On Sat, 12 Aug 100, Ken Williams wrote:
> That seems right - so perhaps it would be better to integrate the
> StatINC functionality into Apache::Reload than the other way around.
> If it's called Apache::Reload, then it might make sense to have a
> directive like 'ReloadMode' that could be set to 'INC' to emulate
> StatINC's behavior, or to 'Files' to explicitly name the files you
> want checked with every request, or to 'Slave' (perhaps a bad choice)
> to only check those files that do the "use Apache::Reload" thingy
> you've written.
Apart from the name "Slave" I think its right on the money :)
I think Required or Imported would be better.
> 'Files' and 'INC' wouldn't be very different in implementation, but
> 'Files' gives you finer-grained control. They both just take a list
> of files and check them with every request.
There is some trickiness to it, such as you want to "do" a file, but
"require" a module (%INC gives you modules (and files) whereas Files mode
you'd probably just specify a bunch of files to stat).
> It might be nice to have only one [main] module for reloading changed
> code files, both because of memespace issues and because the actual
> reloading task seems to be a tricky one, and it would be nice to only
> have to get it right once. But if it doesn't make sense to combine
> the approaches into one module, it doesn't make sense.
One issue that will always remain is the legacy of StatINC...
But this gives me something fun to hack on after I walk the dogs today :)
--
<Matt/>
Fastnet Software Ltd. High Performance Web Specialists
Providing mod_perl, XML, Sybase and Oracle solutions
Email for training and consultancy availability.
http://sergeant.org | AxKit: http://axkit.org
