Stefan Bodewig <[EMAIL PROTECTED]> wrote: > * Allow mappers to be genericised so that particular features can be > modified during mapping. Something similar to
-1 on using mappers, +1 on defining a better way to attach attributes to filesets. > * Allow include/exclude tow work with multiple characteristerics of > a file. +1 but keep a simple syntax for pattern matching. > * provide datatypes through property tag and remove need for > separate free standing entities. -0, i.e. I don't care to much and don't see much point in it. > * provide support for non-hardwired (ie loadable) low-level > components (mappers/itemset-filters/converters). > > * provide support for non-hardwired (ie loadable) converters. +1 on making things more pluggable, we'll have to define the details in the design phase. > * Make all datatypes interfaces to allow them to be customized in > many ways. -1 for that generalized form. We can make them interfaces if we don't need a contract that is stronger than method names. In the case of <path> we have a stronger contract (build.sysclasspath handling) right now, but this might change if we spread responsibility in a different way. > * Set arithmetic for fileset/patternset/*set +1 > * inheritance of ant properties/datatypes/context etc in project > hierarchy +1 - and is a result of the unified namespace. > * inheritance of between ant datatypes. ie fileset A inherits from > fileset B (includes all entries in A). Here inheritance still seems to imply set union, no reason to have it as a separate and ambiguos point. -1 > * Homogenize notion of PATHs and filesets -1, path is ordered, fileset is not - and they should stay that way. Stefan
