There appears to be a typo, since np.dtype is not a dtype, but a function for getting dtypes. Older versions of numpy allow this to slide by assuming an object dtype which get corrected later when writing to/from the h5 file. With at least numpy 1.25, this call results in a process hanging error that stops the compiler cold.
Fix is pretty simple: self.raw_output = np.array(self.timeseries, dtype=np.dtype(self.dtype)) to match the call below it.
Refreshed a labscript installation and bumped up against this line
https://github.com/labscript-suite/labscript/blob/bc7fa4b98d75591988ae2e1c5aee97da86e5067c/labscript/labscript.py#L1725
There appears to be a typo, since
np.dtypeis not a dtype, but a function for getting dtypes. Older versions of numpy allow this to slide by assuming anobjectdtype which get corrected later when writing to/from the h5 file. With at least numpy 1.25, this call results in a process hanging error that stops the compiler cold.Fix is pretty simple:
self.raw_output = np.array(self.timeseries, dtype=np.dtype(self.dtype))to match the call below it.