Re: Newer Versions of Octave Struggle with Character Strings

[Mike Miller]

On Mon, Jul 03, 2017 at 08:03:26 -0700, evansste wrote:
> I've noticed that Octave 4.0.0 and 4.0.3 struggle to handle character strings
> that contain a significant number of elements.  When attempting to handle
> operations with large character strings, these versions return the known
> error message:
> 
> error: input buffer overflow, can't enlarge buffer because scanner uses
> REJECT
> input buffer overflow, can't enlarge buffer because scanner uses REJECT

Fortunately, this issue has already been reported and fixed. If you
update to Octave 4.2 this should be resolved for you.

For others trying to reproduce, the long input line must be read from a
script file. I used the following

    alphabet = char (repmat ('a':'z', 1, 2^15));
    fid = fopen ("foo.m", "wt");
    fprintf (fid, "alphabet = '%s';\n", alphabet);
    fclose (fid);
    clear all
    foo

which crashes 4.0.3 with the reported error but is ok in 4.2.1.

See also bug #45947.

-- 
mike