I have started experiencing some really weird issues after updating to the latest Rhino release? When working in GH, using GH Components inside GHPython, I am now regularly getting ‘fails’ after working for a bit?
For example, If I do something simple like try and call DeconstructBrep in a GHPython, at first it all works great, but then after a while working in the file, it starts returning ‘none’ for everything ? :
If I use ‘normal’ GH Components on the same input geom. it all works fine - so I know it’s not a problem with the geometry or anything - it’s only in the GHPython compo that I get these fails - and only after I’ve been working in the file for a while?
If I do a close and reopen Rhino, it then all goes back to working normally:
but after working for a while, it starts to fail again. I’ve never experienced this before, and it only just started happening this week after I updated to the April 10 release. So far, it seems to be only the ghpythonlib.components methods that are causing this issue? It seems to happen every 2-3 hrs or so.
For reference, the full details of my system are:
System:
Rhino:
Version 7 SR5
(7.5.21100.3001, 2021-04-10)
has anyone else experienced anything similar? Are there settings someplace I need to configure to avoid these issues? Some sort of weird memory issue? I really appreciate any thoughts or input on this issue - it’s starting to really disrupt work and cause an awful lot of confusion.
I’ll see if I can save something out the next time it happens. It’s very peculiar, and as I indicated - a close/restart seems to fix it every-time.
I happened to me again just this morning after working for a few hours - I noticed that it seems to occur when I ‘change’ something as part of a large set (ie; I have a big group of points / surfaces referenced into GH, then I edit one of the set, or add one to the set) That seems to cause this funny issue.
In any event - I’ll see if I can capture it the next time it occurs and I’ll share for sure.
I am still having trouble with this issue unfortunately. I think I have narrowed it down a little bit though.
If I have a Python component in GH that calls some Grasshopper methods (as I showed above), it will mostly work fine. However, sometimes, when I use the ‘UNDO’ command back in the Rhino side (I move something by accident, etc…) on some geometry which was referenced by this GHPython component, this is when I think I often begin to see these failures in my GHPython components. All the GH methods begin returning ‘None’ values, and only a re-start will fix it.
After restart, all mostly works ok (until I work for a bit and accidentally hit Undo…). Re-Referncing the geom doesn’t work - the only thing which seems to clear it up is a full application restart unfortunately.
That seems to be the closest thing to a pattern that I can find when these failures are happening.
Somehow I start to think that this might not have to do with the library, but with some references held in the code itself that calls the library. Is it possible that there are some Data Dams? Or, are you using some third-party add-ons? Or, are you keeping track of objects inside the code, and using IDs?
Unfortunately it’s really hard to narrow down the problem by just having a description. To be more helpful, there should be a repeatable case with a simple definition.
Yes, totally understandable. I’ll see what I can do to recreate this error more reliably and hopefully I can ‘capture’ it at some point. Its not a Data Dam or a any funky third party add-ons - but it could very well be the way I’m referencing in the data - I’ll dive a little deeper into that a see if I can get a cleanup up example to share.
No worries though - as I said, a restart always fixes it. Good enough for the time being.
