Alexander Tsybulsky
Alexander Tsybulsky
Also serialization to xml does not look too efficient. I would probably consider serializing just the tree table. Of course, there should be an efficient way to initialize a new...
Not sure I understood the whole idea (indeed looks complex) but here are some comments: 1) `interface zif_ajson does not include if_serializable either?` - it does. And it does not...
Also +1 to Marc's comment about serilization to json - that would probably be a lot more debuggable. Although slower. So either convenient and readable json or fast but binary...
Yeah, old hanging issue ... I was thinking about specifiying the separator char (optionally) but it looked complex. But what about serialization? And did you measure the performance impact, by...
On my hardware: `21,9010570 seconds` vs `21,9381740 seconds` after So the diff is negligible. Merging, Thanks for a good feature ! :+1:
Certainly a good use-case and interesting idea. However, it is not a "JSON native/intuitive behavior", so I wouldn't implement it in the core class itself. And certainly not with autodetection....
Well, it is certainly a good pattern to implement, no doubt. Just that it is surely must be an utility. PR is welcomed, or I will implement it one day...
Interesting idea. whats the use case for string_table by the way ?
:+1: will have a look during the weekend. "PS: Cosmetic change for type constants" - oops :)
Hi Marc, sorry for the long delay. Just lack of time and energy ... :/ Good PR, merging ... Perf test just for info (minor perf impact - insignificant for...