Advogato blog for drow

10 May 2000

Wed, 10 May 2000 19:15:39 GMT

OK, enough of this silence stuff.

Boot floppies seem to be in pretty good shape now! I just went through a clean install on both i386 and powerpc, and the results were really impressive. The installer has come a long, long way. I've also verified that the powerpc CDs do boot new-world Macintoshes as they ought.

I'm in the middle of installing one last time on crack (formerly crack.linuxppc.org, one of my prouder accomplishments) before I ship it off to become the new powerpc build daemon. It is both relieving and disappointing for me to do this - I'm going to have less than twice the number of computers of my roommate again. But I'm changing roommates in a few days and I think the new one will only have two, so it will all be ok again :)

Had some interesting conversation with Richard Braakman, the Release Manager, about organizing test cycles; there's pretty obviously going to have to be another one after this. We will probably need a new kernel update (for 2.2.15 final), and there are some security update, and some random important fixes.

I've been working on BitchX all week. The windowing layer finally pissed me off enough to merit a complete rewrite; I don't know if Panasync is ever going to accept this code, but I want to at least see how it turns out. I may get a useful console abstraction library out of it. We'll see.

29 Apr 2000

Sat, 29 Apr 2000 16:54:07 GMT

Been a little while - this is because I've gotten something approaching free time in the past few days. I'm not quite sure what it feels like, besides odd.

The CD creation is coming along nicely - we should know in a day or three if they will boot. If they do, we're pretty much set for potato.

Oh, and mdorman, it's Reason Number One mostly, I'd say.

24 Apr 2000

Mon, 24 Apr 2000 03:54:00 GMT

Woop!

We got the rest of the modconf issues tracked down now, hopefully - thanks a lot, tausq. Tasksel takes more gracefully to being control-c'd, too. No real settlement on the broken miBoot issue, but Ben is looking at it; for now I've reverted to the previous version.

Boot floppies 2.2.12 is uploaded for sparc, powerpc, i386, and alpha. The changelog is a fairly hefty piece of work. Thanks go out to Othmar Pasteka for his help in making all the documentation build again, and of course Adam Di Carlo for his usual excellent management. I've been doing a bit of the organization of the past two releases, and I am really getting a feel for why he hates his job.

I'm going to go back to debian-cd boot support tomorrow or the following day, as soon as we have 2.2.12 installed in the archive. YaBoot is a package now, and installed (a big Thank You to our Harried Release Manager, Richard Braakman). Thus I can clean up the code and remove a couple of the worst hacks.

23 Apr 2000

Sun, 23 Apr 2000 07:13:33 GMT

Well, I feel stupid. That build failed miserably - miBoot can not find the kernel! I'm backing down to the previous version of miBoot, 603e problems and all, for the moment. Remind me to test this sort of thing before committing to CVS.

Other than that this will probably be 2.2.12. I'm going to sleep with the build running (it takes about 80min because we have to build a kernel in the middle of it). I'll test and tag in the morning.

23 Apr 2000

Sun, 23 Apr 2000 05:26:54 GMT

And yet more boot floppies. In the past few hours I've uploaded a new yaboot package that is more likely to work on the boot floppies, a new set of kernel images (2.2.15pre19), and new PCMCIA modules for them.

I had to add mktemp to busybox in order to get ybin (YaBoot INstaller) to work on the boot floppies. This was my first run-in with the inside of busybox; I have a few ideas for how it could be improved, but it is an impressively well organized hack. Maybe I'll have time to contribute some of my ideas sometime soon.

Newt got updated today to fix the annoying Control-U brokenness; my congratulations to Enrique Zanardi for being the most responsive of all maintainers I've sent RC bugs to in weeks.

Here goes another build! This is a candidate for 2.2.12 release, if it works.

22 Apr 2000

Sat, 22 Apr 2000 19:51:19 GMT

More boot floppies... lots more. Gets through a full build on powerpc now, and we're up to 50K free - Ben says there are no problems with stripping yaboot or using -Os. I think I am not going to actually change mac-fdisk right now; I'll just document creating a bootstrap and put the rest off till woody, or at least 2.2r1. I still have an offer from Michael Schmitz to give me mac-fdisk; when I have time I will give it an update from the newer upstream version and/or just merge it into util-linux so we can use cfdisk. There's also 'perldisk' but I don't know if that would be appropriate for boot floppies.

Other goal for today is updating kernel packages again. I'm thinking of adding a prep-netboot package soon so that boot-floppies does not need to recompile a kernel every time through make release.

22 Apr 2000

Sat, 22 Apr 2000 06:20:58 GMT

Just signed up for this after peer pressure from the Debian Cabal... oh, wait... I didn't say that, you didn't hear it.

Today was the best day I've had for boot-floppies in quite a while. I got an almost functional Debian/PowerPC set uploaded (barring some screwups with libdb2 and psmisc), and I've fixed a couple dozen things since then - I hope to do another set in two days or so.

Pending for them:

Fix control-U in Newt - done, NMU pending.
Fix dependencies on the base system - done.
Add yaboot and ybin and hfsutils to the root and base disks - done, but requires ftpmaster intervention
Partitioning fixes - simplify creating a bootstrap partition - todo
Bootloaders - Actually try to CONFIGURE yaboot, or at least describe it, during the install - todo
Documentation, documentation, documentation

There's another build running now. Time will tell how much I messed up this time through.

Root.bin still fits on a floppy - by a very lucky 2112 bytes. Can you say "close call"?