On Wed, 2010-12-08 at 21:15 +, Michael Meeks wrote:
Do we enable / run the bridgetests ?
We always build and run the cpp bridge test, yes. Well except for MacOSX
PPC. That particular bridge is broken somewhere.
C.
___
LibreOffice mailing
On Thu, 2010-12-09 at 01:42 +0100, Jan Holesovsky wrote:
http://agner.org./optimize/calling_conventions.pdf
and from what I see on page 19, it seems to me that *BSD defaults to the
'fastcall' variant, right?
We're looking at Table 7. Methods for returning structure, class and
union objects
Hi Caolan,
On 2010-12-09 at 09:43 +, Caolán McNamara wrote:
and from what I see on page 19, it seems to me that *BSD defaults to the
'fastcall' variant, right?
We're looking at Table 7. Methods for returning structure, class and
union objects right ?
The table suggests that
Hey,
Sweet!
Should I do a full build with it on i386 or did you test it already
on the box?
On (2010-12-08 16:49), Caolán McNamara wrote:
So, checking this out on Robert's OpenBSD box and digging into MacOSX
land it appears that Linux has different x86 struct returning rules than
MacOSX and
On Wed, 2010-12-08 at 18:52 +0100, Robert Nagy wrote:
Hey,
Sweet!
Should I do a full build with it on i386 or did you test it already
on the box?
I tested that testtools would build and pass its tests. That *should* be
sufficient to ensure that the uno bridge is 100%, but the testtools is
On Wed, 2010-12-08 at 20:38 +, Caolán McNamara wrote:
I tested that testtools would build and pass its tests. That *should* be
sufficient to ensure that the uno bridge is 100%, but the testtools is
comprised of tests that got added when the last time it was thought that
all conditions
Hi Caolan,
On 2010-12-08 at 16:49 +, Caolán McNamara wrote:
So attached is my proposed change that would allow the testtools tests
to run correctly during the build without being hacked out or disabled,
which is rather important because if those tests fail it implies that
the final