#FRAME_END# token wrong behaviour

[TanJeremy]

Hello all.
We have an issue with the #FRAME_END# token when the frame range is not divisible by the chunk size.

For exemple, if we have 72 frames with a chunk size of 10,
the last job will have the correct #FRAMESPEC# "71,72", but the #FRAME_END# will be 80,
rendering 8 more frame than necessary.

unfortunatly, i don't have knowledge in java to offer you a ready to use "pull request", and only suggest you some idea like
.replaceAll("#END_FRAME#", String.valueOf(Math.min(frameNumber+frame.chunkSize-1), fs.getAll()[ fs.size() - 1]))
i'm not sure the syntax is correct but the idea is there.

Many thanks

Opencue Version Number
0.14.5

[carlosfelgarcia]

Does anyone know if this issue has been addressed? I did have the same issue and I already resolved, it's been under production for a month and not issues. Please let me know, so I can implement a PR for it.

Thanks,

[TanJeremy]

Does anyone know if this issue has been addressed? I did have the same issue and I already resolved, it's been under production for a month and not issues. Please let me know, so I can implement a PR for it.

Thanks,

it doesn't seem to be the case

[KernZag]

Would you mind sending a PR with your fix ?

[carlosfelgarcia]

Would you mind sending a PR with your fix ?

Definitely I can, probably next week as I'm a busy at work.

[antonshack]

Hi there.

Any news regarding that issue ?

I came up with that workaround in the meantime, when providing the command to the layer:

For Windows/CMD:

command = [
        f"for /f %i in ('powershell -command \"[math]::min(#FRAME_END#,{frame_range.end})\"')",
        "do",
    ] + py_cmd

Cheers.

[carlosfelgarcia]

Thanks for reminding me about this one, I just created a draft PR (#1320) as it might need some tests to be created before I put it in review.