Mac Forum / Programming / Perl / May 2006

I have a patch to change this vanilla file read:

  my $text = do { local $/; open my($fh), $file; <$fh> };

to this one using perlio's utf8 layer

  my $text = do { local $/; open my($fh), "< :utf8", $file; <$fh> };

Aside from differences in Perl versions (that's easy enough to handle)
and the stupidity of reading the whole file in one go, is this going to
bite me in butt with encodings? I have next to no expereince with UTF8
problems because we mostly use ASCII around here.

My main concerns are:

  * Will a plist always be utf8? For instance, what about
localizations for Japan, etc? The encoding shows up in the <?xml ...?>
portion, but am I going to have to open the file to find that, then
re-open it?

  * What happens if it's not utf8 but I tell Perl it is?

  * Will whatever I do work on another machine? I wrote this module to
read plist files on my FreeBSD machine, and I don't want to make users
have to be on a Mac (or even the same Mac) to be able to parse a file.

I have some other ideas to get around this, but my inexperience with
character encodings make me cautious (e.g. allows me to not do
anything).

  * Ignore it. A user can get the same thing by reading the file and
using parse_plist which takes a string

  * Add a parse_plist_fh method and let the user make his own
filehandles. I'm going to do that because it's a good idea, and I can
make parse_plist_file redirect to that if its arg is a filehandle.
However, the user then needs to open their own filehandle and know
things I'd like to be able to guess for them.

  * If the plist is always going to utf8, just make the change.

  * Actually hook into the Mac libraries and let the OS parse it and
give me back the data. That's the best solution, but only if I give up
being pure-Perl and portable, which are both very important to me. I
will gladly add something to do that if someone writes it for me, but
alongside the pure Perl version.

  * Can I hook into the Mac stuff with something like Inline::Java?

  * Find a sucker to take over the module and figure it out for me. As
PT Barnum said "There's a new CPAN user registered every second".