Different architectures in the same or different directory trees?

At $work, we maintain a set of Perl modules at a central location for easy inclusion via PERL5LIB. As there is a re-installation ahead and we need to provide the modules for both 32 and 64 bit architecture, we are开发者_StackOverflow中文版 wondering if it's better to install them into the same directory tree, relying on the $archname subdirectories, or keep the two architectures entirely separate and duplicate each module.

I was not very successful at researching the inner workings of the Perl module lookup process involving $archname, maybe someone can point me in the right direction.

In your experience, what are the pros and cons of the two approaches?

---

From perldoc lib:

When using use lib LIST;

For each directory in LIST (called $dir here) the lib module also checks to see if a directory called $dir/$archname/auto exists. If so the $dir/$archname directory is assumed to be a corresponding architecture specific directory and is added to @INC in front of $dir.

lib.pm also checks if directories called $dir/$version and $dir/$version/$archname exist and adds these directories to @INC.

IMHO, it is more idiomatic - and dare I say, neater - to use the per-architecture subdirectories, like Perl's standard libraries would.

However, it night be more straightforward to manage per-architecture-entire-tree of your own libraries, though not by a large margin once you create a few basic tools/scripts to do so.

---

Build the modules separately on each system so that you only get the files needed there. Or use a packaging system that distinguishes between architectures. Don't try to provide the files for all architectures to all systems.