heliocentric / evoke Goto Github PK

Update copies everything into a ram staging area by default. 
Unfortunately, if there is a low memory condition, this will overflow, 
leading update to believe the file is corrupted, and error.

Update needs to be able to query the boot device for free space, the ram 
disk for free space, and determine which is available to use. We should 
never /commit/ the final files to disk until after verification, however.

Since the filesystem tasting and remount tasting code works, mounter 
should now grow the ability to specify geoms, files, and directories 
without needing a protocol to be specified.

This will involve changing the logic to testing if the operand is a 
directory, a file, or character device, and making sure the second operand 
is a valid directory.

filetype has extraneous data, including extra ELF output, and 'setuid' if 
the file's setuid bit is set.

What steps will reproduce the problem?
1. mounter ufs:/dev/ad0s1a /mem/mnt


What is the expected output? What do you see instead?

A normal mount, however, it errors with 'device not configured' and fails

Please use labels and text to provide additional information.

netstat and sockstat do not work correctly on the FreeBSD 8.0 kernel. This 
is due to some structure size changes in 8.0 that causes the 7.2 netstat 
to break. This needs to be fixed before 0.2

The 8.0 kernel boots slowly on SMP machines, starting with the DDB output. 
This is most likely due to some of the debugging options that are in 
effect. Unable to reproduce on 8.0-RC2.

What steps will reproduce the problem?
1.
2.
3.

What is the expected output? What do you see instead?


Please use labels and text to provide additional information.

While testing XGGI, I noticed a disparity in how libggi seems to act under 
Xvnc and Xorg.

Under Xvnc, the color channels are flipped; red is blue, and blue is red. 
Other then this, all applications that I have tried have worked fine.

Under Xorg, the color channels are as they should be, however there seems 
to be some addressing problems with regards to drawing. Some things are 
offset 0x0 in the window, and moving windows (as well as hitting 'enter' 
in applications such as xterm), cause a jerky up and down movement.

Until I can figure out why the difference exists, I cannot fix the 'true' 
bug in this code, so that I can bring XGGI to a point where it is viable.

There are cursor related artifacts in both, I believe this is a problem in 
xggi proper.