<oracle>
begin
if '#item.displayoption#' = 'link' then
   htp.p('<a href="#item.url#">View item in current window</a>');
elsif '#item.displayoption#' = 'linktonewwindow' then
   htp.p('<a href="#item.url#" target="_blank">View item in new window</a>');
else
   htp.p('Display item directly in page');
   htp.br;
   htp.p('#item.content#');
end if;
end;
</oracle>

On the surface that looks very good, but what if you got regions that shall support both links and text items ? The problem of referring to #item.content# (or the #item.content.value#) multiple times, is that the generated code allocates space for the length of #item.content# each time it is referred. For instance, if your region contains both links and a 15k text item, and you refer to #item.content# three times, the space used will be 45k + the size of the links. This is far above the PL/SQL 32k limit, and your portal page will not render properly, and you will get the following error thrown in the browser:

Internal Error (WWC-00006)
Execution of the PL/SQL-block failed: , (WWC-40015)
Preference path not found: oracle.portal.wsrp.<YOUR_PORTLET_1_NAME> (WWC-51000)
Preference path not found: oracle.portal.wsrp.<YOUR_PORTLET_2_NAME> (WWC-51000)
Preference path not found: oracle.portal.wsrp.<YOUR_PORTLET_3_NAME> (WWC-51000)
Preference path not found: oracle.portal.wsrp.<YOUR_PORTLET_4_NAME> (WWC-51000)
and so on, for each of the portlets mounted on your page.

The solution is to use more PL/SQL code, and less HTML Content Layout Substitution tags:

<oracle>
DECLARE
  p_target VARCHAR2(6);
  p_content VARCHAR2(32000);
BEGIN
  p_content := SUBSTR('#item.content.value#',1,32000);
  htp.p('#ITEM.STATUSINHTML#');
  if '#item.displayoption#' IN ('link','linktonewwindow') then
    if '#item.displayoption#' = 'link' then
      p_target := '_self';
    elsif '#item.displayoption#' = 'linktonewwindow' then
      p_target := '_blank';
    end if;
    htp.p('<a href="'||p_content||'" target="'||p_target||'">#ITEM.TITLE.VALUE#</a>');
  else
    htp.br;
    htp.p(p_content);
  end if;
END;
</oracle>

This way, the potentially huge #item.content.value# is referred to only once, and it’s also made sure it never exceeds 32k. (You cannot create text items larger than 32k through the Rich Text Editor or through the Portal PL/SQL API. But since text items are saved in CLOBs, they could potentially have been populated through a custom API and be larger than 32k).

To improve performance with Oracle Text based searches in Oracle Portal, you should on a regular basis make sure to syncronize and optimize Portal text indexes (if this is not allready set up during installation).

sqlplus portal/<password>

exec wwv_context.sync
exec wwv_context.optimize
exit

Unless you make an full export with grants you won’t get all grants when doing a export/import with Oracle. So if you don’t have a list of all grants across those schemas you imported, you will have a huge job of compiling and establishing which grants are missing.

The solution to this is to find those grants through the data dictionary on the source system. For instance, if you want to find all grants made by the PORTAL schema, the query would be like this:


SET HEADING OFF
SET PAGES 999
SPOOL grants.sql
SELECT 'GRANT '||privilege||' ON '||owner||'.'||table_name||' TO '||grantee||DECODE(grantable,'YES',' WITH GRANT OPTION;',';')
FROM DBA_TAB_PRIVS
WHERE grantor = 'PORTAL';
SPOOL OFF
EXIT


Now you have everything you need in the grants.sql and this script can now be run on the target system.

To avoid quoting quotes in string, Oracle 10g offers the Q-Quote technique.

Let’s say yoy want to select the following from dual: I’m into rock’n’roll

This would normally mean you should quote the three quotes, but with the Q-Quote technique it’s much simpler:

SQL> SELECT q'[I'm into rock'n'roll]' FROM dual;

VersionTracker is much like download.com, except it also has a large collection of software for Mac OS X. You browse by categories, and can sort by licenses, so it’s easy to find software at a decent price (for instance free).

I was however not satisfied with any of the FTP clients I found for Mac. As a devoted FileZilla user, I didn’t think any of the FTP/SCP clients for Mac was living up to the standard. Furtunatly the FileZilla project has decided to release FileZilla on Linux and Mac too, and there is allready a beta available

A really nice open source text editor with syntax highlighting is Smultron.
]]> http://far2fish.wordpress.com/2009/05/25/software-for-your-mac/feed/ 0 far2fish :) http://far2fish.wordpress.com/2009/05/25/mac-keyboard-combinations-on-a-windows-keyboard/ http://far2fish.wordpress.com/2009/05/25/mac-keyboard-combinations-on-a-windows-keyboard/#comments Mon, 25 May 2009 10:54:22 +0000 far2fish http://develix.net/?p=32 ]]>

I recently bought a Mac mini and a Belkin KVM switch, which enables me to use my existing monitor, keyboard and mouse with both my current Windows PC and my new Mac mini. There are however quite a few characters I couldn’t find when using the Windows keyboard on Mac, so I will use this thread to remember them. Thanks to Allan for supplying many of the keystrokes for a Windows keyboard with Danish layout.

1. Commercial at @ : ALT Gr-* (just ALT-* in some applications)
2. Pipe | : ALT-i
3. Left brace { : SHIFT-ALT-8
4. Right brace } : SHIFT-ALT-9
5. Apple Command Key: Windows Key
6. Less than < : §
7. Greater than > : SHIFT-§
8. Copy : Windows Key-C
9. Cut : Windows Key-X
10. Paste : Windows Key-V
11. Quit Application : Windows Key-Q
12. Dollar $ : <
13. Euro € : SHIFT-4
14. Backslash \ : SHIFT-ALT-7

I also found this article, which suggests a software solution to the problem. When I get a Mac keyboard, I should study this. The layout of a Mac keyboard can also be useful.
]]> http://far2fish.wordpress.com/2009/05/25/mac-keyboard-combinations-on-a-windows-keyboard/feed/ 0 far2fish http://far2fish.wordpress.com/2009/05/25/taking-screenshots-on-a-mac/ http://far2fish.wordpress.com/2009/05/25/taking-screenshots-on-a-mac/#comments Mon, 25 May 2009 10:53:48 +0000 far2fish http://develix.net/?p=30 ]]>

• Command-Shift-3: Take a screenshot of the screen, and save it as a file on the desktop
• Command-Shift-4, then select an area: Take a screenshot of an area and save it as a file on the desktop
• Command-Shift-4, then space, then click a window: Take a screenshot of a window and save it as a file on the desktop
• Command-Control-Shift-3: Take a screenshot of the screen, and save it to the clipboard
• Command-Control-Shift-4, then select an area: Take a screenshot of an area and save it to the clipboard
• Command-Control-Shift-4, then space, then click a window: Take a screenshot of a window and save it to the clipboard

These screenshot tricks where found in MacRumors.
]]> http://far2fish.wordpress.com/2009/05/25/taking-screenshots-on-a-mac/feed/ 0 far2fish http://far2fish.wordpress.com/2009/05/25/add-a-new-harddisk-to-linux-using-parted/ http://far2fish.wordpress.com/2009/05/25/add-a-new-harddisk-to-linux-using-parted/#comments Mon, 25 May 2009 10:53:17 +0000 far2fish http://develix.net/?p=28 ]]>

If you have added a new harddrive, and want to take advantage of it in Linux, parted is a great tool to get you started. Parted is like fdisk, but a bit more challenging. Please note that when following this tutorial you will loose any data on the harddrive in question.

Finding the device name

The first task is to find the devicename of your new harddrive. SCSI disks are usually given device names like /dev/sda, /dev/sdb, /dev/sdc, /dev/sdd and so on (dependening of how many phsyical disks you have got. IDE disks are usually given device names like /dev/hda, /dev/hdb, /dev/hdc, /dev/hdd and so on. In this tutorial we will image we have added a second SCSI disk, we will treat it as /dev/sdb

Using parted

As root, type
parted /dev/sdb
and parted will now startup, expecting to perform work on our new harddrive.

First we create a disk label
(parted) mklabel msdos

Then we need to find the size of the disk, or more correctly the start- and end sector of the disk
(parted) print

The print command should display something like this:

Disk geometry for /dev/sdb: 0.000-102400.000 megabytes
Disk label type: msdos
Minor Start End Type Filesystem Flags

The geometry data will come in handy in our next command where we create a filesystem on the disk
(parted) mkpart primary ext2 0.000 102400.000

We can also make extra sure that we get a ext2 disk with the following command (please note that parted doesn’t support ext3):
(parted) mkfs 1 ext2

Now exit parted:
(parted) quit

Mounting the new drive

First create the mountpoint that you want to use for the new drive:
# mkdir /u01

Change the filesystem from ext2 to ext3:
# tune2fs -j /dev/sdb1
Now edit /etc/fstab and add the following line:
/dev/sdb1 /u01 ext3 defaults 1 1

The mount the new drive on mountpoint /u01:
# mount -a

Finally check that the new drive is mounted:
# df -k

Filesystem 1K-blocks Used Available Use% Mounted on
/dev/sda3 5526416 2081024 3164660 40% /
/dev/sda1 101086 8613 87254 9% /boot
none 387852 0 387852 0% /dev/shm
/dev/sdb1 10317828 33460 9760252 1% /u01