The default key comparer used by QuackingDictionary, and its child class Row, can't currently be changed.
My changes allow for a custom StringComparer to be specified upon instantiation. It is then preserved over row cloning and copying.
This can be useful in various scenarios:
If case sensitivity is required for keys
If better performance is needed (my particular usecase)
My commit should be entirely backwards compatible and doesn't change existing behaviour. Nevertheless, for an easy performance improvement accross all users of the library, at very little cost, I would suggest changing the default value for QuackingDictionary.DefaultComparer to StringComparer.OrdinalIgnoreCase after my changes are pulled in. This should be nearly identical except in very remote corner cases, and tests on my computer have shown that it is up to 10x faster for hashing (which happens a lot in Rhino ETL applications).
The default key comparer used by QuackingDictionary, and its child class Row, can't currently be changed.
My changes allow for a custom StringComparer to be specified upon instantiation. It is then preserved over row cloning and copying. This can be useful in various scenarios:
My commit should be entirely backwards compatible and doesn't change existing behaviour. Nevertheless, for an easy performance improvement accross all users of the library, at very little cost, I would suggest changing the default value for QuackingDictionary.DefaultComparer to StringComparer.OrdinalIgnoreCase after my changes are pulled in. This should be nearly identical except in very remote corner cases, and tests on my computer have shown that it is up to 10x faster for hashing (which happens a lot in Rhino ETL applications).