Here are some issues not discussed in the kernel documentation, that need to be resolved before technical users who know nothing about the kernel (like me) will be able to test the kernel in any meaningful way...
How to configure support for Virtual Terminals?
Using an install straight from the Debian Woody DVD-R, it can compile, boot and get to X. However there is absolutely no output from "loading kernel...", to the start of X, and cntl-alt-f1 gets be to an unchanged screen, not a login prompt. I know that VT support is not enabled by default (why?), but enabling VT and console on VT does not make any difference. Same thing happens with Mandrake 9.1.
Are compile errors for default configuration OK?
I thought I may have accidentally removed something required for VT support. So I made a completely default install, i.e. "make config" and hold down enter for all questions. However, this would not even compile on gcc 3.95.4 from Debian Woody. Are drivers that don't even compile enabled by default, or is the statement in the docs that any gcc 3.95.x where x>3 will be ok, out of date?
Would getting the.config file from a working install of 2.6-test help?
If so do you have a link?
Why does make modules_install complain about missing dependencies?
Why doesn't it just make the dependencies? What are we meant to do about this?
Also, is NTFS write support ready for end-user testing in non-mission critical situations?
It worked under a completely unpatched Debian Woody DVD-R install, so I would say yes. I had some problem with VT support, but I think that was a configuration issue with the kernel rather than a dependancy problem.
I think that the linux community should use more "Aggressive Negotiations" with SCO. Some progress has been made with the german libel lawsuit. Since most of the technology they claim is stolen is owned by others, from JFS to SysV (http://www.novell.com/news/press/archive/2003/05/ pr03033.html) it seems that some of their more personal accusations could be libelous. Since the IBM lawsuit does not involve copyright violations, forcing a lawsuit on this issue could force them to put or shut up much earlier.
Also since the LVM code chosen was not IBM's, it is rather surprising that it is "copied" from SCO. Proving that it was SCO that pirated the code could be useful from a PR point of view. We should not merely sighing when SCO claims that Linux is like napster. We should note that the only reason that SCO got away with it for so long was because like napster that closed their source.
The best case senario is not SCO going away after giving Linux a beating in the press, it is people associating piracy with the Bolshevik like secrecy of closed source software.
"Windows XP? Isn't that like that SCO, and Napster? Sorry, Windows Paper Clip Server 2000 seems just right for the job but we just can't take the risk that they are hiding IP violations under their NDAs."
"Linux needed millions of developers world wide to reach enterprise level. Windows 2050 could not have reached enterprise without the code that IBM stole from linux. Yer honour, IBM has clearly terminiated the GPL, and has lost the right to distribute GPL-AIX. Thus windows licenses are now invalid, and all customers must switch to linux. The customers knew they were using closed source that had not be subject to peer review for IP violations, so they not innocent parties."
Mwahahaha! Attack is the best defense! Mwahahaha!
Slashdot Mirror
How to configure support for Virtual Terminals?
Using an install straight from the Debian Woody DVD-R, it can compile, boot and get to X. However there is absolutely no output from "loading kernel...", to the start of X, and cntl-alt-f1 gets be to an unchanged screen, not a login prompt. I know that VT support is not enabled by default (why?), but enabling VT and console on VT does not make any difference. Same thing happens with Mandrake 9.1.
Are compile errors for default configuration OK?
I thought I may have accidentally removed something required for VT support. So I made a completely default install, i.e. "make config" and hold down enter for all questions. However, this would not even compile on gcc 3.95.4 from Debian Woody. Are drivers that don't even compile enabled by default, or is the statement in the docs that any gcc 3.95.x where x>3 will be ok, out of date?
Would getting the .config file from a working install of 2.6-test help?
If so do you have a link?
Why does make modules_install complain about missing dependencies?
Why doesn't it just make the dependencies? What are we meant to do about this?
Also, is NTFS write support ready for end-user testing in non-mission critical situations?
It worked under a completely unpatched Debian Woody DVD-R install, so I would say yes. I had some problem with VT support, but I think that was a configuration issue with the kernel rather than a dependancy problem.
I think that the linux community should use more "Aggressive Negotiations" with SCO. Some progress has been made with the german libel lawsuit. Since most of the technology they claim is stolen is owned by others, from JFS to SysV (http://www.novell.com/news/press/archive/2003/05/ pr03033.html) it seems that some of their more personal accusations could be libelous. Since the IBM lawsuit does not involve copyright violations, forcing a lawsuit on this issue could force them to put or shut up much earlier.
Also since the LVM code chosen was not IBM's, it is rather surprising that it is "copied" from SCO. Proving that it was SCO that pirated the code could be useful from a PR point of view. We should not merely sighing when SCO claims that Linux is like napster. We should note that the only reason that SCO got away with it for so long was because like napster that closed their source.
The best case senario is not SCO going away after giving Linux a beating in the press, it is people associating piracy with the Bolshevik like secrecy of closed source software.
"Windows XP? Isn't that like that SCO, and Napster? Sorry, Windows Paper Clip Server 2000 seems just right for the job but we just can't take the risk that they are hiding IP violations under their NDAs."
"Linux needed millions of developers world wide to reach enterprise level. Windows 2050 could not have reached enterprise without the code that IBM stole from linux. Yer honour, IBM has clearly terminiated the GPL, and has lost the right to distribute GPL-AIX. Thus windows licenses are now invalid, and all customers must switch to linux. The customers knew they were using closed source that had not be subject to peer review for IP violations, so they not innocent parties."
Mwahahaha! Attack is the best defense! Mwahahaha!