Blanking at end of sequnce when playing v2017.1 thru 2017.8 fseq on FPP [SOLVED]

[gadgetsmith]

I'm working on some lighting for a school drama club production.  I've been using the newest release of xLights (2017.7) to create effects and play them on a FPP.  All my sequences are designed to run in a 30 second loop and repeat until the scene ends when the lighting kid hits 'stop'.  What i'm finding is that any sequence produced on v2017.x will 'blackout' at the end of the 30 seconds, rather than continue on with the 'repeat' box checked.  Sequences created in v2016.56 do not have this issue.  I'm not sure if there is a new way things are being rendered that mess with the FPP's handling of 'repeat' sequences, or a new setting someplace that I don't yet know about, or user error, or some bug.

Any ideas?

Thanks,
Charlie

SOLUTION: Couple lines of code created a situation in which the last two frames of the fseq were not rendered. dkulp submitted a change to fix this issue.

[keithsw1111]

So you rendered the same sequence in 56 and got no blanking then you did it in 7 and uploaded fseq with no other changes and now it blanks?

[gadgetsmith]

yes, correct.

[mararunr]

Same for me.  If I use 2016.51 (production version) I get no blanking, but if I use 2017.7 (latest beta) I get blanking.

[Gilrock]

You should be able to expand to the node level and see if the data is missing.  Also if you open the .fseq file in a binary editor I would expect the end of the file to be a block of zeroes if the data is blanking at the end.  If neither of those show zero then it's whatever is playing the file causing the blanking.

[Gilrock]

Anyone use Beyond Compare?  Should be able to drop a simple On effect across the entire sequence and a little past the end and then render it in both versions and do a binary compare to see differences.

[gadgetsmith]

I've done that with a simple 30 second sequence, so I'll check the fseq.

[keithsw1111]

Can you try xschedule again using v.8.

[sean]

Can you try xschedule again using v.8.

Keith I thought the original post was fseq files created from old versions vs 2017.7 as they run on the FPP.
Since FPP is running the fseq, i would think it would have to be a difference in the fseq.

Maybe save two fseq files and do a comparison between them. i have some utility that compares binary files.

[keithsw1111]

I changed xschedule which was blanking and now it shouldn't be.

[gadgetsmith]

I downloaded Beyond Compare as mentioned in Gil's post.  The attached screenshot is the end of the comparison, and indeed the 2017.7 file contains many zeros at the end.

[gadgetsmith]

I changed xschedule which was blanking and now it shouldn't be.

As indicated by Sean, i'm not using the scheduler, i'm using FPP.  It appears something has changed in the fseq file, and if so, maybe the FPP guys need to change something in the software to get the 2017 fseq files to play without blanking?  It sounds like you changed something in the scheduler to keep the files from blanking on the end?

To everyone.  Thank you for looking into this.

[gadgetsmith]

I'm replying here, but see now that jholmes is posting on Falcon and GitHub, so I need to digest that information.  Thought there was something to the zero's at the end of the 2017 fseq, and not the 2016.56 fseq.

[keithsw1111]

Gadgetsmith. The only files we have received did not demonstrate the problem you describe. Maybe you can share yours.

[gadgetsmith]

doh. yea, I'll get them uploaded here.