that’s whacked

build 23 done!

Wednesday, Sep 28. 2005  –  Category: OpenSolaris

I’ve finished the source drop/bfu archives/closed-bins/etc. for Nevada build 23. It should be up on SDLC soon (tonight hopefully).

I’ve also got a putbacks log of all the changes that happened between builds 22 & 23. I’ve got the previous log of the changes between builds 20 & 22 as well.

forcing the issue

Wednesday, Sep 28. 2005  –  Category: OpenSolaris, OpenSource

I was thinking about user interfaces to all the various commands and utilities I use today, and it amazes me how many programs have a -f, or –force option.

What’s even more astounding is how often I have to use it.

Why? Doesn’t this speak to a more inherent design flaw? Anytime I have to “–force” my program’s behaviour from its normal codepath, then I’m bypassing the program’s logic. This is a concession to the “user always knows best” paradigm. In an ideal world, our software and hardware should act in conjunction to take these kinds of choices away from users. Yeah yeah, the Matrix, blah blah. But seriously, I shouldn’t have to override my program’s default decisions - the program should know what I want.

This is a design flaw that many Unix OS’s suffer from, Solaris included. We have a group, KISS (Keep it Simple Solaris) that is working on this, and other approachability issues. I for one would love to see all these stupid “–force” and “-f” options go away.

don’t panic?

Monday, Sep 26. 2005  –  Category: Musings

panic[cpu288]/thread=2a10001fcc0: RAMI KJ@SPDIáPRT1SPDpCPU0CONðCPU1CONðCPU0DCB¨ ðCPU1DCB¨CPU0STK @CPU1STK ,@CPU0XIR L@CPU1XIR l@DLDBUFR@@JTAGDAhÿÿÿÿøJýIá!ºPST0MBXÃPST1MBXÃXIR1CMh ÿÿÿÿJýIáÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿI§JCOøSCTLDCDhøSTARLDpanic[cpu3213/thr=ad=300a293eb40: JýLMP8anic[cpu291]/thread=2a100441cc0: JXLMP8MP8TRAP: type=31 rp=30 panic[cpu321]/thread=300a293eb40: BAD TRAP: type=34 rp=18c6970 addr=ffffffffffffffff mmu_fsr=0

panic: entering debugger (continue to reboot) Type ‘go’ to resume {120} ok

hrm… well. that’s one of the more cryptic panic messages i’ve certainly seen.

courant.com | JetBlue Airliner Lands Safely After 3-Hour Ordeal

Thursday, Sep 22. 2005  –  Category: Musings

courant.com | JetBlue Airliner Lands Safely After 3-Hour Ordeal

damn. not to sound crass or anything, but that pilot has some serious balls.

SEED

Wednesday, Sep 21. 2005  –  Category: Sun

this is about a week late, but oh well…

last week we had our SEED showcase, the end of year meeting for our SEED program (one of Sun’s mentoring programs, and how i met geoff, my mentor).

some highlights? roughly in order…

• hearing Tom Chalfant’s talk on PTS (Product Technical Support). i always wondered what this guy did.
• hearing Keith Whisnant talk about system telemetry (used in our big machines like the Starcat). i actually heard his talk at the UCSD recruiting event last year, but Keith is one of the more engaging speakers i’ve heard come out of our engineering departments. you can tell he’s really enthusiastic about his work.
• the panel of DEs (2), VPs (2), and director (1) talk about ‘how to become a Sun executive’. this one was really interesting. one of the questions i first asked geoff when we started our mentoring relationship was whether breadth or depth was more important in his career. he answered (if i recall correctly, i don’t have my notebook in front of me ATM) that both were necessary. this may not be surprising given he’s the DE of HR (his words, not mine). interestingly, the two DE’s (Shueling Chang-Shantz, Poorna’s mentor, and Mike Wookey) both answered resoundingly depth was more important. whereas the 2 VPs & director answered breadth, with certainly some depth in your area of expertise. so now the question for the SEEDlings is, do we follow the technical track (aspiring to become DEs) or follow the managerial track? tough choice… and one i’m sure i’ll blog about in the future.
• the tour of the Sunnyvale compute ranch. holy crap this place is cool. this is the datacentre of your dreams. clean rows of machines, neatly run cabling, one system admin, etc. etc. part of it is helped by a uniform architecture (all fairly homogeneous SPARC machines), and doing a focused task: job queues of EDA tasks. so not quite your general purpose compute ranch - but still pretty darn cool.
• Ron Goldman’s talk on open source at Sun.
• Radia Perlman’s talk on her encrypted file system.

i should blog more details i know, but maybe in the future… for now i’ve got to go get build 23 delivered..

Slashdot | IBM Training Employees To Leave IBM?

Friday, Sep 16. 2005  –  Category: Computers

Slashdot | IBM Training Employees To Leave IBM?

wow. talk about far-sighted. this is really really cool… i know IBM is our evil competitor and all, but kudos to them. i think this is a great idea…

what comes after 22?

Thursday, Sep 15. 2005  –  Category: OpenSolaris

well it depends. historically (numerically?), 23 would come next. but in terms of OpenSolaris build deliveries, we seem to have some sort of aversion to the odd numbers (having gone from 18 to 20 to 22).

i’m happy to say i’m actually trying to stay on top of things and deliver 23. 22 took a little longer than i would have liked due to the crypto code opening. (and i still managed to screw that up slightly. thanks Juergen.

so what’s up for 23?
i’m planning on opening up the cryptokit stuff (which should eliminate that dprov warning on your nightly log files), as well as cmd/cmd-inet, which should open up rsh, telnet, rlogin, rcp, etc.

there are, of course, many more things on the board to open up - but kupfer is working on some stuff in the usr/src tree, specifically trying to putback all our tonic changes to the main Nevada gate (which should enable future deliveries to be nightly instead of weekly or bi-weekly), and i don’t want to tread on his toes and mess things up for him.

phewww

Monday, Sep 12. 2005  –  Category: Musings

thank god for backups. my blog is back.

putback logs for snv_22

Monday, Sep 12. 2005  –  Category: OpenSolaris

i just posted the putback logs of everything that changed between our 20050818 (build 20) and 20050909 (build 22) source drops.

i’m a retard

Monday, Sep 12. 2005  –  Category: Musings

let it be known, thou shouldst tread carefully when messing around with MySQL and the DROP command.

sigh.

yes. i deleted my blog.

yes, i really am stupid.

yes, i intend on trying to recover it tonight from a backup.