Example usage:

(make-thumbnail-from-stream some-image-url (str “~/my_thumbs/” (some-image-id) “.png”) 100)

Enjoy!

Converting the Rails function “simple_format” from Ruby

def simple_format(text, html_options={}, options={})
        text = text ? text.to_str : ”
        text = text.dup if text.frozen?
        start_tag = tag(’p', html_options, true)
        text.gsub!(/\r\n?/, “\n”)                    # \r\n and \r -> \n
        text.gsub!(/\n\n+/, “</p>\n\n#{start_tag}”)  # 2+ newline  -> paragraph
        text.gsub!(/([^\n]\n)(?=[^\n])/, ‘\1<br />’) # 1 newline   -> br
        text.insert 0, start_tag
        text.concat(”</p>”)
        text = sanitize(text) unless options[:sanitize] == false
        text
      end

to Clojure (leaving out options)

(defn simple-format [text]
  (let [patterns 
      {#"\r\n?" "\n"  #"\n\n+" "</p>\n\n<p>" #"([^\n]\n)(?=[^\n])” (str #”$1″  “<br />”)}
      ]
    (str “<p>” (reduce #(clojure.string/replace %1  (key %2) (val %2)) text patterns) “</p>”)))

illustrates the fundamental naïveté of imperative programming with its silly reassignments, incidental state and mutability. It’s laughing at this and marvelling at the sheer brilliance of Clojure in equal measure that make Clojure so much fun.

Weighing the options:

One could just use an old version of the less gem that still supported the —watch option, i.e. the 1.x versions. However, the gem won’t track included/imported files, the obsolete more plugin was required for this in Ruby on Rails.

Peeking over the platform wall, the dotless compiler does all this well under Windows and it’s very fast. However, running this app with mono on Linux will not work with the —watch option, it will stop watching files after all and any file changes. This may be due to different backends of the mono FileSystemWatcher implementation on different Linux platforms, so YMMV.

Meanwhile, the author of less.js has declined to support the —watch option and has rejected at least one merge request with this functionality.

Solution:

Alas,  there’s a fork that supports watching the main file as well as all includes:

https://github.com/wvl/less.js

It works great and is stable. All you have to do is clone the repo and add bin/lessc to your path.

Now with a node installation (since recently there is a Windows installer) you should be good to go with silent and effortless server side compilation of less on all platforms.

http://blog.morroni.com/2011/06/30/debugging-scala-projects-in-intellij/comment-page-1/

You may have to restart the IDE to get the debugger to stop when running from the SBT console. Yes, you can actually run your tests on the SBT console that comes with the SBT plugin for IntelliJ (test-only etc.), set breakpoints on your test code, and the debugger will open when it hits the break points. This is really cool.

class ChangeAllColumnsToUnderscore < ActiveRecord::Migration
  include ActiveRecord::ConnectionAdapters::SchemaStatements
  require ‘active_record/connection_adapters/abstract_adapter’

  #not needed in Rails 3
  #pilfered from here:
  #http://api.rubyonrails.org/classes/ActiveRecord/ConnectionAdapters/SchemaStatements.html#method-i-column_exists-3F
  def self.column_exists?(table_name, column_name, type = nil, options = {})
    columns(table_name).any?{ |c| c.name == column_name.to_s &&
        (!type                 || c.type == type) &&
        (!options[:limit]      || c.limit == options[:limit]) &&
        (!options[:precision]  || c.precision == options[:precision]) &&
        (!options[:scale]      || c.scale == options[:scale]) }
  end

  def self.up
    tables.reject {|t| t == “schema_migrations”}.each do |table|
        columns(table.to_sym).each do |column|
          unless self.column_exists?(table.to_sym, column.name.underscore.to_sym)
            rename_column table.to_sym, column.name.to_sym, column.name.underscore.to_sym
          end
        end
        unless table_exists?(table.underscore.to_sym)
          rename_table table, table.underscore
        end
    end
  end

  def self.down
    tables.reject {|t| t == “schema_migrations”}.each do |table|
      columns(table.to_sym).each do |column|
        unless column_exists?(table.to_sym, column.name.camelize.to_sym)
          rename_column table.to_sym, column.name.to_sym, column.name.camelize.to_sym
        end
      end
      unless table_exists?(table.camelize.to_sym)
        rename_table table, table.camelize
      end
    end
  end
end

Hope it helps.

zypper addrepo http://download.opensuse.org/repositories/home:elvigia/openSUSE_11.4/home:elvigia.repo

zypper refresh

zypper install gitosis

Then continue as outlined here, slightly adjusted for OpenSuse with user creation:

groupadd git

useradd git -d /home/git -c ‘git version control’ -s /bin/sh -g git -r -m

sudo -H -u git gitosis-init < /tmp/id_rsa.pub

Then just proceed with

git clone git@YOUR_SERVER_HOSTNAME:gitosis-admin.git

on your local machine etc. etc.

apt-get install rubygems

Then revert to an older version (for example 1.3.6):

gem install rubygems-update -v=1.3.6

update_rubygems

Check your version with

gem env

In SQL Server we have the “secret” sproc sp_MSforeachtable for this purpose, but how to bend the stricter syntax of plpgslq to achieve something similar? After a number of hours of trial and error, here’s is a script that actually appears to work, at least with PostgreSQL 9.0 and probably earlier versions:

CREATE OR REPLACE FUNCTION add_version_column_to_all_tables()
RETURNS VOID
AS $$
DECLARE
    my_row    RECORD;
BEGIN      
    FOR my_row IN
        SELECT table_name
        FROM   information_schema.tables
        WHERE  table_schema = ‘public’
       
    LOOP
    IF NOT EXISTS
    (
    SELECT attname FROM pg_attribute WHERE attrelid =
    (SELECT oid FROM pg_class WHERE relname =  my_row.table_name )
     AND attname = ‘version’
     )
     THEN
        EXECUTE(’ALTER TABLE ‘ || my_row.table_name || ‘ ADD COLUMN version int NOT NULL DEFAULT 0;’);
        END IF;
    END LOOP;
END
$$
LANGUAGE plpgsql;

SELECT add_version_column_to_all_tables();

Im sure this script can be optimized, please post any improvements in the columns.

Hope it helps.

Lx-Office ist ein komplettes und recht ausgereiftes Buchhaltungs und ERP-System für Linux, ein Fork von SQL Ledger.

Installation

Mit Firefox runterladen und mit Package-Manger installieren; oder shell:

wget http://sourceforge.net/projects/lx-office/files/Lx-Office%20ERP/2.6.1/lx-office-erp_2.6.1-0-all.deb

sudo dpkg -i lx-office-erp_2.6.1–0–all.deb

Wahrscheinlich git es einige fehlende Dependencies, die im unmittelbaren Anschluss nun mit

sudo apt-get install -f

installiert werden können.

Einstellungen für Apache

Mein Webserver ist mit Phusion Passenger für Ruby on Rails ausgestattet, die Perl Skripten von LX-Office benötigen also ihren eigenen VirtualHost mit eigenem Port.

Wer nicht Passenger/Rails fährt, braucht die Einstellungen betreffs VirtualHost nicht vorzunehmen, die Einrichtung eines eigenen Ports könnte aber mittelfristig ohnehin flexibler sein.

In /etc/apache2/ports.conf:

NameVirtualHost *:8080
Listen 8080

In /etc/apache2/sites-available/default ans Ende

<VirtualHost *:8080> 
 AddHandler cgi-script .pl
 Alias /lx-office-erp/ /usr/lib/lx-office-erp/
 <Directory /usr/lib/lx-office-erp>
  Order Deny,Allow
  Allow from All
  Options +ExecCGI +Includes +FollowSymlinks
 </Directory>
 <Directory /usr/lib/lx-office-erp/users>
  Order Deny,Allow
  Deny from All
 </Directory>
</VirtualHost>

Dann

sudo a2dissite

und

sudo a2ensite

und Apache Neustart.

Einstellungen für Postgresql

Wenn noch nicht passiert, PL/pgSQL in template1 installieren:

psql -U postgres template1

createlang plpgsql template1

In postgresql.conf

tcpip_socket = true

zu setzen funktioniert nicht mehr in Postgresql 8.1 +, es muss jetzt stattdessen

listen_addresses = ‘*’

gesetzt werden.

In pg_hba.conf

Ich hatte guten Erfolg mit

local all all trust
host all all 127.0.0.1 255.0.0.0 trust
host all all ::1 ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff trust

am Ende der Datei wie in der Debian Anleitung empfohlen. Ansonsten ist diese Anleitung weitgehend veraltet.

Dann geht’s weiter mit

http://localhost:8080/lx-office-erp/admin.pl

und die Datenbank(en) werden initialisiert. Das ist eigentlich alles selbsterklärend, nur sollte man nicht vergessen, den angelegten Benutzer einer Gruppe mit Vollzugriff zuzuordnen.

YMMV

When you generate your local driver with for example

wsdl2ruby.rb --wsdl MyService.wsdl --classdef MyService --type client --force

and you instantiate your proxy driver like so

ms = MyService.new
ms.wiredump_dev = STDERR

you simply add

ms.streamhandler.accept_encoding_gzip = true

That’s all. No need for custom headers etc.

And: zlib (compression library used by soap4r) is built into Ruby since version 1.8. No need for any gems or other shenenigans.

Link List

As an added year-end bonus, here’s a list of links that helped me with SOAP and ruby over the years:

Excellent SAP article:

http://www.sdn.sap.com/irj/scn/go/portal/prtroot/docs/library/uuid/505ac48e-0578-2a10-60be-c5ffe14a0134

Correct usage of wsdl2ruby (must be wsdl2ruby.rb)

http://www.hostingrails.com/236/Using-Web-Services-Description-Language-wsdl2ruby-and-soap4r-as-a-clientI

ActionWebService:

http://www.datanoise.com/articles/2008/7/2/actionwebservice-is-back

Links:

http://www.brendonwilson.com/blog/2006/04/02/ruby-soap4r-wsdl-hell/

WSDL/Ruby:

http://codeidol.com/other/rubyckbk/Web-Services-and-Distributed-Programming/Using-a-WSDL-File-to-Make-SOAP-Calls-Easier/

Very Helpful Detailed Article:

http://www.ibm.com/developerworks/opensource/library/os-ws-rubyrails/index.html

xsltproc examples:

http://linux.byexamples.com/archives/463/xslt-processor-command-line/

xsd2wsdl:

http://jira.springframework.org/secure/attachment/12306/XsdToWsdl.xsl

linked from here:

http://jira.springframework.org/browse/SWS-79

Don’t forget to enable wiredump:

fe.wiredump_dev = STDERR

Get rid of NS1:

http://www.pluitsolutions.com/2007/08/10/remove-n1-namespace-for-soap4r/