It is not possible to extract an array of bytea.

[pinver]

Actually the PostgreSQL bytea type is handled by the D ubyte[] type, like the varchar, text is handled by the D string type.

The readArray method of 'Message' is using methods like 'arrayDimensions!T' or 'readDimension' that have specials check for handling strings, with the isSomeString template, but can't actually handle the case where the PostgreSQL column to read has a type bytea[], that must be read by a ubyte[][] array.

Suggestion? There's some workaround of this is an intended behaviours? Must I provide a pull request with the adaptations to 'readDimension', 'arrayDimensions', 'multiArrayElemType' to handle the case?

[pszturmaj]

isSomeString are neccessary because ddb treats strings as base types. Otherwise strings would be treated as arrays of chars. As for ubyte[][] problem, pull request is welcome.

[pinver]

Exactly, so the question is: should we also thread ubyte[] as a base type, for handling bytea? Or maybe immutable(ubyte)[], for symmetry with string?

[etcimon]

Or maybe immutable(ubyte)[], for symmetry with string?

I've never seen a data type like that before, but if it's serialized or binary data (which it probably is) you'll need a ubyte[] to put it through some engines like MsgPackD, so I guess the immutable would be somewhat impractical.