Testing MinGW Octave-3.4.2 binaries

[Philip Nienhuis]

Tatsuro,


Last week I've done a fair amount of work with your Octave-3.4.2-MinGW 
binary.
I'm quite happy with it, it works good, and it is faster than the 3.2.4 
MinGW binary I've been using up till now.

But IMO it might still be too unstable for building a binary installer 
package out of it - see below.


The few issues I found so far are these:

1. "clear all" leads to a complete crash. For a split second 
"segmentation violation" is shown before all traces of Octave are gone.
I think thus bug has been there already since the Octave-3.2.4-MinGW or 
even Octave-3.2.3-MinGW binaries by Benjamin Lindner.

2.  A Matlab script, made by a colleague, that I'm porting to Octave 
creates 34 pictures. On some PC's, using fltk, doing "close all" leads 
to a similar crash as 1. (above)
By closing manually, or by repeated "close" commands, one plot window 
after the other, the crash occurs after 10-20 plots have been closed.
If needed I can supply the script. An octave-core file is generated but 
I can't do much with it.
While the script may be wrong (though it works OK with gnuplot), Octave 
shouldn't crash that hard.

3. With that script, again on some PCs only, using fltk, titles appear 
only on the first plot. Using gnuplot all plots get the proper titles.

4. Again, on some PCs only, text output in fltk pics is gibberish.
May be due to lacking fonts - I haven't had time to investigate. Gnuplot 
plots are fine.

5. The bug I reported earlier (edit.m complaining about "unrecognized \P 
\O sequences") is a bit of a showstopper. Even when the EDITOR command 
echoes the full path to the editor program with double backslashes, edit 
<file> still complains (again, on some PCs only...).


Issue 1. is already known (though I couldn't spot the bug entry).

I'm a bit reluctant about entering bug tracker entries for issues 2-4 as 
not all of these are fully repeatable on all PCs I tested on.

Issue #5 has already been reported.


Philip