[% setvar title Cache byte-compiled programs and modules %]
To see what is currently happening visit http://www.perl6.org/
Cache byte-compiled programs and modules
Maintainer: Simon Cozens <email@example.com> Date: 25 Sep 2000 Mailing List: firstname.lastname@example.org Number: 301 Version: 1 Status: Developing
Python does it. Java does it. It's time we did it.
When Perl runs a program, it's compiled to bytecode and then the bytecode is executed. Now, what you could easily do is stop after compiling and dump out the bytecode to save you compiling it next time the program is run. In Perl 5, this is slow, because it's quicker to recompile than to read in the bytecode. In Perl 6 it'll be quick, and I'll have some suggestions as to how to speed it up soon.
So, we check for the existence of a
.plc file before running a
program; if the
.plc file is newer than the program, we use that
instead. If there isn't a
.plc file or it's older than the program,
recompile and dump the bytecode to a
.plc file. Naturally, this gives
us the best speedup for modules which change very, very infrequently,
rather than programs which can change a lot during development. Maybe
you only want to do this for modules, then.
Perl 5.6 has the elements of support for
.plc files, but it doesn't
tell anyone about this: have a look at
pp_require in pp_ctl.c.
(Hmm, it's not in 5.7.0, but I could have sworn I've seen it in there in
I also threw out a module called
ByteCache onto p5p a few months ago,
which implemented the thing as a coderef on
@INC. Implementing it in
pp_require is a cleaner way of doing it, however.