> This seems a very ambitious plan.
> I don't see how all this could be finished before 2002 at best.
Perhaps, but I personally support since I think it is nessary,
we really can't call it Wine 1.0 without at least most of these
implemented.
> If so, there is a possibility to have a 'moving target'
> effect, for example, if Ms
> merges its OS at some point, full Unicode support will probably become
> mandatory...
How do you expect them to pull this off? All (or at least most) old
application must still run otherwise people won't upgrade, and
application makers aren't likely to abandon their customers by
mandating them to install the new OS.
Personally, I ran Windows 95 OSR2 until at few month ago when I
bought a new computer that runs Windows 98. I have no wish to
run Windows 2000 or whatever especially if programs will break.
The _only_ thing I am happy with in Windows 98 compared to Windows
95 OSR2 is that WinAmp won't skip when I start the Microsoft
Terminal Server Client. But then I do that _once_ per day.
If I won't upgrade how will they make other people to upgrade?
I was perfectly happy with Windows 95 OSR2.
> Is it not a risk to want that much, given that a 'stable'
> Wine could bring more corporate support ?
If we release Wine 1.0 it _must_ work properly, we do not
want to scare cooperate people away, as I said in a earlier
mail, porting things like Corel Office, and similar applications
where as long as what comes out of the printer looks good everything
is OK. Note looks good not nessarily is correct. That is one thing,
but for wide cooperate acceptance it _must_ be correct.
With that in mind I have, this weekend, been working one a brute
force crash and burn regression tester called winapi_test. Some
simple tests seems to work on Linux, Solaris is still broken :-(
Currently I am tring to port the generated files to Windows now
in order be able make a comparision.
winapi_test is an extention to winapi_check, that instead of static
source analysis, generates dynamics tests with exception handlers
that are supposed to be possible to run under both under WineLib
and Windows.
Oh, as a side note while we are at it. Any ideas on where
to put the generated files and how to automatically make
them easily compiled?
Currenly as a kludge I put the files in wine/tests/winapi_test
with a Makefile.in file that is referenced in configure.in.
However the Makefile.in will, since what files that will be
in wine/tests/winapi_test depends and what tests that exists
and that is not fixes, so it has to be autogenerated. That
is not a problem in itself but since configure has be run
the generate the Makefile from the Makefile.in this is kind
of kludgy, especially since we definitly don't want the
generated files permanently in the tree since they are
already are 2Mb together and subject to heavy change.
What I want to do in principle is to generate a fully
functionally WineLib application, that can be compiled
on Windows as well.
Perhaps I should have a configure of my own?
As I side note perhaps the applications in
wine/programs should have that as well.
I'm open to suggestions. Comments?