Dear Morten,

answer to your last question first:

> What exactly is the "GNUmeric portable app version"?
http://portableapps.com/apps/office/gnumeric_portable

You can obtain LibreOfficePortable there as well, for example. In this
context, "portable" means that you can run a "portable" app from a USB
stick attached to any PC running a recent version of MS-Win, and take 
all your personal settings with you; no administrative privileges 
neccessary for installation, no registry entries etc. very handy for 
the "mobile workforce"... of course, portable apps might use code of 
Win standard libraries. 

> We are not supposed to have anything like magic brackets.
> We don't even have code for it.
I appreciate this opinion very much, as I hate "unkown unknows"...

Here's the link to the paper in which W. Kahan complains about this 
phenomenon:
http://www.cs.berkeley.edu/~wkahan/Mindless.pdf
The "magic bracket" subject is treated in "ยง2: Errors Designed Not To 
Be Found" on page 3 ff. I used the 1/2^n approach in order to find out
how many trailing bits are cleared.

> Could you please file a bug for this at bugzilla.gnome.org with a test
> sheet?
I promise to do that this weekend; I'm a bit busy right now because 
I'll leave for a two days business trip tomorrow morning. Please find 
attached the Win-version of my sheet. My Linux-GNUmeric was able to 
open it (sorry, but I'm on Win now and in a hurry...).

All the best
Schorsch

Attachment: DP-tests.gnumeric
Description: Binary data

_______________________________________________
gnumeric-list mailing list
gnumeric-list@gnome.org
https://mail.gnome.org/mailman/listinfo/gnumeric-list

Reply via email to