I’m using Mathematica 11.3 on macOS 10.13.3 (High Sierra). Anyway, when an error is detected in a cell in Mathematica 11.3, it causes the previous cells to become unevaluated. Then I’m forced to evaluate to the entire notebook again from the top. Is this something can be disabled and/or reverted back to 11.2 behavior?
From the image, all the previous evaluated cells are not evaluated at the point of the error and it’s not clear to me as to how to disable this functionality.
Has the kernel crashed?
After the failure, one can still evaluate other expressions but all cells prior to the error become unevaluated. If the cells prior to the error evaluated successfully, why do the have a grey circle instead of the standard number? Anyway, this issue and several others have been reported to Wolfram support regarding Mathematica 11.3.
I do not think that the kernel crashed because I can evaluate cells after the error.
It has the look of a kernel crash, possibly a bug in the software but impossible to say without a full working (as in misbehaving) example. If that was sent to Wolfram support then there is a reasonable chance of diagnosis. For any possibility of diagnosis in the Community forum you’d need to post a MWE here as well. If this involves proprietary data, maybe it can be replaced with synthetic data and still elicit the trouble.
Hi, I can almost reproduce this in any notebook by inserting an error within one of the cells and watch prior successful cells become unevaluated when error has been detected in the notebook. Here are the steps to reproduce:
expect - the successful cells to remain evaluated
actual - the successful cells prior to the error change from evaluated to unevaluated (i.e. grey circle)
Furthermore, when one fixes the cell in which an error has been introduced, evaluating just the fixed cell’s expression will still generate an error because the dependent cells have not been evaluated. Please see the attached notebook for an example of this issue.
