If there were a clear path that showed which patches the Mandrake guys had applied, in what order, then it might be possible to construct a hybrid kernel that retained as many Mandrake features as possible but solved your specific problem. And the information gained would be useful to Mandrake. But I have never found a document that detailed the patches applied by the Mandrake guys. Unless it exists, Mandrake are missing an opportunity to harness community input.
Cheers,
Bruno Prior
Vincent Meyer, MD wrote:
On Sunday 30 March 2003 02:49 pm, simon wrote:
I know this is not a "Real" bug report, but I'm starting to get very unsure of the kernel source supplied with mandrake. I know this kernel has hundreds of patches added to it that are not in the main tree. Although I understand the need to support lots of hardware (I know I've got some difficult stuff) I'm sure it is very wrong to do it like this (lots of patches on top of pre releases of the next kernel).
A solution would be to get an "Alan Cox" to work for/with mandrake as a lieutenant for Linux Torvalds or Marcelo or both. That way people will trust these kernels and they will be tested very thoroughly. As it is now, whenever there's a kernel bug, you couldn't possibly have the manpower/knowledge to find the source of it in the big pile of patches you have included.
Hmmm.. Well, I trust the kernels that come out of Mandrakesoft. I also think it's kind of condesending to assume that our kernel maintainers aren't smart enough to troubleshoot what they create. I personally don't much care if the Mandrake kernel has someone's "trust" or seal of approval. I've generally had pretty good luck with it, and when I've had problems, they've usually been diagnosed and corrected fairly quickly.
And if you want a "standard" kernel, it's out there - download it and use it!
Getting an "Alan" and having him as a trusted part of the kernel community will probably take years, unless you get Alan himself ;-), but it'll be worth it!
Worth it? How so?
V.
