Back to the main page.
Bug 2910 - uint64 arithmetic operations can crash Matlab
| Status | CLOSED FIXED |
| Reported | 2015-06-18 11:28:00 +0200 |
| Modified | 2019-08-10 12:33:29 +0200 |
| Product: | FieldTrip |
| Component: | core |
| Version: | unspecified |
| Hardware: | PC |
| Operating System: | Mac OS |
| Importance: | P5 normal |
| Assigned to: | |
| URL: | |
| Tags: | |
| Depends on: | |
| Blocks: | |
| See also: |
nno - 2015-06-18 11:28:55 +0200
The following code, when run from FieldTrip's fileio/@uint64 directory, makes Matlab crash: a=uint64(zeros(0,1)); max(a) (and similarly for 'min' instead of 'max'). Also, FT's @uint64 has a few minor issues does not support input sizes that builtin Matlab / Octave versions support (e.g. non-vector inputs for min/max). For this reason it may be better to use builtin functions when available (*). Suggested fix: https://github.com/fieldtrip/fieldtrip/pull/64 (*) I am aware of bug 1354 marked as "wontfix", yet think that my proposed fix can be used without adverse side effects. In particular, the proposal that users who don't want to use FT's uint64 operations should delete the directory would require manual intervention by the user that can be avoided.
nno - 2016-02-07 20:20:58 +0100
Compatibility with GNU Octave has increased significantly with PR 78 [1] which was merged into master. [1] https://github.com/fieldtrip/fieldtrip/pull/78