There's an improved edition of the feedreader available. Here's what it looks like. One thing worth remembering is that installing the upgrade will override your subscriptions in the old feedreader, so make a back up if you have a huge list. Please note this bug , incase your feeds are not beeing updated.

The feedreader and blogger are slowly beeing tied together to provide an interesting publishing platform because of optional dependancies and increased management features.

The OpenACS blog title has changed from "OpenACS Hints and Tips Blog" to just "OpenACS Blog", as it has become apparent that it is no longer just being used for Hints and Tips.

There's an interesting discussion on the forum about how to liven up the community, and get things moving in the right direction.

Hi folks,

Why not turn this blog into a a stimulating web space. One thing I feel we need to change as soon as possible is the default stylesheet and add some color to it. The folks over at Coudal are constantly playing with great header pictures.

It can't be that complicated for us to add a header image, and offer a proper default 3 collumn blog layout as an option. Photodb already has code that allows pictures to be displayed on the photodb frontpage, so all we need is an extension to blog. With the stupid blog calendar removed, and the max content b

Less is more.
Who's got the code to make it as simple as possible to add the picture and update, without digging into the templating system? (combination of blog and photobd maybe) with the max content bug fixed this would create a realy nice blog package.

Maybe it's an interesting pet project for a developer, whereby duplicate data entry options are eliminated. Makes a lot of sense to me, migrating images with text. There's realy no reason to have seperate packages for such simple tasks.

Next question, who's got the best picture. Why no start a competition?

Just a thought
Ben
Questions?, chat with me now


Added bonus:

- Less documentation needed. (end-users don't read manuals!)
- Reduces bugs by combing packages as part of spring cleanup.
- More userfriendly
- More incentives for community members to add improvements.

This blog lists all available packages. The topic URL's refer to the apropiate sections of bug tracker. In order to create a roadmap for the spring cleaning everybody is free to add comments to each topic. This will result in a consolidated bug report per package, and make it easier for the package maintainer to fix things.

Once bugs are fixed, the posting can be moved draft mode, so we still have documentation about what was changed and why.

If the categories would be used throughout the toolkit, the list can also be used a a documentation manual.

This blog also is an easy newbie answer to the question, what's OpenACS, as everything is available from one page.

- this is a work in progress -

Vinod has posted a diff proc for comparing and displaying changes between two revisions of text. Check it out here.

Feel free to add any feature request.

# Retrieve package_id of organizations package
set node_id [site_node::get_node_id_from_object_id -object_id [ad_conn subsite_id]]
set package_id [site_node::get_children -node_id $node_id -package_key organizations -element package_id]

Nima has made a script available that helps integrate an existing package into dotLRN. Details of this script are available here.

This forum posting provides some excellent pointers for expanding and using acs-subsite based member roles.

How to configure AOLServer behind a proxy (Taken from this forum posting)

If you have a installation of OpenACS that is behind a proxy you will soon realize some stuff will break:

* Installation of modules from OpenACS repository
* News Agregator
* Any application using ns_httpopen, ns_sock...

In order to fix this you have to change two files:

* config.tcl
* [AOLSERVER]/modules/tcl/http.tcl

Changing your server configuration

Go to the section ns/parameters and add these lines under the section:

ns_param proxy your-cache.server
ns_param proxy_port xxxx

Changing the proc ns_httpopen

In the file [AOLSERVER]/modules/tcl/http.tcl go to the line

set fds [ns_sockopen -nonblock $host $port]

And change it to

set proxy [ns_config ns/parameters proxy $host]
set proxy_port [ns_config ns/parameters proxy_port $port]
set fds [ns_sockopen -nonblock $proxy $proxy_port]
set uri http://$host$uri

Now if you restart your server it will work just fine.

Thanks to Stefan Vogel for the fix [1].

[1] http://www.mail-archive.com/aolserver@listserv.aol.com/msg06933.html

Something like this should do the trick

-groupby { values { "Day" { { groupby time_stamp } { orderby time_stamp,desc} } } }

The list is { groupby_name { { groupby column } { orderby column,order } } }

Add this to the query as any other filter with template::list::filter_where_clauses.

You can pass as many lists in to the values parameters of group by as columns you want to group on. If you add the listfilters tag to your ADP file, it will show the available grouping options as filters. Make sure you add groupby:optional in ad_page_contract.

Another trick is to hide the column when it is used to grouping. Just dynamically set the hide_p paramter for the list element that is in groupby.

To replace the annoying ^M characters you can search and replace. The following representation holds true:

^M = C-q C-m

Resulating in this sequence:

M-%
Query replace: C-q C-m with: C-q C-j

I just came accross this instruction about how del.icio.us allows you to keep a rolling collection of bookmarks and reuse them on your website by retrieving them back via RSS. Jeffrey Veen describes how to do this easily using another external tool, but, hey, you've got your on server and OpenACS − this quick tutorial by Steffen Tiedemann Christensen shows you how to add del.icio.us links to your OpenACS server using an adp include.

Again using Launchd in MacOS X Tiger you can do away with daemontools.

Just copy/paste the following into a file called org.openacs.OpenACS.plist.

################# COPY BELOW ###########################
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple Computer//DTD PLIST 1.0//EN" "http://www.apple.com/DTD
s/PropertyList-1.0.dtd">
<plist version="1.0">
    <dict>
        <key>Label</key>
        <string>org.openacs.OpenACS</string>
        <key>OnDemand</key>
        <false/>
        <key>ProgramArguments</key>
        <array>
            <string>/var/lib/aolserver/openacs/etc/daemontools/run</string>
        </array>
        <key>ServiceDescription</key>
        <string>OpenACS Service</string>
        <key>UserName</key>
        <string>ncarroll</string>
        <key>GroupName</key>
        <string>web</string>
    </dict>
</plist>
################# COPY ABOVE ###########################

Then use the command below to start up the service:

$sudo launchctl load /Library/LaunchDaemons/org.openacs.OpenACS.plist

And the command below to stop the service:

sudo launchctl unload /Library/LaunchDaemons/org.openacs.OpenACS.plist

Couldn't be simpler.

I just installed PostgreSQL on Mac OS X 10.4 Tiger, and found that the PostgreSQL StartupItem provided in the acs-core-docs no longer works.  Instead use Tiger's Launchd system, which is so much simpler than StartupItems in previous Mac OS X releases.

Creating a PostgreSQL Launchd file allows your system to start up postmaster when your Mac boots.

This tip shows you how to create this file.

Use a text editor such as emacs to create the PostgreSQL Launchd file:

$ emacs /Library/LaunchDaemons/org.postgresql.PostgreSQL.plist

Copy and paste the following into the file.

################# COPY BELOW ###########################
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple Computer//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
    <dict>
        <key>Label</key>
        <string>org.postgresql.PostgreSQL</string>
        <key>OnDemand</key>
        <false/>
        <key>ProgramArguments</key>
        <array>
            <string>/usr/local/pgsql/bin/pg_ctl</string>
            <string>-D</string>
            <string>/usr/local/pgsql/data</string>
            <string>-l</string>
            <string>/usr/local/pgsql/data/server.log</string>
            <string>start</string>
        </array>
        <key>ServiceDescription</key>
        <string>PostgreSQL Server</string>
        <key>UserName</key>
        <string>postgres</string>
        <key>GroupName</key>
        <string>web</string>
    </dict>
</plist>
################# COPY BELOW ###########################

Next reboot your Mac and check to see if postmaster started up.

Or you can try the following commands:

$ sudo launchctl load /Library/LaunchDaemons/org.postgresql.PostgreSQL.plist
$ sudo launchctl start org.postgresql.PostgreSQL

I make silly mistakes when coding that I don't notice right away, and then a few days later I'll notice that something's not working as expected and it could take hours for me to figure out what's wrong. So...how do I save me from myself? Good question," says daveb, "Keep logs of what you do" (As in, backed up postgres. fixed bug 22. Installed photo-albem, and so on.) Even better, those who run a command-line through the emacs editor, can auto-log their shell buffers to file. After you enter shell-mode with "M-x shell" or "C-x C-m" or "C-x RET" or whatever (To see what the keyboard short-cut is on any of the command names, "M-x where-is"), the buffer is called "*shell*" and it has no file associated with it. All you have to do is write that buffer into a file that you pick and you'll have a complete log of what you did -- I usually save the file as ~/log/20050605_shell or something. Optionally, you can then rename the buffer back to "*shell*" (because writing to a file will automatically rename the buffer name to the file name.) To write a file, "M-x write-file" which is usually bound to "C-x C-w". To rename the buffer, "M-x rename-buffer". You need to make sure that your file is in auto-save-mode and that you have selected an appropriate value of auto-save-interval which is, I think, the number of key strokes between auto saves. I use 100.

One thing I've done is create a script in my ~/bin directory called debug. The contents are simply:

tail -f /path/to/aolserver-error.log |grep "--"

Then while I'm working on OpenACS, I use ns_log Notice "--myvar: $myvar" to see pseudo-breakpoints in the code.

This guide should help you with configuring OpenACS Notifications to use Postfix instead of the qmail MTA on Macos X. However the instructions should be similar for OpenACS installations on other OSes.

Once postfix is installed, all you need to do is enable it to start as your default Mailserver in /etc/hostconfig. Proceed with configuring Postfix, then onfigure OpenACS Notification Parameters to use Postfix. Finally, restart Postfix for the configurations to take effect. You may also need to restart your OpenACS server as well.

NOTE 1: Replace hostname.mydomain.tld with the hostname of your server. This should be something like openacs.org.

NOTE 2: I've assumed that the OpenACS server will run as the service0 user.

NOTE 3: You need to create the Maildir directory in the home directory of the service0 user.

$ mkdir /Users/service0/Maildir

Edit /etc/hostconfig to enable Postfix on Panther

MAILSERVER=-YES-

Edit /etc/postfix/main.cf

myhostname=hostname.mydomain.tld
myorigin=$myhostname
inet_interfaces=$myhostname, localhost
mynetworks_style=host
virtual_maps=regexp:/etc/postfix/virtual
home_mailbox=Maildir/

Edit /etc/postfix/virtual

Add the following regular expressions:

/^notification-(.*)@hostname\.mydomain\.tld/ service0
/^bounce-(.*)@hostname\.mydomain\.tld/ service0

Edit /etc/postfix/master.cf

Uncomment line 77 and change:

#smtp inet n - n - - smtpd
to
smtp inet n - n - - smtpd

OpenACS: Configure Notifications Parameters

EmailDomain: hostname.mydomain.tld
EmailQmailQueue: /Users/service0/Maildir
EmailReplyAddressPrefix: notification
EmailQmailQueueScanP: 1

OpenACS: Configure Mail Services Lite Parameters

BounceDomain: hostname.mydomain.tld
BounceMailDir: /Users/service0/Maildir
EnvelopePrefix: bounce

Restart Postfix

$ sudo Postfix stop
$ sudo Postfix start

What about sending a bulk mail message to all community members that we have created an end users welog?

Instead of linking to other websites, I am using the blogroll feature in Larsblogger to create shortcuts to the most important functionality of the site I am creating. It would be usefull to have an option to add additional blogrolls, just like we can add more category trees. That way the blog keeps the traditional blog features also.

The "How to" category could be used to create goal centred approach similar to the wasy the folks at Basecamp are roling out "chunks" of usefull features. like their Tada and Backpack tools, people are likely to be looking for.

One of the most useful tools in Larsblogger is the bookmarklet. On eTNweblog I monitor travel industry news, and adding stories to my blog is now a matter of seconds. I receive a Google allert, check the the story for relevance, highlight the the first 5 sentences and click on bookmarklet in my browser toolbar, which automaticaly pastes header, highlighted body copy and URL to the my blog. Works like a charm. All you have to do is drag the bookmarklet (go to blog admin) .

At the moment there are 2 stylesheets for larsblogger. This default one, and one I think only available if you install DotLRN, with a much more attractive calendar display. Who knows where it is located, and is there a simple way to change between the 2 designs? I also think we should have a third default stylesheet based on 3 collumn with header picture.

Just activated categories. If you have ideas for more topics, let me know.

As a designer / manager using OpenACS, I find myself wanting more information on the best ways to use OpenACS (rather than develop it). I'd love it if there are people out there managing sites who have tips or hints for working with OpenACS from a administrator's perspective. I agree with Ben Koot's post - I feel like it is an iceberg, and I don't really know how to get the most out of OACS yet. Thanks!

OpenACS has some really powerful features that we would like to expose through this blog. We will try to present features that will be of immediate use to OpenACS developers. However before diving into the deep end, a few basics of OpenACS will be covered within this posting for reference.

OpenACS Packages

OpenACS has transformed from a monolithic code base (3.x and below) to a neater modularised code base in the latest releases (4.x and above). Applications that use the OpenACS framework are now developed in self-contained modules called "packages". All packages are structured using the Model-View-Controller (MVC) pattern.

There are core packages such as the kernel, templating and authentication packages that give you the basis for a web application that supports user, group and application management. You can extend the functionality of a vanilla installation of OpenACS by plugging in already developed packages such as a calendar or wiki through a package installer interface. Or if you can't find a package that provides the functionality you need from OpenACS's extensive package catalog, then you can easily develop your own package.

A package contains an XML based .info file containing meta-data for the package. The core packages use the .info file for dependency checking, as well as for defining properties and specifying call-back commands. The package can also have an SQL directory that contains all SQL files for generating the data model for the application. There is a TCL directory that contains all library procedures for the application. Finally, there is a WWW directory that contains all the presentation logic for the Web application. The structure of the notes package is as follows:

notes

    + notes.info

    + sql

        + oracle

        + postgresql

            * notes-create.sql

            * notes-drop.sql

    + tcl

        * notes-apm-callbacks.tcl

        * notes-init.tcl

        * notes-procs.tcl

    + www

        * index.adp

        * index.tcl

        * index.xql

        * index-postgresql.xql

        * index-oracle.xql

In the SQL directory, the notes-create.sql file is called only once when the application is installed, and is used to create the data model in a database. Currently there is support for both Oracle and PostgreSQL in OpenACS, so depending on which RDBMS you are using, the notes-create.sql will be executed from the SQL sub-directory that corresponds to the RDBMS used.

In the TCL directory, the notes-apm-callbacks.tcl file contains procs that get called when the application is installed/mounted. The notes-init.tcl contains TCL procs that initiate the application when AOLServer starts up. The notes-procs.tcl contains TCL procs that can be used by files in WWW or by other packages.

In the WWW directory, each web page consists for an XQL file, a TCL file and an ADP file. The ADP file contains HTML and templating tags for presentation of data from the data base. The XQL file is an XML based file that contains SQL statements that can be used by the TCL file. SQL statements common to all database systems can be placed in file.xql, whereas RDBMS specific SQL statements can be placed in file-postgresql.xql or file-oracle.xql. The TCL file contains the logic for executing SQL statements and then pushing the returned data to the ADP template file.

That just about covers OpenACS and packages from a high-level overview. In future posts we will start blogging about more useful features in OpenACS.

This is a blog intended for use by OpenACS users. Share your latest tips and hints here!