On Apr 21, 2005, at 6:55 PM, Tracy R Reed wrote:
The fact that Linus doesn't believe in kernel debuggers and forces that
decision on everybody.
Nobody else seems to believe in them either as I am not aware of anyone
really using one. Linus can't force you to not use a kernel debugger can he?
Yes, he can. Primarily by not providing any kernel "hooks" to do things like get information, get symbols, query the kernel, smack processes from within gdb, etc.
Dropping into a debugger from the FreeBSD kernel and pounding away is a really good way to figure out what the heck happened. Dropping into a debugger on the Linux kernel is horrifically painful.
The whole VM subsystem mess
I hear a lot of people complain about the vm but I've never understood why. It has always worked great for me in the stable kernel versions.
Forkbombs. Overzealous malloc() allocation. System thrash under light load but heavy memory use.
Linus says that Perforce and many others are not suitable for the
massively distributed development effort that is the linux kernel. Does
he have a point there? I doubt any Perforce user has so many people all
working on the same project scattered all over the world.
I would probably disagree: http://www.perforce.com/perforce/customers/byindustry.html
The fact that Perforce is used by several games companies and a couple of movie houses means that it probably *dwarfs* the Linux kernel in terms of amount tracked and distributed development.
Linus is just being pissy and won't admit that he is wrong. This is a known character trait. Sometimes it stands him in good stead; sometimes it works against him.
My respect for him has dropped a huge amount primarily over his reaction to Andrew Tridgell rather than anything else. Linus has basically been sounding *exactly* like a Microsoft parrot with the reverse engineering whining.
Oh, and in case somebody missed it: Tridge demonstrated his "advanced reverse engineering techniques"-- telnet <repository> 5000 and then type help.
http://www.theregister.co.uk/2005/04/21/tridgell_bitkeeper_howto/
Based partly on the discussions in this list my company has chosen SVN.
Although I liked darcs too. We had originally decided to go with CVS but
the improvements that Lan cited plus the nice web interfaces that come
with bonuses to us.
Cool. Glad the discussions were helpful.
I know that I would certainly go with SVN over CVS if I was making the decision for a company.
Any easy way to tell how many authors the BSD and Linux kernels have in
common? If you could somehow pull an author list from both, sort, and
unique them against each other.
No easy way that I can think of. I know of this just because I tended to be surprised at seeing quite a few well-known Linux names in the repository. Also, quite a few of the subsystem people try to make sure that things are interoperable. I bumped into this far more when I was rewriting rpc.lockd for FreeBSD.
-a -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
