROLLUP command
[Corion]: My current approach for subtotals involves rerunning the given query, with the hint to the user that they should use a temporary table if they want better performance.
[Corion]: But I could create that temporary table in the module and use it for the improved perfomance directly instead.
[Corion]: And the question is, what would be better/preferred ;-)
[Corion]: Hmm - not exactly like the ROLLUP command. Ah well.

How do I use this? | Other CB clients
Other Users?
Others pondering the Monastery: (10)
As of 2017-02-23 15:26 GMT
Find Nodes?
    Voting Booth?
    Before electricity was invented, what was the Electric Eel called?

    Results (347 votes). Check out past polls.