Make ByteBuffer.description useful and pretty

[weissi]

Right now, ByteBuffer's description is ugly and not very useful (doesn't print the bytes). With #2856 being merged, I'm proposing to make ByteBuffer's description the following:

[ hex bytes, maximum 64 or something ](length bytes)

so hello world would be

[68656c6c6f20776f726c64](11 bytes)

and 256 0 bytes would be

[00000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000...00000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000](256 bytes)

the empty string would be [](0 bytes)

[FranzBusch]

Onboard with this idea. The current description is not really helpful in most cases and I always do a hexDump manually while debugging. We should also change the debugDescription of ByteBuffer while we are it.

[weissi]

Onboard with this idea. The current description is not really helpful in most cases and I always do a hexDump manually while debugging. We should also change the debugDescription of ByteBuffer while we are it.

Normally I'd say that we shouldn't change debugDescription (because that actually has the bytes and when debugging a ByteBuffer the other stuff is interesting too). However, if you do Array<ByteBuffer>.description then that will invoke the element's debugDescription!?... So yeah, I guess we should just make .description == .debugDescription

[FranzBusch]

Yes CustromStringConvertible and CustomDebugStringConvertible are a bit messy. The latter is used by collection types to print their contents. So you should always implement both. At least that's my latest understanding @dnadoba dove into this a bit a while ago.

[supersonicbyte]

I can pick this one up also 👋

[supersonicbyte]

@weissi since @FranzBusch mentioned here that () are preferred, should the output of ByteBuffer's description be wrapped with () or should we stick with []? IMO, [] feels more natural for ByteBuffer.

[dnadoba]

Normally I'd say that we shouldn't change debugDescription (because that actually has the bytes and when debugging a ByteBuffer the other stuff is interesting too). However, if you do Array.description then that will invoke the element's debugDescription!?... So yeah, I guess we should just make .description == .debugDescription

Good! Then please approve: #2495

[supersonicbyte]

Here it is