https://bugs.documentfoundation.org/show_bug.cgi?id=99730
Eike Rathke <[email protected]> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |ASSIGNED
Assignee|[email protected] |[email protected]
|desktop.org |
Ever confirmed|0 |1
--- Comment #3 from Eike Rathke <[email protected]> ---
The reason I did it that way probably was that the determinant for the specific
matrix set SHOULD be 0.0 even with our LU decomposition, and so far WAS, except
for the 32 bit case and now librenedine's case. Which is odd. Btw, what system
still uses gcc 4.8.5? Even Debian Jessie has 4.9.2 ...
So, I'm not against a general error margin of 10e-12, which we also use at
other places, if this helps to build. I'm just sad we have to lower the bar to
cater for inferior systems ;-)
--
You are receiving this mail because:
You are the assignee for the bug._______________________________________________
Libreoffice-bugs mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/libreoffice-bugs
