Fix end frame replacement to acknowledge the chunk size.

[carlosfelgarcia]

Link the Issue(s) this Pull Request is related to.
Fixes #1129

Summarize your change.
Simple change to replace the "FRAME_END" variable, acknowledging the chunk size by selecting the minimum between the last frame and last frame in the chunk.

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: carlosfelgarcia / name: Carlos Garcia (88fa02c, 58f28b3, 9bbd632)

[anton-ubi]

Hi there, thank you for doing this!

We met the same problem recently and this is nice to see this being addressed.

I wonder if the logic is correct though. Wouldn't it be the indexes we want to compare, rather than the frame numbers themselves ?

Let's take that example:

chunkSize = 3
frameRange = "1,2,3,4,99"  # Note the frame 99 at the end

With that configuration, I'd expect the frame set to be divided into two chunks:

[1, 2, 3] and [4, 99] where #FRAME_START# and #FRAME_END# are respectively 1 and 3 for chunk 1, and 4 and 99 for chunk 2.

Now with the current proposal, #FRAME_END# of the 2nd chunk is 6 (frameNumber + chunkSize - 1 being less than 99). That seems to be an incorrect behavior.

---

I'd suggest to follow the behavior of the getChunk method of the FrameSet java object, as introduced by @donalm here
The comparaison is made on the indexes rather than the values. We take the lesser one, and only then we convert to the corresponding frame value.

What do you think ?

[carlosfelgarcia]

Yes, you are absolutely right, will revisit in the coming days.
Thank you!

[DiegoTavares]

Hi @carlosfelgarcia any update on this?

[carlosfelgarcia]

Hi Diego, this has totally got out of my radar, I just updated the code.

[DiegoTavares]

Can you rebase this from master to make sure you get the fixes to failing unit tests merged in?