Rhaamo's world

Photon: A Rails3 Photo Gallery

2010-08-12T00:00:00+02:00

What is Photon, features and how to download.

What is Photon ?

Photon is a simple Photo Gallery writen in Rails3.

Features

Multi file upload in one time
Upload files from remote URL (WIP)
Local file storage (no S3)
Thumbnails and others sizes processed in background (DelayedJob)
Support of JPGs, GIFs and PNGs
Partial EXIF support (hackable to add more)
GPS EXIF Tags support and Mini-Map (via OSM) display

Get the sources!

Go here: Photon on GitHub

Or:

$ git clone git://github.com/rhaamo/photon.git

Deployment

$ cd photon
$ rake db:migrate
$ rake db:seed # optional, you will need to create manually an "admin" account
               # see the db/seeds.rb file for that

Maemo5/N900 Got RNotify frontend of libnotify

2010-06-14T00:00:00+02:00

I’ve finally got RNotify on the N900, this is how to get it.

Use it at your own risks, it can eat your kitten or baby

Edit: I’ve fixed the gtk2 package, now working.

Intro

I want to use RNotify in a ruby app for my N900 but… There is no gtk2 nor glib nor RNotify in the maemo repositories…

I’ve rebuilt theses packages and direct-dependencies from the debian’s one and provide it. (also note this is my firsts debian’s packages rebuild)

All packages are build with ‘debian/optify’ to ‘auto’ and i’ve run ‘maemo-optify-deb debname.deb’ for all debs, hope it’s good.

Requirements

I do not remember exactly what to install, use it at your own risks etc. Contact me or comment for anything to add here :)

You need the maemo-extras-devel (search the web for this) and probably maemo-extras then (i assume you are probably a geek-hacker and have installed ‘rootsh’)

Install this:

# apt-get install ruby irb rdoc wget

Then got all debs:

# cd /home/user/MyDocs
# mkdir root
# cd root
(there is more space here)
# wget -A.deb -nH -m -l 1 http://pub.rhaamo.li/Debian/armel/

Install all packages with dpkg -i packagename.deb in order:

_armel.deb then _all.deb and in order:

libglib → libcairo → libpango → libatk → libgdk → libgtk2 → libnotify → rubygems (optional but usefull)

Usage

Not too hard, but you really need to do require ‘glib2’ before require ‘rnotify’ or you will get a beautifull fail.

require 'glib2'
require 'rnotify'

Notify.init("test")
notif = Notify::Notification.new('X', nil, nil, nil)
notif.timeout = 5000
notif.update("notif test", "message de test", nil)
notif.show

That’s all, the docs are here

Using acts_as_versioned with rails3

2010-06-12T00:00:00+02:00

Simple usage of acts as versioned plugin with rails3.

Intro

In this post i will use some values from one of my projects:

Model name: “paste”
Some columns inside the model etc.

Don’t lost to adapt for you…

Installation

Just put inside your Gemfile:

gem 'acts_as_versioned', :git => "git://github.com/xspond/acts_as_versioned.git", :branch => "rails3"

Config

In the model to acts_as_versioned, add:

acts_as_versioned :table_name => :paste_versions

Add a migration:

class AddVersionedPaste < ActiveRecord::Migration
  def self.up
    Paste.create_versioned_table
  end

  def self.down
    Paste.drop_versioned_table
  end
end

Then run rake db:migrate , i need to run it twice, it fails on Paste.create\versioned\table the first time (don’t know why).

Using

I will use it from rails console but it’s the same from inside the app.

At this point from the www i create a paste, then edit it.

Our ‘paste’:

yespaste dev> a=Paste.find_by_id(1)
=> #<Paste id: 1, pid: "lrns2Lmxfo", language: "plaintext", expiration: 1276347188, expiration_choosed: 3600, comment: "foo v2", paste: "foo v2", private: 0, no_bbcode: 0, user_id: 1, created_at: "2010-06-12 11:47:26", updated_at: "2010-06-12 11:53:08", version: 2>

As you can see, it’s the ‘v2’, i’ve put ‘v1’ before the edit, so it’s good.

But the ‘v1’ paste isn’t lost, see:

yespaste dev> a.versions
=> [#<Paste::Version id: 1, paste_id: 1, version: 1, pid: "N45mzQmCUG", language: "plaintext", expiration: 1276346846, expiration_choosed: 3600, comment: "foo v1", paste: "foo v1", private: 0, no_bbcode: 0, user_id: 1, created_at: "2010-06-12 11:47:26", updated_at: "2010-06-12 11:47:26">, #<Paste::Version id: 2, paste_id: 1, version: 2, pid: "lrns2Lmxfo", language: "plaintext", expiration: 1276347188, expiration_choosed: 3600, comment: "foo v2", paste: "foo v2", private: 0, no_bbcode: 0, user_id: 1, created_at: "2010-06-12 11:47:26", updated_at: "2010-06-12 11:53:08">]

No, not lost, just versioned, so you can also do:

yespaste dev> a.versions.find_by_version(1)
=> #<Paste::Version id: 1, paste_id: 1, version: 1, pid: "N45mzQmCUG", language: "plaintext", expiration: 1276346846, expiration_choosed: 3600, comment: "prout v1", paste: "proutprout v1", private: 0, no_bbcode: 0, user_id: 1, created_at: "2010-06-12 11:47:26", updated_at: "2010-06-12 11:47:26">

Migration du blog vers Jekyll

2010-06-05T00:00:00+02:00

Moi aussi j’ai vu Jekyll on Heroku et je me suis dit, “pourquoi pas”, mon CBlog datant des hommes préhistoriques et voulant jouer avec des rubys, je me suis lancé.

Comme souvent, je suis aller piquer un design chez Bender ;), maintenant reste à le remplir le nouveau blog.

Les commentaires arriveront bientôt (merci ange), et quelques anciens posts utiles aussi, les autres seront archivées et disponibles bientôt.

EDIT: Ils sont la :)
EDIT: J’ai refait de zéro le design :)

Custom HTML-Maker for RBBCode

2010-02-20T00:00:00+01:00

How to write a custom HTML-Maker for RBBCode

For YesPaste i use RBBCode for BBCode parsing and need to have a custom tag: code.

By default TagNode doesn’t have pure (read: non parsed) text, after reading the source code i got it:

class YesPasteHtmlMaker < RbbCode::HtmlMaker
  def html_from_code_tag(node)
    if node.value.nil?
      lang = 'plaintext'
    else
      lang = node.value
    end
    inner_html = node.children.inject('') do |output, child|
      output + make_html(child).gsub(&quot;&lt;br/&gt;&quot;, &quot;\n&quot;)
    end
    tokens = CodeRay.scan(inner_html, lang).div(:line_numbers =&gt; :table)
    return tokens
  end
end

def BBCode_render(text)
  html_maker = YesPasteHtmlMaker.new
  parser = RbbCode::Parser.new(:html_maker =&gt; html_maker)
  html = parser.parse(text)
  return html
end

Mimetype d'un fichier en C

2009-12-08T00:00:00+01:00

Obtenir le mimetype d’un fichier en C

La lib magic est parfaite pour ça. man 3 libmagic sous FreeBSD.

Petit exemple:

BSDmakefile:

PROG=           mimetype
LDFLAGS+=       -lmagic
CFLAGS+=        -Wall -g -ggdb3
NO_MAN=         yes

.include <bsd.prog.mk>

Code:

#include <magic.h>
#include <stdio.h>
#include <stdlib.h>
#include <err.h>

int
main(int argc, char *argv[])
{
    const char* program_name = argv[0];
    char *filename = argv[1];

    if (filename == NULL) {
        fprintf(stderr, "Usage: %s file\n", program_name);
        return(1);
    }

    struct magic_set *magic = magic_open(MAGIC_MIME_TYPE);
    if (magic == NULL)
        err(1, "magic_set");

    magic_load(magic, NULL);

    printf("%s: %s\n", filename, magic_file(magic, filename));

    magic_close(magic);

    return(0);
}

OpenBSD en domU HVM et carte réseau giga

2009-10-11T00:00:00+02:00

Jusqu’a maintenant la realteck foirais sur un blah0: watchdog timeout et il fallait utiliser la ne2k_pci qui présentait le désavantage d’être en 10BaseT (pas cool).

Maintenant, depuis Xen 3.3.1 ou peut-être un poil avant, on peut utiliser e1000.

Et ça fonctionne très bien !

# ifconfig em0
em0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
        lladdr 00:16:3e:15:7e:b4
         priority: 0
         groups: egress
         media: Ethernet autoselect (1000baseT full-duplex)
         status: active
         inet 42.42.42.42 netmask 0xffffffc0 broadcast 42.42.42.42
         inet6 ...%em0 prefixlen 64 scopeid 0x1
# dmesg|grep em0
em0 at pci0 dev 4 function 0 "Intel PRO/1000MT (82540EM)" rev 0x03: apic 1 int 5 (irq 5), address 00:16:3e:15:7e:b4