Re: Again, seeking advocate
Roger Leigh wrote:
However, I think that the long wait is a good idea in general, since
it requires sponsoring of your work, which will hopefully pick up any
problems you have. While Michael Banck was sponsoring my packages for
the last six months, his testing of my packages did find some
problems, so it was certainly worthwhile. The wait also demonstrates
that applicants have at least a medium-term level of commitment to the
Project. In my case, I had been maintaining the Debian packages
upstream for over a year prior to my applying, so the NM process
wasn't exactly a challenge.
Hey good news, I have been sponsored so I'm in the queue now. In the meantime
I have a lot of good suggestions for my package which I intend to incorporate
into the scheme. I believe I will increment my version to 1.8 and convert back
to a regular upstream packaging scheme, with the Debian specific diff against
it. Really that is the only way to do it I think if you want to be able to
run it on the autobuilders.
Now I need to get some sound support going in Hurd so I can port this package
to it, or is somebody looking at this already? The last time I built OSKit/Mach
I don't recall any sound options in OSKit. Which reminds me that I need to get
going about recompiling it again.
I switched over to a new motherboard because the SMP board started going bad.
This new one is an Asus A7N8X Deluxe with the "AMD Viper" style chipset. I
can confirm that the latest Hurd tarball will definately not boot.
I can, however, make an OSKit/Mach serial cross-debugging setup because I have
successfully compiled an Arm -> X86/Hurd cross-gdb for my little Netwinder
webserver, and I have a null-modem cable. I guess that will be my concurrent
Hurd development project at the moment. Can't do much with Hurd until I can
get it to boot again :-)
That is pretty neat using the serial cross-debugger... single stepping through
the Mach kernel and all that.
I also wouldn't mind toying with one of the L4 variants, Notably Fiasco seems
nifty in that it is written entirely in C. There is also a rather interesting
real time kernel called MarteOS which is written mostly in Ada95 and links
against portions of OSKit. These little kernels have nowhere near the hardware
support that gnumach does right now though, haven't had much luck getting them
to boot yet, though I finally have a floppy drive again :-)
Which of the Hurd servers are the most essential, and which are the most trivial?