Steve132
Read and write JSONLab/JData flavored optimized N-D array header
see https://github.com/ubjson/universal-binary-json/issues/61#issuecomment-492445324
Fixed a few gcc warnings, added the missing ubjw_begin_ndarray() function.
valgrind now gives me a lost buffer inside priv_ubjr_read_raw_array() which I modified, I suspect it is related to this line, not sure if that call releases the memory associated with the temp variable.
Why does your spec differ from the existing ND array functionality extension inside UBJ (using the @ character) It looks to provide the same functionality exactly, with the same ordering.
I guess it was probably because it was implemented in my JSONLab toolbox about 1 year before your proposal of the @ format, which I only noticed it many years later. At that point, there is already a sizable user community for jsonlab, and I don't want it to break backward compatibility. I agree these ideas are rather similar.
There was a draft 13 posted online...but no, the array syntax was never "accepted" per se. The debate about how to move forward with UBJ for the draft 1.0 spec (whether to go in the direction you went or whether to do a more schema-based solution like protobuf) never really solidified into anything and the project sort of died. I took it over intending to move it forward benevolent dictator style towards a 1.0 spec that was very very similar to what you ended up for JData, but then my life got really complicated and I wasn't able to push as hard as I wanted.
I think the right step forward for the project would be to agree on an array spec, agree on other changes, and move forward to a 1.0 spec, along with reference implementations in as many languages as possible and a test suite. The problem is how to do so and maintain backwards compatibility and how far to go. The other problem is that the original designer of the project who passed it on to me didn't give me control of ubjson.org
On Tue, May 12, 2020 at 9:55 PM Qianqian Fang [email protected] wrote:
maybe I overlooked, I did not see the #+@ syntax in the Draft 12 of the spec https://github.com/ubjson/universal-binary-json/blob/master/spec12/container-types.html, was it accepted?
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/Steve132/ubj/pull/8#issuecomment-627697785, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAGWIMDCIGIQ2WTX3WS522DRRH4YPANCNFSM4HM7IINA .