conflict between reffect_effect_switch and refl_num in legacy wrapper

[kecnry]

secondary luminosities need to be updated within phoebe1. Apparently this happens automatically for detached but not contact systems?

For the 2.0 release - we either need to get to the bottom of this or throw an error saying we do not support calling phoebe1 for contact systems (which also makes testing difficult). We also need to understand why this isn't causing issues for the detached case.

[bpablo]

This bug report is actually incorrect. What is actually happening is the following:

In newer versions of phoebe2 we have gotten rid of the reffect_effect_switch. Basically the assumption is that if a star has reflections then reflection is on. Most of the time this is fine, except if you have multiple reflections in your file but have the reffect_effect_switch turned off. Then phoebe2 reads it in sees that there are multiple reflections and passes a system to legacy with the reflection switch turned on. This is pretty straightforward, but the fix isn't. There are in my mind two ways this situation can be handled:

1.) output a warning saying something like "reflection switch is off but multiple reflections are listed. Since phoebe2 doesn't distinguish the reflection will be on by default. If you do not wish this to be so then set refl_num to 0."
2.) If reflection effect is off pass the number of reflections to phoebe2 as 0 regardless of what the number actually is.

Which is preferable?

[aprsa]

I vote the latter. If the reflection effect is off, then the number of reflections is a dummy variable.

That said, reflection in WD is never truly off: the "simple" reflection model is always used, it just doesn't do a detailed computation of multiple reflections. Thus, the "reflection" switch in phoebe 1 should be understood as "detailed reflection" switch. How are you dealing with switch=1, number of reflections=1?

[bpablo]

The current logic is as follows:

refl_num = 0:
albedo both components is set to zero
reflection effect is set to 0

refl_num = 1:
reflection effect is set to 0

refl_num= multiple:
reflection effect is set to 1

[aprsa]

When reflection is on, and refl_num=1, phoebe 1 computes reflection using the inverse square law, so we are not consistent with the previous functionality, but at the same time we don't support this "simple" reflection model. I am adding @horvatm to this discussion.

[bpablo]

The wrapper has now been changed to the second option: automatically set refl_num to 0 when reflection switch is off. It also raises a warning alerting you to this fact. I am not closing this thread though, as there is still some question as to how to handle refl_num=1.

[kecnry]

great, then I'll just switch from 2.0 to future since we decided to defer the "simple" reflection case from this release

[kecnry]

@bpablo - can you check or remind us of the status of this? I know we changed the behavior slightly for 2.1 - was that enough to consider this closed or not?