Good luck, and a happy New Year to all of you.
Cat_7
EDIT: updates in the code:
-Add SSSE3 optimizations (Intel Core 2 CPUs and newer). This brings an overall +10% performance improvement in AltiVec Fractal Carbon.
-Detect SSE 4.1
-Add support for 64-bit MacOS X for PowerPC (untested beyond "it compiles").
-Exception handler code[1] always contains the fault address nowadays. So make it the default fast path but keep provisions to check that at run-time first. This yields a nearly 4x improvement in SIGSEGV recovery.
-Fix build with new MacOS X extfs implementation (in C++).
Last edited by Cat_7 on Wed Jan 02, 2008 12:19 pm, edited 1 time in total.
Apparently most problems with the "Unix" shared folder are solved. Files, applications, custom icons, all show their icons properly and I could even launch applications from inside the shared folder.
One problem with the shared folder: When in MacOS in SheepShaver a bunch of files and folders from the shared folder are trashed and then the trash is emptied, the files are deleted as expected, but the folders can not be deleted.
It is still not possible to run SheepShaver full screen, the screen stays black.
(Intel Core 2 Duo iMac, Mac OS X 10.4.11, SheepShaver running Mac OS 9.0.4)
First my best wishes for 2008 and thanks for the new version, is seems to me a litle bit more stable then the previous version.
It's a pity that still very short after I connect to other computers in my network the connection is broken.
I hope that one day slirp will be replaced by a far more stable version and that at the same time it will be possible to make an alias of a connected server on the Desktop and that it will be possible to connect to that server by just clicking on the alias instead of entering the full IP address in the Chooser each time a connection is needed.
That part is really annoying so I hope that this year will be the year that SheepShaver will tackle all those big and little isssues.
Does the fixing of the UNIX folder sharing mean that you can leave the shared disk image on the desktop on OSX and even have it opened while you have SS launched?
I wouldn't recommend that.
What it means is that the Unix shared drive is functioning better. So if you designate a folder to be shared through the external file system option in the GUI there are fewer errors to expect.
There is a bug in SheepShaver that will cause SheepShaver to freeze when the cursor has been moved in and out of the SheepShaver window a certain number of times. The bug has not yet been resolved.
You can move the cursor, but not to often. I forgot the number of times one has to move the cursor in and out of the window to make SheepShaver freeze. Somewhere around thirty?
that's fine. I downloaded IE for os9 and then I won't to have to move the cursor away from os9 window.
Another thing: I realized there's a limit to load only 3 disks images in SS, right?
There is a bug in SheepShaver that will cause SheepShaver to freeze when the cursor has been moved in and out of the SheepShaver window a certain number of times. The bug has not yet been resolved.
I'd pay to see this fixed. Sheepshaver is just the way to go. Most 68k programs run fine on it. Most PPC programs run fine.
Ronald P. Regensburg wrote:You can move the cursor, but not to often. I forgot the number of times one has to move the cursor in and out of the window to make SheepShaver freeze. Somewhere around thirty?
I don't think it's an exact number, but the average is somewhere around 80. When you have moved the mouse in and out the window more than 60 times, it's time to save your work and restart SS before it hangs. Sometimes you can tell when the next 'exit' will be fatal, if the keyboard starts missing or repeating strokes.
Actually it doesn't completely hang, but the mouse and keyboard input stops working (which in practice is nearly as bad as a complete system freeze of course).
You can move the cursor, but not to often. I forgot the number of times one has to move the cursor in and out of the window to make SheepShaver freeze. Somewhere around thirty?
I have a fix for that problem, which results from an interaction between SheepShaver and SDL. The fix is in code from BasiliskII. Can somebody tell me who to contact about getting the change into the source tree?
I also have a workaround for the "text cursor off by a few pixels" problem, and a bug fix in the external filesystem code.
I'm currently building with SDL release 1.2.10; later releases (up through 1.2.13) do not work for me -- they exhibit the problem that the cursor never changes. Has anybody built SheepShaver with a later release of SDL and not had that problem? I believe the SDL release installed by Fink is 1.2.9.
Thanks Cat_7; I will post bug reports at sourceforge and see what happens.
I don't want to encourage people to download and run SheepShaver builds created by random unknown people on the Internet (i.e., me), but those who like to live dangerously and want to beta-test my changes can download my current build at http://www.nomicro.com/SheepShaver.html.
Please read the caveats there. You can contact me by responding to this thread.
Nice!
In your build the problem of SheepShaver becoming unresponsive to mouse or keyboard input is indeed solved. This, and the correction of the I-beam cursor position is enough for me to use this build for the time being.
For me, the most annoying problem left is the the cursor not displaying applications' cursor images.
I still wonder how the cursor problems were introduced in SheepShaver while none of these SheepShaver cursor problems exist in BasiliskII. Can it be related to the cursor in BasiliskII not moving out of the window at all?
Ronald P. Regensburg wrote:For me, the most annoying problem left is the the cursor not displaying applications' cursor images.
If my understanding is correct, cursor changes will be seen only if made via the video driver's "set hardware cursor" function. I do see this happening when (e.g.) I run MacMoney, which uses a few different cursors. Directly changing the cursor image in low memory would not have any effect since nobody is looking at it. (I suppose you could poll it; the BeOs video implementation appears to do so.)
Ronald P. Regensburg wrote: still wonder how the cursor problems were introduced in SheepShaver while none of these SheepShaver cursor problems exist in BasiliskII. Can it be related to the cursor in BasiliskII not moving out of the window at all?
The cursor hot-spot issue is specific to the SheepShaver video driver, and the hanging issue occurs only when the cursor moves in/out of the window, or the application activates/deactivates.
Binder wrote:Any chance of an Windows compiled build?
I don't have the facilities to do that, but my changes are available for anybody who has the ability to build from source code.
Ronald P. Regensburg wrote:For me, the most annoying problem left is the the cursor not displaying applications' cursor images.
If my understanding is correct, cursor changes will be seen only if made via the video driver's "set hardware cursor" function. I do see this happening when (e.g.) I run MacMoney, which uses a few different cursors. Directly changing the cursor image in low memory would not have any effect since nobody is looking at it. (I suppose you could poll it; the BeOs video implementation appears to do so.)
I am sorry, but I do not understand what you are talking about. I only know that the changing cursor works fine in BasiliskII, but does not work in SheepShaver. The default cursor disappears, leaving a kind of speckled shadow, but the cursor image that should appear does not.
Ronald P. Regensburg wrote: still wonder how the cursor problems were introduced in SheepShaver while none of these SheepShaver cursor problems exist in BasiliskII. Can it be related to the cursor in BasiliskII not moving out of the window at all?
The cursor hot-spot issue is specific to the SheepShaver video driver, and the hanging issue occurs only when the cursor moves in/out of the window, or the application activates/deactivates.
I have mentioned this over and over again, but again and again the essence of what I am trying to tell seems to be ignored. The cursor behaves very differently in BasiliskII. In BasiliskII the cursor cannot move outside the window. Outside the window the MacOSX cursor takes over while the MacOS cursor remains at the border inside the BasiliskII window. I wonder why this is done differently in SheepShaver and whether several cursor problems that exist(ed) in SheepShaver are somehow related to this difference. Also, for me the BasiliskII behavior feels more natural than the SheepShaver behavior where the cursor moves freely between the two different environments.
Edit: A picture showing the BasiliskII cursor behavior:
