« January 2007 »
SunMonTueWedThuFriSat
 
1
2
3
6
7
16
20
27
30
31
   
       
Today
XML

Tom Haynes

loghyr.com
excfb.com

Blogs to Gander At

Navigation

Editing

AllMarks

Referers

Today's Page Hits: 1968

Powered by Roller Weblogger.

statcounter.com

clustrmaps.com

Locations of visitors to this page

technorati.com

www.alesti.org

Add to Alesti RSS Reader

South Park as I was 10 years ago

South Park Fantasy

South Park today

South Park Reality

I have more hair and it isn't so grey. :->

10 years ago, really

Toon Tom

Today, literally

Tom Today

Site notes

This page validates as XHTML 1.0, and will look much better in a browser that supports web standards, but it is accessible to any browser or Internet device. It was created using techniques detailed at glish.com/css/.

« Previous day (Jan 25, 2007) | Main | Next day (Jan 27, 2007) »
20070126 Friday January 26, 2007
One of my old managers is a Jedi Padawan

I'm not making this up - one of the skins in Jedi Knight: Jedi Academy by Lucas Arts is an old boss of mine - Paul Crozier. If you know Paul, you would have to agree:

Not shown

The fun part will be passing this on to his current direct reports and letting them speculate on whether he is on the Light or Dark Side.


Originally posted on Kool Aid Served Daily
Copyright (C) 2007, Kool Aid Served Daily
How to tell you unpacked the wrong architecture for the closed binaries

Okay, I must have done this twice, on different machines. The make fails, telling me that it can't build the closed source. I've unpacked the closed source, I don't think it is in the right place, but it does agree with what the opensolaris.sh states. The logs are pretty generic in the mail message:

The following command caused the error:                                                                                        
dmake: Warning: Don't know how to make target `/zoo/ws/onnv-gate/proto/root_i386/lib/libc_i18n.a'                              
dmake: Warning: Target `libc.so.1' not remade because of errors                                  

I think it is the wrong SUNWonbld, the wrong compile options, etc. I start over and get it to work on the slow machine - why? I don't know. But I sure do try and clone everything on the fast machine.

On the slow machine, this can be a 3-5 hour process. On the fast one, 1-2 hours. 30 minutes if I don't clobber everything.

I still don't know what tweaked me on to the problem. I thought it was something in the log file, but in looking back, I can't see jack. I can see this in my history file:

    38  13:15   ls -la closed/
    39  13:15   ls -la closed/root_sparc/

It is easy to spot if you know I'm building on an w2100z, i.e., an amd64 box.

So the other clue in the error logs is:

==== Build warnings (DEBUG) ====

Warning: closed binaries not properly unpacked.
                echo "Warning: closed binaries not" \
                echo "Warning: libc_i18n.a must be built" \

First, make sure you did unpack the closed encumbered stuff. If you did, then make sure you did it for the right architecture. I suspect if you have them both there, things will just work.


Originally posted on Kool Aid Served Daily
Copyright (C) 2007, Kool Aid Served Daily
First stabs at building OpenSolaris from source

One thing I noticed in the BFU instructions (see Section 5 - Installing and Testing ON and Dennis Clark at How To Build NON-debug OpenSolaris - { updated for snv_20060925 }) was that you need to set some environment variables before you can proceed. Actually, it seems like the Developer's Reference at OpenSolaris.org (the Section 5 link above) doesn't mention this at all. Dennis provides the following:

# PATH=/opt/SUNWspro/bin:/opt/onbld/bin:/usr/ccs/bin:/usr/sbin:/usr/bin
# export PATH
# FASTFS=/opt/onbld/bin/`uname -p`/fastfs;export FASTFS
# BFULD=/opt/onbld/bin/`uname -p`/bfuld;export BFULD
# GZIPBIN=/usr/bin/gzip;export GZIPBIN 

Now I got bit by this at work, in preparing the Connectathon 2007 build boxes. This will lead to an additional warning:

/tmp/bfu.628850[2428]: /net/onnv.eng/export/gate/public/bin/acr: cannot open
chmod: WARNING: can't access /tmp/bfubin/acr

You could bypass this by entering an additional environment variable:

# ACR=/opt/onbld/bin/acr

You will still get a warning about the following:

Marking converted services as enabled ...
cp: cannot access /net/greenline.eng/meta0/smf/post-5090532/sysidtool.xml
bfu: could not copy /net/greenline.eng/meta0/smf/post-5090532/sysidtool.xml
cp: cannot access /net/greenline.eng/meta0/smf/post-5090532/kdmconfig.xml
bfu: could not copy /net/greenline.eng/meta0/smf/post-5090532/kdmconfig.xml

Okay, despite these warning messages, how did it come out?

[tdh@mrx grub]> uname -a
SunOS mrx 5.11 kanigix-20060123-nd i86pc i386 i86pc

The BFU script is such a big deal in OpenSolaris development and testing. You shouldn't have to setup these environment variables and you shouldn't get the warning messages. Time to roll up some sleeves and suggest a fix.


Originally posted on Kool Aid Served Daily
Copyright (C) 2007, Kool Aid Served Daily

Copyright (C) 2007, Kool Aid Served Daily