Implement Map<> URIHandle#time() which returns all supported time metadata at once.
Rename content() to text()
Create content() which returns binary data as byte[], or ByteBuffer, or something
Implement `Map<> URIHandle#time()` which returns all supported time metadata at once.
Rename `content()` to `text()`
Create `content()` which returns binary data as byte[], or ByteBuffer, or something
When we talk about Map<> URIHandle#time() it would be better to just
don't add new method
old method should load all timestamps, if it is supported and memoize them
next calls should just read memoized values
Update of memoized content should be triggered if you ask again about URITime which was used before for memoization. It means, if you as two times for modification, it will be fetched two times.
To implement this, we should store internally:
memoized result
URITime used to memoize
When we talk about `Map<> URIHandle#time()` it would be better to just
- don't add new method
- old method should load all timestamps, if it is supported and memoize them
- next calls should just read memoized values
Update of memoized content should be triggered if you ask again about URITime which was used before for memoization. It means, if you as two times for modification, it will be fetched two times.
To implement this, we should store internally:
- memoized result
- URITime used to memoize
Implement
Map<> URIHandle#time()which returns all supported time metadata at once.Rename
content()totext()Create
content()which returns binary data as byte[], or ByteBuffer, or somethingWhen we talk about
Map<> URIHandle#time()it would be better to justUpdate of memoized content should be triggered if you ask again about URITime which was used before for memoization. It means, if you as two times for modification, it will be fetched two times.
To implement this, we should store internally:
Honestly, current implementations of time are good enough. Skip this part.