Contributing a feature - is it relevant?

[danni-m]

Contributing a feature - is it relevant?

[danni-m]

First of all, contributing is always welcomed.

Regarding the features:

1. This is supported at the moment for the latest timestamp only, Im planning of adding compression to the data, once we have compression its harder to update existing timestamp without a big overhead.
   having said that, with the simple chunk interface we can support updating without a problem, I would be willing to accept a PR for that as long as it kept in the chunk level.
2. back filling is a problem, currently the chunk design is implmented in a way that adding a timestamp in a middle of a chunk is a not easy. you would need to move data around and possibly create a new chunk.
   I can see an easy implementation if the timestamps you are backfilling are older then the oldest timestamp at the moment.

Another possible implementation for this would be done in the client side where you copy all the data from a key and re-insert it to another key.

[shaiker]

I made a PR for (1) - #21

regarding (2) - I guess I'll do it on the client as you suggested :-)

Cheers.

P.S.
Just out of curiosity - Perhaps I'm just not familiar enough with how Redis manages modules, but I noticed that in ChunkItertorGetNext and SeriesItertorGetNext you're doing memcpy and creating a copy of a Sample, but never freeing that copy's memory. Doesn't it create memory leaks ?

[danni-m]

Regarding the memory, memcpy used for copying data around, it doesn't allocate more memory.
in SeriesItertorGetNext we create a variable internalSample on the stack, this variable location is passed to SeriesItertorGetNext, when we exit the function this variable will be deleted from the stack.

Generally, the caller to ChunkItertorGetNext and SeriesItertorGetNext are incharge of allocating and releasing the memory.