Alex Nitz

@josh-willis Poke (sorry, don't know when you are back form vacation. @spxiwh Any idea when we can move forward on this one?

Noting this is on hold awaiting changes following discussions on a better implementation with @josh-willis. I'll get back to this when my task list is sufficiently clear (so unknown).

@josh-willis Can you look at this? It seem quite worrying.

@josh-willis I can reproduce this with the following as well (essentially, the same as posted, just minor simplifications and switched to PhenomD as PhenomA has backward Fourier conventions). ``` import...

@josh-willis However, the following *does* give the correct result on the same environment. ``` import pylab from pycbc import types, fft, waveform delta_t = 1 / 4096 delta_f = 1.0...

@joao-aveiro Thanks for reporting this. I think you've identified the issue correctly. If you have a suggestion we have happy to accept PRs. I think the main issue here is...

I think looking at this those datasets are using the bulk data naming convention before it was changed. We could try a fallback here in case the standard changes? Perhaps...

@joao-aveiro You are correct that that would be the better behavior. If you think you can add that I think we'd be happy to accept a PR. The underlying library...

@spxiwh @titodalcanton Maybe it's worth adding a unittest for this class and the coinc one at this time? Especially with these functions, I worry about small bugs creeping in.

@spxiwh I think the changes you've made though are OK in general. I've read through and I think I understand that you just preallocate some buffer size and then only...