Open richardjgowers opened 1 year ago
honestly at this rate we should just drop the TRZ reader for v3?
Yeah I doubt anyone is actually using it in anger
I do think this is more symptomatic of a class of potential bug where if a Reader is relying on information stored inside Timestep
they might make mistakes when inside a ChainReader, since this is breaking the assumption that nobody else is editing the Timestep. I might do a crawl through existing readers and see if this assumption is being made elsewhere.
The only upside is that it crashes...
(Also seeing that there's still Python 2 code inside TRZ indicates that it might not be the most looked at code...)
Something I've noticed while tinkering with ChainReader and transformations. I think this is because TRZ makes seeks relative to the current state of
ts.frame
, which is misleading inside a chained trajectoryCode to reproduce the behavior
raises
Current version of MDAnalysis
python -c "import MDAnalysis as mda; print(mda.__version__)"
)python -V
)?