Find file
Fetching contributors…
Cannot retrieve contributors at this time
185 lines (108 sloc) 4.67 KB
* Also think about different debug flags, e.g. DEBUG_THIS, DEBUG_THAT
UPDATE: Badger::Debug now does this - just need to hook class in
* make debug hook work better: allow import of other items, e.g.
debug => ':dump'. Also generate DEBUG constant using any existing
value of $DEBUG.
* AUTOLOAD method to delegate to real class if possible?
* add 'hook' subroutine attribute
* add a way for B::C subclasses to register hooks that automatically
get called on import, e.g. for B::Factory::Class to push base()
* Remove class($class, $target) second argument in export hooks
because UBER should now handle that - needs testing
UPDATE - no, I think it's still required to make The Right Thing happen.
* Make CLASS static.
* Change message() to snippet()
NOTE - already changed Badger::Base XXX_msg() method to explicitly
package-scope message() to Badger::Base so that problem is mitigated.
A subclass can now define a message() method and it will still Just Work[tm]
* encoded()/decoded() methods (e.g. for utf8)
* update docs to clarify the fact that utf8, etc., are available as named
codecs via Encode/Encoding modules.
* Just a basic implementation at moment. Merge in TT3 config, AppConfig
and other stuff.
* Add 'constant' as alias for 'method'. e.g. 'FOO|class:FOO|constant:FOO'
* See if we can make this a low-level mixin that we can import into
Badger::Utils et al.
* make debug enabler export :debug and :dump into module
* Finish writing it / cleaning it up and release.
* Do we still want to add a higher-level catch() method?
* Subclass out into base class (returns loaded module name), object
creator, object creator + cache (e.g. for hub), or whatever.
* have it bind fetch methods to item/items if undefined, e.g
node => item, nodes => items.
* support multi-element names, e.g.
* directory files/directories/dirs as grep across children
* need to handle encoding better in read_file() and write_file() methods.
* read-only filesystem option
* write_file() should have an option for writing to a temporary file and
renaming into place to avoid race conditions.
* I'm considering refactoring this. I'd like to unify the different
filesystem-specific formats to an underlying URI-based one. It would
require a lens (combined parser + presenter) for each O/S. Unix (inc
Mac OSX) are simple, Win32 isn't hard. Delegate to File::Spec for
everything else.
Some info about Win32 path <-> uri translation, snarfed from
For the UNC Windows file path
\\laptop\My Documents\FileSchemeURIs.doc
The corresponding valid file URI in Windows is the following:
For the local Windows file path
C:\Documents and Settings\davris\FileSchemeURIs.doc
The corresponding valid file URI in Windows is:
Functions currently used from File::Spec are: catpath, catdir,
canonpath, splitpath, splitdir, filename_is_absolute, abs2rel,
no_upwards. Most of those can be moved out to lenses. Path
manipulation becomes much easier internally when the path can be
stored as a list of path nodes.
* Add code to detect Log::Dispatch objects and forward messages.
* This is available in the git repository but not yet included in the released
* unify parameter parsing and identity definition between filesystem and
database storage modules.
* get/put which take data / return id
* fetch/store which returns
* create/destroy connect/disconnect open/close
* Could do with some proper testing. Although all the other test scripts
do tend to thrash it quite well...
* Change if_env to use Badger::Logic so we can specify things like:
* hashlike, listlike, numlike
* add path to any bases missing it
* document
Longer Terms Goals / Larger Projects
* incorporate re-write of AppConfig into Badger::Config
* consider doing the same with Pod::POM into Badger::Pod and combining
with Badger::Docs. (update: doing now)
* Finish cleaning Badger-Web and release
* Finish refactoring Badger-Database and release