Jade's weblog Archive

Let's begin the diagnosis.

First of all, Macintouch has a very worrysome article on Firewire in 10.3. It looks like there were some very serious bugs in the firmware or in 10.3 that are causing data loss and corruption in external Firewire disks. Great. Let's cross our fingers that that is not what happened... although I am concerned it might be exactly that.

http://www.macintouch.com/panfirewire.html

Well, it looks like I am in luck. The drive I have is a dual connection, so I was able to connect to it via USB. Perhaps the Firewire cable went bad?

ps -ef --forest | less

Cool, except it doesn't work with Preview. You must open your final label in Acrobat Reader.

But after I figured that out, it worked beautifully!

Useful for Ajax web applications.

Fun.

Okay, let's see:

cd
mkdir bash3
cd /usr/local/src
wget http://ftp.gnu.org/gnu/bash/bash-3.0.tar.gz
tar xzf bash-3.0.tar.gz
cd bash-3.0
./configure --prefix=/home/jade/bash3
make install

Wow, wasn't that fun.

Now use /home/jade/bash3/bin/bash instead of the normal bash and it should work.

It didn't.

Tonight, I took the test again, and found that I was an ENTJ this time.

But only barely. I am almost perfectly on the border for E and T (1% preference), and only have a slight preference for the J as well (22%). For the N portion, I have a distinctively expressed (75%) preference for N.

I think the description for the ENTJ is pretty accurate.

So the types I am now closest to (besides ENTJ) are:

ENFJ Teacher
INTJ Mastermind
ENTP Inventor (where I came from ten years ago)

sqlplus: error while loading shared libraries: /u01/app/oracle/oracle/product/10.2.0/db_1/lib/libnnz10.so: cannot restore segment prot after reloc: Permission denied

solution:
root
tail -f /var/log/audit/audit.log | tee oracle.log
oracle
sqlplus / as sysdba
root
crtl c to abort the tail
audit2allow -M oracle < oracle.log
semodule -i oracle.pp
that's it

Got errors in the upgrade on my staging server, so rebuild the database and did every stinking upgrade script manually. I should note that I do appreciate the work that these developers do, and many of these folks are very good programmers. But I did want to post where the problems occured, and who would be responsible for fixing them.

• acs-content-repository (author davis@xarg.net) -- harmless because it recreates these foreign key constraints.
  psql -f upgrade-5.2.0d2-5.2.0d3.sql
  psql:upgrade-5.2.0d2-5.2.0d3.sql:4: ERROR: constraint "cr_folder_type_map_fldr_fk" does not exist
  ALTER TABLE
  psql:upgrade-5.2.0d2-5.2.0d3.sql:9: ERROR: constraint "cr_folder_type_map_typ_fk" does not exist
  ALTER TABLE

• acs-content-repository (author timoh or daveb)
  psql -f upgrade-5.2.0a1-5.2.0a2.sql
  psql:upgrade-5.2.0a1-5.2.0a2.sql:221: ERROR: cannot drop view download_archive_descsi because other objects depend on it
  HINT: Use DROP ... CASCADE to drop the dependent objects too.

  I tried changing the text from create view to create or replace
  view in case that would help. I have no idea if it will.

• acs-content-repository (author daveb)
  psql -f upgrade-5.2.1d1-5.2.1d2.sql
  psql:upgrade-5.2.1d1-5.2.1d2.sql:930: ERROR: cannot drop view download_archive_descsi because other objects depend on it
  HINT: Use DROP ... CASCADE to drop the dependent objects too.

• news-aggregator (this package never worked for me anyway, so I dropped it using their drop scripts and rebuilt it. There was a typo in one of the drop scripts.) It seems like this might actually be functional now... cool.

arch_commit: unable to acquire revision lock (could not rename file.)
tree: /home/safe4all-dev/safe4all-dev
revision: jade@safe4all.org--2005/openacs--dev--5.1.0--patch-11

Tried to get the permissions right, but that wasn't the problem.

The issue is that you need a directory in

/{archives\}/2005/openacs/openacs--dev
/openacs--dev--5.1.0/patch-(highest-number-here)

called ++revision-lock
and a directory under that called
+contents

If that's not there, no dice.

So I'm going through the annoying process of switching.

By the way, if someone from Tektonic reads this, feel free to reply by leaving a comment on this posting. :)

Jade

So I created a little script to send the email gradually (so as to not hit Gmail's filters):

> cat mysendmail
sleep 10
/usr/sbin/sendmail myemailaddress@gmail.com

Then I send it through

grep -v ^Delivered-To: /home/safe/testmbox | formail -s /home/safe/mysendmail &

The solution I've used is:

sqlplus user/pass
purge recycle bin;

then recompile invalid objects

Courtesy of MB