List of known problems:
- Apache2 incompatible. An Apache2::Wyrd framework is in the works, though
it may be released as a context-sensitive wrapper class.
- Concurrent access modes (CDB_<FOO>)of the BerkeleyDB module have
not been widely tested, and the BerkeleyDB Index is scheduled for
demolition before the 1.0 release.
List of known problems which will not be fixed, and why:
- Until there's a good reason to do so, Apache::Wyrd will continue to
use the "old-style" file semantics, since they are slightly more
efficient than the IO::<FOO> packages. A general attempt has been
made to make it reasonably easy to override these methods if you
must use another method of dealing with files. Please report a bug
to the author if the offending code is not easy enough to override.
- Wyrds should be case-insensitive: too cycle-costly and only encourages
sloppiness. Attributes are already case-insensitive.
- Should be able to put HTML within the attributes inside the Wyrd tag:
Only a few instances can't be done with Attribute Wyrds, and those will
likely require an additional mini-interpreter "language" like the
Apache::Wyrd::Services::Setter conditional mini-language for escaping
HTML: This would likely be cycle-costly.
- Wyrds should not encourage polluting the system-wide module
namespace: Guilty as charged.
- It should work more like PHP or Mason: No need to duplicate effort. If
PHP or Mason works for you, use it!
- The _self_parse method which is integral to the Wyrd object is
functionally incompatible with the idea of a "non-parsed-header"
document: It makes more sense to use something else hand-coded if you
need to do that.
- The labor cost of converting the framework to D.C.'s spiffy new "Best
Practices" at this point outweigh the benefits. Even the Apache2 port
will not need sufficient rewriting. Current style is consistent where
required, and many of the more important best practices are already used.