Wine Traffic
Eric Pouech
This is the sixth release of the Wine's kernel cousin publication. It's
main goal is to distribute widely what's going on around Wine (the Un*x
windows emulator).
Well this week has been very quiet on wine-devel, as if most of the
developpers were on vacation at the same time. Of course, there was the
O'Reilly conference in Monterey (CA) where Alexandre Juillard was giving a
talk. So, I barely had enough material to put an issue together.
Marcus Meissner will be giving a talk at the Linux Kongress in Augsburg,
Germany.
Uwe Bonnes
Uwe Bonnes reported the following question: after a block of memory
has been freed by HeapFree(), what should be the output of
IsBadReadPtr() on this very block.
Ian Schmidt investigated on Windows 98 and reported: If the size passed to IsBadReadPtr() is 4k or less it returns zero
(memory is OK). If the size is greater than 4k it returns nonzero. So it
seems like the first page of the pointer is OK immediately after HeapFree()
but the rest isn't. The actual allocation size and presence/absence of other
allocated blocks in the heap don't seem to matter.
This remain to be patched in Wine.
Uwe Bonnes
Patrik Stridval found an article regarding CPU performance, but which
contains a benchmark between Windows and Wine on the same hardware (
http://www.cpureview.com/rev_k7-linux1_a.html).
Here's an excerpt of the results (on the new Athlon 600 from AMD):
Recently I've started playing with Wine 990815;
one of the programs I tried to run is an old favorite of mine; WinTach. I
was pleasantly surprised to see this old 16 bit benchmark application work
perfectly. The results with a yellow background are obtained under Wine, the
other row was obtained under Windows 98SE on the same hardware.
| Mhz |
Driver |
VSync |
WP |
CAD |
Spread |
Paint |
Score |
| 600 |
Wine |
on |
980.78 |
1490.02 |
490.02 |
255.48 |
804.19 |
| 600 |
Win98 |
OFF |
2043.49 |
1976.52 |
2102.37 |
651.87 |
1693.56 |
Patrik asked if anybody else done benchmarking
of Wine vs Windows?
", and started wondering on the cause of the
speed difference (X Windows, Wine itself...), but anyway was satisfied with
the results at this stage of Wine development.
Uwe Bonnes proposed if some benchmarks were to be done to also integrate
figures from wmware.
Uwe Bonnes reported some feedback he got from the cygwin mailinglist (via
http://www.cygnus.com/cygwin), and chances of seeing Mumit Khan's patches
for anonymous unions/structs being incorporated in gcc 3.0. Mumit wrote:
Some of the patches are in, and some (notably the C anon aggregates one,
and a few others) are not.
The more users lobby, the more its chances are for inclusion; so please
voice your interest in gcc@gcc.gnu.org!
So Uwe asked people from the wine community to raise
their voices.