| 1 |
Hanno Meyer-Thurow wrote: |
| 2 |
|
| 3 |
>Why I call it 64bit hacks? |
| 4 |
>That patchset for 64bit support lets OpenOffice.org compile on amd64 |
| 5 |
>but then runtime is still unstable. One thing is to make OpenOffice.org |
| 6 |
>compile on amd64. The other thing is to make OpenOffice.org runtime |
| 7 |
>stable. That means some code just does not need fixing but a full |
| 8 |
>rewrite from scratch until it works nicely on amd64. |
| 9 |
> |
| 10 |
>If I say something wrong anyone please correct me! Thanks. |
| 11 |
> |
| 12 |
> |
| 13 |
Well, I disagree that the code needs a full rewrite. Why? All it needs |
| 14 |
is correcting all places where it attempts to store pointers in 32-bit |
| 15 |
variables. Admittedly, in the core of UNO this might happen a LOT (due |
| 16 |
to its nature it does a lot of direct memory manipulation), but there's |
| 17 |
no real reason why it should be re-written from scratch. |
| 18 |
If the UNO runtime violated other portability issues (such as code |
| 19 |
execution protection), it wouldn't run at all in 64-bit mode, perhaps |
| 20 |
not even when compiled as 32-bit. |
| 21 |
|
| 22 |
Sebastian Redl |
| 23 |
-- |
| 24 |
gentoo-amd64@g.o mailing list |
Archives