andersen.berlin

Rsync Real Progress

Thu, 09 Apr 2020 00:00:00 GMT

Looking a little bit through the man pages of rsync I found two flags to show the overall status of the currently running rsync process.

--info=progress2: display the overall status of the process
--no-inc-recursive: crawl all files and folders upfront

Unless --no-inc-recursive is used, rsync looks only a few thousand files ahead and recalculates the overall number of files causing a jump back to a lower completion percentage, which causes headaches ...

Hydra to Test-drive Mono Fonts

Sun, 09 Feb 2020 00:00:00 GMT

This hydra maps h and k for cycling through Mono fonts installed on your system:

(require 's)
(require 'dash)

(defun leoc/s-contains-mono (name)
  (and (s-contains-p "mono" name t)
       (not (s-contains-p "italic" name t))
       (not (s-contains-p "bold" name t))))

(setq leoc/frame-font-loop
      (-filter
       'leoc/s-contains-mono
       (x-list-fonts "*" nil (selected-frame))))
(setq leoc/current-frame-font-index 0)

(defun leoc/cycle-frame-font (direction)
  (let* ((current-font-index leoc/current-frame-font-index)
	 (next-font-index (mod (+ current-font-index direction) (length leoc/frame-font-loop)))
	 (next-font (nth next-font-index leoc/frame-font-loop)))
    (message "Setting new frame font (index %S): %S" leoc/current-frame-font-index next-font)
    (set-frame-font next-font)
    (if (null (assq 'font default-frame-alist))
	(add-to-list 'default-frame-alist
		     `(font . ,next-font))
      (setcdr (assq 'font default-frame-alist) next-font))
    (setq leoc/current-frame-font-index next-font-index)))

(defun leoc/cycle-next-frame-font ()
  "Cycle to next font in list."
  (interactive)
  (leoc/cycle-frame-font +1))

(defun leoc/cycle-prev-frame-font ()
  "Cycle to previous font in list."
  (interactive)
  (leoc/cycle-frame-font -1))

(defhydra hydra-frame-font nil
  "Cycle through mono fonts"
  ("k" leoc/cycle-next-frame-font "Next font")
  ("j" leoc/cycle-prev-frame-font "Previous font"))

Operations on Bash Variables and the Usual Suspects

Fri, 03 Jan 2020 00:00:00 GMT

Lately I have been catching myself looking up bash variable operations too often. This is something I want to keep in the back of my head.

The Bash Beginners Guide gives a good overview on Operations on variables.

TL;DR

Operation	Description
`${#STR}`	Length of the value of `$STR`
`${STR:-test}`	Default `$STR` to "test"
`[ -z "${INT:-}" ] && INT=80`	Default `$INT` to value `80`
`${STR:OFFSET}`	Keep everything after `OFFSET`
`${STR:OFFSET:LENGTH}`	Keep `LENGTH` characters after `OFFSET`
`${STR/PATTERN/STRING}`	Replace first occurence of `STRING`
`${STR//PATTERN/STRING}`	Replace all occurences of `STRING`
`${VAR#WORD}`	Remove first occurence of `WORD`
`${VAR##WORD}`	Remove all occurences of `WORD`
`${VAR#PATTERN}`	Remove shortest match of `PATTERN`
`${VAR##PATTERN}`	Remove longest match of `PATTERN`
`${VAR%WORD}`	Keep first occurence of `WORD`
`${VAR%%WORD}`	Keep all occurences of `WORD`
`${VAR%PATTERN}`	Keep shortest match of `PATTERN`
`${VAR%%PATTERN}`	Keep longest match of `PATTERN`

Usual Suspects

Operation	Description	Effectively
`${VAR##*/}`	Remove path until filename	Keep filename
`${VAR%/*}`	Keep the path without filename	Keep dirname

Ira Glass on Taste and Talent

Mon, 19 Aug 2019 00:00:00 GMT

A quote by Ira Glass left me thinking about my feeling how everything I do does not compare to what I ought to do in the beginning...

Nobody tells this to people who are beginners, I wish someone told me. All of us who do creative work, we get into it because we have good taste. But there is this gap. For the first couple years you make stuff, it’s just not that good. It’s trying to be good, it has potential, but it’s not. But your taste, the thing that got you into the game, is still killer. And your taste is why your work disappoints you. A lot of people never get past this phase, they quit. Most people I know who do interesting, creative work went through years of this. We know our work doesn’t have this special thing that we want it to have. We all go through this. And if you are just starting out or you are still in this phase, you gotta know its normal and the most important thing you can do is do a lot of work. Put yourself on a deadline so that every week you will finish one story. It is only by going through a volume of work that you will close that gap, and your work will be as good as your ambitions. And I took longer to figure out how to do this than anyone I’ve ever met. It’s gonna take awhile. It’s normal to take awhile. You’ve just gotta fight your way through.

I need some more time to think about it ...

Base58 for readability

Thu, 08 Nov 2018 00:00:00 GMT

When reading Bitcoin & Blockchain by Andreas M. Antonopoulos I stumbled upon Base58.

I cannot deny that I thought back with shame on all those times I tried to implement procedures that generated random human readable identifiers for different platforms.

Let's rejoice and use Base58!

Meteor and React Hooks

Tue, 06 Nov 2018 00:00:00 GMT

Since react hooks have been introduced I tried to rephrase my components with `react-hooks` and I came to the conclusion that they help me to write less code that is much more readable.

Currently I am working again with Meteor and to my joy Meteor 1.8 was just released, providing react templates. As of now it is suggested to use the `withTracker` HOC from meteor/react-meteor-data to subscribe to and fetch data.

Although that can be a quite clean solution I wanted to give `react-hooks` a try.

Via `meteor create --react <<projectname>>` this component is generated:

import React, { Component } from 'react';

import { withTracker } from 'meteor/react-meteor-data';

import Links from '../api/links';

class Info extends Component {
  render() {
    const links = this.props.links.map(
      link => this.makeLink(link)
    );

    return (
      <div>
        <h2>Learn Meteor!</h2>
        <ul>{ links }</ul>
      </div>
    );
  }

  makeLink(link) {
    return (
      <li key={link._id}>
        <a href={link.url} target="_blank">{link.title}</a>
      </li>
    );
  }
}

export default InfoContainer = withTracker(() => {
  return {
    links: Links.find().fetch(),
  };
})(Info);

My proposal for an API that uses `react-hooks` looks like that:

import { useMeteorSubscription, useMeteorData } from 'meteor/react-meteor-hooks';

import Links from '../api/links';

export const Info = function (props) {
  const loading = useMeteorSubscription('links');
  const links = useMeteorData(() => Links.find().fetch());

  if(loading) return (<div>Loading links ...</div>);

  return (
    <ul>
      {links.map((link) => (
        <li key={link._id}>
          <a href={link.url} target="_blank">{link.title}</a>
        </li>
      ))}
    </ul>
  );
}

I created a pull request and a sample project.

useMeteorSubscription

import { Meteor } from 'meteor/meteor';
import { Tracker } from 'meteor/tracker';
import { useEffect, useState } from 'react';

let useMeteorSubscription;

if (Meteor.isServer) {
  // When rendering on the server, we don't want to use the Tracker.
  // The subscription is always ready on the server.
  useMeteorSubscription = () => true;
} else {
  useMeteorSubscription = (publication, ...parameters) => {
    const [loading, setLoading] = useState(true);
    let handle, computation;

    const cleanUp = () => {
      handle && handle.stop();
      handle = null;
      computation && computation.stop();
      computation = null;
    }

    useEffect(() => {
      if(computation) cleanUp();

      Tracker.autorun((currentComputation) => {
        computation = currentComputation;

        handle = Meteor.subscribe(publication, ...parameters);
        setLoading(!handle.ready());
      });

      return cleanUp;
    }, [publication, ...parameters]);

    return loading;
  }
}

export default useMeteorSubscription;

useMeteorData

import { Meteor } from 'meteor/meteor';
import { Tracker } from 'meteor/tracker';
import { useEffect, useState } from 'react';

let useMeteorData;

if (Meteor.isServer) {
  // When rendering on the server, we don't want to use the Tracker.
  // We only do the first rendering on the server so we can get the data right away
  useMeteorData = getMeteorData => getMeteorData();
} else {
  useMeteorData = (getMeteorData, inputs = []) => {
    const [meteorData, setMeteorData] = useState(getMeteorData());
    let computation;

    const cleanUp = () => {
      computation.stop();
      computation = null;
    }

    useEffect(() => {
      if(computation) cleanUp();

      Tracker.autorun((currentComputation) => {
        computation = currentComputation;
        setMeteorData(getMeteorData());
      });

      return cleanUp;
    }, inputs);

    return meteorData;
  }
}
export default useMeteorData;

Manage Docker Images Disk Usage

Thu, 17 May 2018 00:00:00 GMT

A while ago I was finding that ncdu / shows that docker is using a lot of disk space. I showed how to remove volumes in Manage Docker Volume Disk Size. But volumes are not the only thing that can claim disk space.

A usual suspect are images that are not used anymore. You can get a list of images that are not tagged anymore via:

docker images -q --filter "dangling=true"

To delete them all, hit:

docker rmi $(docker images -q --filter "dangling=true")

Kernel Panic: No working init found

Fri, 11 May 2018 00:00:00 GMT

For the second time I performed a `pacman -Syyu` to update my system and yet again with the kernel update an unbootable system was produced.

TL;DR

Pacman seemed to have deleted the /usr/lib64 symlink so that mkinitcpio built an incomplete initramfs.

Check /boot/initramfs-linux.img is too small (~ 2.3 MB)
Ensure existence of symlink ln -s /usr/lib64 -> /usr/lib
Rebuild initramfs via mkinitcpio -p linux
Check /boot/initramfs-linux.img is about 7.4 MB
Reboot

Text and Links and Stuff

The error message upon restart read the following:

ACPI Error: Method parse/execution failed \_PR.CPU0._PDC, AE_AML_OPERAND_TYPE (20180105/psparse-550)
Failed to execute /init (error -2)
Kernel panic - not syncing: No working init found.  Try passing init= option to kernel.  See Linux Documentation/admin-guide/init.rst for guidance.

When I booted into a archlinux console from the installation USB stick and chroot~ed into the system I looked into the ~/boot directory and found the following:

$ ls -alh /boot
total 42M
drwxr-xr-x  6 root root 1.0K Jan  1  1970 .
drwxr-xr-x 20 root root 4.0K Apr 12 12:42 ..
drwxr-xr-x  4 root root 1.0K May 11 09:59 EFI
drwxr-xr-x  6 root root 1.0K Apr 12 12:52 grub
-rwxr-xr-x  1 root root  28M May 11 10:19 initramfs-linux-fallback.img
-rwxr-xr-x  1 root root 2.3M May 11 10:19 initramfs-linux.img
-rwxr-xr-x  1 root root 1.6M May  7 23:11 intel-ucode.img
drwxr-xr-x  3 root root 1.0K Apr 26  2017 loader
drwxr-xr-x  2 root root 1.0K Apr 12 12:27 syslinux
-rwxr-xr-x  1 root root 5.0M May  9 14:21 vmlinuz-linux

The peculiar thing here is, that the size of the initramfs-linux.img is too small. Normally it is something like 7.4 MB in size!

Some searching revealed that pacman might remove the /usr/lib64 symlink: Archlinux forums.

So I created the link and rebuilt the initramfs:

ln -s /usr/lib /usr/lib64
mkinitcpio -p linux

After that everything seemed to be fine again:

$ ls -alh /boot
total 42M
drwxr-xr-x  6 root root 1.0K Jan  1  1970 .
drwxr-xr-x 20 root root 4.0K Apr 12 12:42 ..
drwxr-xr-x  4 root root 1.0K May 11 09:59 EFI
drwxr-xr-x  6 root root 1.0K Apr 12 12:52 grub
-rwxr-xr-x  1 root root  28M May 11 10:19 initramfs-linux-fallback.img
-rwxr-xr-x  1 root root 7.4M May 11 10:19 initramfs-linux.img
-rwxr-xr-x  1 root root 1.6M May  7 23:11 intel-ucode.img
drwxr-xr-x  3 root root 1.0K Apr 26  2017 loader
drwxr-xr-x  2 root root 1.0K Apr 12 12:27 syslinux
-rwxr-xr-x  1 root root 5.0M May  9 14:21 vmlinuz-linux

Booting works!

Fix thinkpad graphicscard hang after archlinux update

Sun, 08 Oct 2017 00:00:00 GMT

The kernel developer decided to change the default for intel_iommu.

Some machines have problems after the update. Either not booting at all or throwing messages like these around:

Oct 08 13:26:28 eigen kernel: DMAR: DRHD: handling fault status reg 3 Oct 08 13:26:28 eigen kernel: DMAR: [DMA Write] Request device [00:02.0] fault addr 3a87000 [fault reason 23] Unknown Oct 08 13:26:28 eigen kernel: DMAR: DRHD: handling fault status reg 3 Oct 08 13:26:28 eigen kernel: DMAR: [DMA Write] Request device [00:02.0] fault addr 3a84000 [fault reason 23] Unknown Oct 08 13:26:28 eigen kernel: DMAR: DRHD: handling fault status reg 3 Oct 08 13:26:28 eigen kernel: DMAR: [DMA Read] Request device [00:02.0] fault addr 7ea000 [fault reason 23] Unknown Oct 08 13:26:31 eigen kernel: [drm:fw_domains_get [i915]] *ERROR* render: timed out waiting for forcewake ack to clear.

Oct 08 13:49:23 eigen kernel: DMAR: DRHD: handling fault status reg 3 Oct 08 13:49:23 eigen kernel: DMAR: [DMA Write] Request device [00:02.0] fault addr 3568000 [fault reason 23] Unknown Oct 08 13:49:23 eigen kernel: DMAR: DRHD: handling fault status reg 3 Oct 08 13:49:23 eigen kernel: DMAR: [DMA Write] Request device [00:02.0] fault addr 35aa000 [fault reason 23] Unknown Oct 08 13:49:23 eigen kernel: DMAR: DRHD: handling fault status reg 2 Oct 08 13:49:23 eigen kernel: DMAR: [DMA Write] Request device [00:02.0] fault addr 35ec000 [fault reason 23] Unknown Oct 08 13:49:32 eigen kernel: [drm] GPU HANG: ecode 8:0:0x85dffffb, in chromium [1112], reason: Hang on rcs0, action: reset Oct 08 13:49:32 eigen kernel: [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. Oct 08 13:49:32 eigen kernel: [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel Oct 08 13:49:32 eigen kernel: [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. Oct 08 13:49:32 eigen kernel: [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it. Oct 08 13:49:32 eigen kernel: [drm] GPU crash dump saved to /sys/class/drm/card0/error Oct 08 13:49:32 eigen kernel: drm/i915: Resetting chip after gpu hang Oct 08 13:49:32 eigen systemd[1]: Starting autorandr execution hook... Oct 08 13:49:40 eigen kernel: drm/i915: Resetting chip after gpu hang

I found a hint on how to fix this issue in the Archlinux forums.

Try booting linux with intel_iommu=off or intel_iommu=igfx_off (in the grub menu hit e to edit the boot command)

To make it permanent add the respective flag to /etc/default/grub and run:

sudo grub-mkconfig -o /boot/grub/grub.cfg

This should fix the issue.

Disabling pointer events

Wed, 27 Sep 2017 00:00:00 GMT

I just found out that you can make a page element "unclickable". Meaning that any click on this element will be registered by the element below this respective element.

This is especially useful if you want to put an overlay on top of something (absolutely positioned) and want to be the underlying layer to be clickable.

.unclickable {
  pointer-events: none;
}

How to refile after org capture

Sun, 24 Sep 2017 00:00:00 GMT

I am using org-protocol and capture things from within my browser and from my window manager (awesomewm). Some templates are generic TODO headings and need to be refiled to the correct location.

This simple script refiles certain templates (e.g. "a" and "c") when finalizing the capturing process.

 (defun leoc/refile-maybe ()
   (let ((key (org-capture-get :key)))
     (when (member key '("a" "c"))
(org-refile))))

 (add-hook 'org-capture-before-finalize-hook 'leoc/refile-maybe)

Install Ruby < 2.3 on Archlinux and OpenSSL 1.1.x

Sun, 24 Sep 2017 00:00:00 GMT

Archlinux updated OpenSSL to 1.1.x. Earlier versions of ruby (< 2.3) depend on OpenSSL 1.0. Thus we have to install openssl-1.0 via pacman and set the environment variables before installing one of the older rubies.

export PKG_CONFIG_PATH=/usr/lib/openssl-1.0/pkgconfig
export CFLAGS="$CFLAGS -I/usr/include/openssl-1.0"
export LDFLAGS="$LDFLAGS -L/usr/lib/openssl-1.0 -lssl"
rvm reinstall ruby-2.1

Store incremental backups with rsync and hardlinks

Tue, 15 Aug 2017 00:00:00 GMT

This script stores backups on a remote server in your network via rsync. To spare disk space and speed up the backup process the rsync command uses hardlinks to only transfer incremental backups.

Credits to morhekil, who wrote a shell script to backup to a local folder: https://gist.github.com/morhekil/8382294.

#!/usr/bin/env ruby

$host = 'ssh-hostname'
$rsync_bin = '/usr/bin/rsync'
$local_backup_path = '/home/arthur'
$remote_backup_path = '/backup/thinkpad'
$relative_link_target = '..'
$verbose = true
$dry_run = false
$excludes = [
  'arthur/projects/**/tmp',
  'arthur/projects/**/log',
  'arthur/.cache'
]

def ssh(command)
  puts "--- ssh dmt \"#{command}\"" if $verbose
  `ssh #{$host} "#{command}"`.strip
end

ssh("mkdir -p \"#{$remote_backup_path}\"")
timestamp = ssh('date +"%Y.%m.%d.%H.%M.%S"')
remote_target_path = "#{$remote_backup_path}/#{timestamp}"

last_backup = ssh("ls -1 #{$remote_backup_path} | tail -n1")
link_target_path = File.join($relative_link_target, last_backup)

rsync_options = ''
rsync_options += ' -v' if $verbose
rsync_options += ' -n' if $dry_run
rsync_options += " --link-dest=#{link_target_path}" unless link_target_path == ''
$excludes.each do |exclude|
  rsync_options += " --exclude '#{exclude}'"
end

puts "--- #{$rsync_bin} -a --progress --delete#{rsync_options} '#{$local_backup_path}' #{$host}:#{remote_target_path}"
system("#{$rsync_bin} -a --progress --delete #{rsync_options} '#{$local_backup_path}' #{$host}:#{remote_target_path}")

I will work out a way to delete old backups later.

Manage Docker Volume Disk Size

Mon, 17 Jul 2017 00:00:00 GMT

Today I was confused about the missing disk space on my development machine. I ran ncdu / only to find out that a lot of the disk is used by docker volumes.

So I began investigating which volumes are actually taking this much space:

docker system df -v

Which showed a lot of local volumes that were taking a lot of GB. These were the result of containers which mounted volumes to the wrong endpoint within the container, so that the volumes grew heavily.

I deleted the largest container first (40 GB):

docker volume rm <VOLUME_NAME>

But then decided to find dangling volumes:

docker volume ls -q -f dangling=true

These can be easily removed via:

docker volume rm `docker volume ls -q -f dangling=true`

This should be executed from time to time, to remove volumes which surreptitiously took a lot of your disk space.

Let It Be Ledgit

Fri, 19 Oct 2012 00:00:00 GMT

First I called it LedgeIt. Then I thought ledge isn't even a real english verb, so why not make it some more artsy-fartsy. Call it ledgit, seems legit to me. And now I found out, that this term is in the urban dictionary. I think it's the right name for that script.

A friend of mine introduced me to the magnificent commandline accounting tool ledger. I like the idea a lot, but I miss the comfort of having all banking account transactions synchronized automatically. So I wrote ledgit.

Ledgit provides a simple procedure. First it reads your ledger files and creates an index, so it can later on approximate the right ledger accounts and transaction names for the synchronized entries. Then it invokes the account handlers, which logs into the online banking website, downloads the last transactions as CSV and parses the data into a hash structure, that the generic ledgit method may use to update your ledger files.

I would love to see some people writing handlers for different online banking websites. It's basically surfing the website via the mechanize gem.

To get a clue of the API that handler class has to implement, have a look at leoc/ledgit/lib/handler/dkb.rb.

Markup.el - The First Step

Fri, 07 Sep 2012 00:00:00 GMT

I had some time yesterday and decided to port the cl-markup package to emacs lisp.

What I came up with I called `markup.el`.

You can check out the code on GitHub. Now you can easily generate HTML or XML from emacs lisp.

The easiest thing you can do is generate simple tags.

(markup (:p "This is a simple example")) ;; => "<p>This is a simple example</p>"

But you can easily create a full html5 document aswell:

(markup-html5
  (:div :style "border: 1px solid #000"
     "Some content of the div!"))
;; => "<!DOCTYPE html><html><div style=\"border: 1px solid #000\">Some content of the div!</div></html>"

You can even mix in some other expressions or variables:

(let ((maximum 5))
  (markup-html5
   (:body
    (:ul
     (loop for x from 1 to maximum
           collect
           (markup (:li  x " / " maximum)))))))
;; => "<!DOCTYPE html><html><body><ul><li>1 / 5</li><li>2 / 5</li><li>3 / 5</li><li>4 / 5</li><li>5 / 5</li></ul></body></html>"

This is my first step towards an emacs lisp static website generator. The next step will be to extract the meta information from org files and integrate their content into layouts that have been described in markup.el compatible files.

Org-mode, GTD and the Pomodoro technique

Wed, 22 Aug 2012 00:00:00 GMT

I have been struggling with keeping myself organized for some time now. I think just as much as anyone else out there. I have tried different todo apps and different techniques.

Before I switched to a ThinkPad as my main working horse I had a MacBook. Those Mac devs definitly know how to make apps nice and shiny, but no app really fitted my needs. They may be polished and integrate in OS X' overall looks, but they implement a likely limited usability of Macintoshs aswell.

I have been using Emacs as my main development environment for some time on the mac so I gave org-mode a try. And just as much as I learned to love Emacs I learned to love org-mode.

As technique to force myself to focus on tasks that my brain refuses to keep focus on I became accustomed to the pomodoro technique. Some people may condemn it as a stupid timer flow and yes it's nothing else. But at least anyone that hears the term knows what I'm talking about.

Let me introduce you to my org-mode configuration.

Most of the ideas come from Bernt Hansen. Check out his ingenious documentation.

For my configuration I took some of his ideas and tried to adapt them to my needs. The results can be found in the org-helpers.el package. I cleaned up some code and made some functions more generic. Because I decided to go with the pomodoro technique I did not take over the clocking functions. For the pomodoro functions I did a rewrite of the org-pomodoro package which was written by Marcin Kozey. Have a look at my Github project (leoc/org-pomodoro).

In the following sections I will explain my .emacs.d/leoc/leoc-org.el.

The file extension associations, bindings and requirements are trivial.

(add-to-list 'auto-mode-alist '("\\.org$" . org-mode))
(add-to-list 'auto-mode-alist '("\\.org_archive$" . org-mode))

(global-set-key (kbd "C-c l") 'org-store-link)
(global-set-key (kbd "C-c a") 'org-agenda)
(global-set-key (kbd "C-c b") 'org-iswitchb)

(require 'org-habit)
(require 'org-helpers)
(require 'org-pomodoro)

Because I want to use pomodoro technique I clock into full pomodoros. That makes clocking out obsolete. If I want to start a new pomodoro I have to kill the running one first. This is done by the org-pomodoro function.

(global-set-key (kbd "C-c C-x C-i") 'org-pomodoro)
(global-set-key (kbd "C-c C-x C-o") 'org-pomodoro)

Tasks may be TODO, actionable tasks are declared as NEXT and finished tasks are DONE. Then there are tasks that are inactive. Eithere they are WAITING for an event to happen, they are on HOLD for an unspecified amount of time (which means they have been worked on), they are declared for SOMEDAY in the future or they are CANCELLED.

(setq org-todo-keywords
      '((sequence "TODO(t)" "NEXT(n)" "|" "DONE(d!/!)")
        (sequence "WAITING(w@/!)" "HOLD(h@/!)" "SOMEDAY(o)" "|" "CANCELLED(c@/!)")))

The agenda tasks are split into different files. I added their archive files aswell, because I like to have a look in the logs aswell.

(setq org-agenda-files (list "~/.org/tasks.org"
                             "~/.org/tasks.org_archive"
                             "~/.org/projects.org"
                             "~/.org/projects.org_archive"
                             "~/.org/job.org"
                             "~/.org/job.org_archive"
                             "~/.org/calendar.org"))

I have a file called inbox.org where I put anything that comes to my mind. When I

(setq org-capture-templates
      '(("r" "Todo" entry (file+headline "~/.org/inbox.org" "Inbox")
         "* TODO %?")
        ("j" "Journal" entry (file+datetree "~/.org/journal.org")
         (file "~/.org/templates/review"))))

For those two capture templates I assigned specific key bindings.

(define-key global-map "\C-cr"
  (lambda () (interactive) (org-capture nil "r")))
(define-key global-map "\C-cj"
  (lambda () (interactive) (org-capture nil "j")))

To configure the habits I have found the following to be quite useful.

;; display teh tags farther right
(setq org-agenda-tags-column -102)
;; display the org-habit graph right of the tags
(setq org-habit-graph-column 102)
(setq org-habit-following-days 7)
(setq org-habit-preceding-days 21)

In the agenda mode I like to have the possibility to restrict to projects or subtrees at point. When widened the agenda only shows the top-level projects. When I restrict to a project alls subprojects and subtasks are shown. Furthermore I want to be able to quickly start pomodoros for tasks at point.

(defun custom-org-agenda-mode-defaults ()
  (org-defkey org-agenda-mode-map "W" 'oh/agenda-remove-restriction)
  (org-defkey org-agenda-mode-map "N" 'oh/agenda-restrict-to-subtree)
  (org-defkey org-agenda-mode-map "P" 'oh/agenda-restrict-to-project)
  (org-defkey org-agenda-mode-map "q" 'bury-buffer)
  (org-defkey org-agenda-mode-map "I" 'org-pomodoro)
  (org-defkey org-agenda-mode-map "O" 'org-pomodoro)
  (org-defkey org-agenda-mode-map (kbd "C-c C-x C-i") 'org-pomodoro)
  (org-defkey org-agenda-mode-map (kbd "C-c C-x C-o") 'org-pomodoro))

(add-hook 'org-agenda-mode-hook 'custom-org-agenda-mode-defaults 'append)

Now we come to the heart of my configuration. The custom agenda commands. I have the same overview as Bernt Hansen does. It works pretty well. Tasks declared SOMEDAY are not shown in any agenda. I use the main agenda a most of the time, which is a combination of all the different views. First there is the daily agenda, then I have the tasks to refile. Then the List of stuck projects and the list of actionable NEXT tasks.

Then there is a list of available tasks, which shows only single tasks unless the agenda is restricted to a subtree. If it is, then the available tasks for this subtree are shown.

As described earlier the list of projects lists only the top-level projects, unless you have restricted the view to a subtree. Then the subprojects of that particular subtree are shown.

Last, there is a list of "Waiting or Postponed Tasks", where tasks declared WAITING or HOLD are shown.

(setq org-agenda-custom-commands
      '(("a" "Agenda"
       ((agenda "" nil)
          (alltodo ""
                   ((org-agenda-overriding-header "Tasks to Refile")
                    (org-agenda-files '("~/.org/inbox.org"))
                    (org-agenda-skip-function
                     '(oh/agenda-skip :headline-if-restricted-and '(todo)))))
          (tags-todo "-CANCELLED/!-HOLD-WAITING"
                     ((org-agenda-overriding-header "Stuck Projects")
                      (org-agenda-skip-function
                       '(oh/agenda-skip :subtree-if '(inactive non-project non-stuck-project habit scheduled deadline)))))
          (tags-todo "-WAITING-CANCELLED/!NEXT"
                     ((org-agenda-overriding-header "Next Tasks")
                      (org-agenda-skip-function
                       '(oh/agenda-skip :subtree-if '(inactive project habit scheduled deadline)))
                      (org-tags-match-list-sublevels t)
                      (org-agenda-sorting-strategy '(todo-state-down effort-up category-keep))))
          (tags-todo "-CANCELLED/!-NEXT-HOLD-WAITING"
                     ((org-agenda-overriding-header "Available Tasks")
                      (org-agenda-skip-function
                       '(oh/agenda-skip :headline-if '(project)
                                        :subtree-if '(inactive habit scheduled deadline)
                                        :subtree-if-unrestricted-and '(subtask)
                                        :subtree-if-restricted-and '(single-task)))
                      (org-agenda-sorting-strategy '(category-keep))))
          (tags-todo "-CANCELLED/!"
                     ((org-agenda-overriding-header "Currently Active Projects")
                      (org-agenda-skip-function
                       '(oh/agenda-skip :subtree-if '(non-project stuck-project inactive habit)
                                        :headline-if-unrestricted-and '(subproject)
                                        :headline-if-restricted-and '(top-project)))
                      (org-agenda-sorting-strategy '(category-keep))))
          (tags-todo "-CANCELLED/!WAITING|HOLD"
                     ((org-agenda-overriding-header "Waiting and Postponed Tasks")
                      (org-agenda-skip-function
                       '(oh/agenda-skip :subtree-if '(project habit))))))
         nil)
        ("r" "Tasks to Refile" alltodo ""
         ((org-agenda-overriding-header "Tasks to Refile")
          (org-agenda-files '("~/.org/inbox.org"))))
        ("#" "Stuck Projects" tags-todo "-CANCELLED/!-HOLD-WAITING"
         ((org-agenda-overriding-header "Stuck Projects")
          (org-agenda-skip-function
           '(oh/agenda-skip :subtree-if '(inactive non-project non-stuck-project
                                          habit scheduled deadline)))))
        ("n" "Next Tasks" tags-todo "-WAITING-CANCELLED/!NEXT"
         ((org-agenda-overriding-header "Next Tasks")
          (org-agenda-skip-function
           '(oh/agenda-skip :subtree-if '(inactive project habit scheduled deadline)))
          (org-tags-match-list-sublevels t)
          (org-agenda-sorting-strategy '(todo-state-down effort-up category-keep))))
        ("R" "Tasks" tags-todo "-CANCELLED/!-NEXT-HOLD-WAITING"
         ((org-agenda-overriding-header "Available Tasks")
          (org-agenda-skip-function
           '(oh/agenda-skip :headline-if '(project)
                            :subtree-if '(inactive habit scheduled deadline)
                            :subtree-if-unrestricted-and '(subtask)
                            :subtree-if-restricted-and '(single-task)))
          (org-agenda-sorting-strategy '(category-keep))))
        ("p" "Projects" tags-todo "-CANCELLED/!"
         ((org-agenda-overriding-header "Currently Active Projects")
          (org-agenda-skip-function
           '(oh/agenda-skip :subtree-if '(non-project inactive habit)))
              (org-agenda-sorting-strategy '(category-keep))
              (org-tags-match-list-sublevels 'indented)))
        ("w" "Waiting Tasks" tags-todo "-CANCELLED/!WAITING|HOLD"
         ((org-agenda-overriding-header "Waiting and Postponed Tasks")
          (org-agenda-skip-function '(oh/agenda-skip :subtree-if '(project habit)))))))

The main helper method here is the oh/agenda-skip functions, which allows to generically skip headlines or subtrees, whether restricted or unrestricted.

Gem: EM::FS

Fri, 06 Jul 2012 00:00:00 GMT

Lately I needed to access the filesystem in a non-blocking way with eventmachine. So I wrote a gem to easily do this.

Check it out its Github Page.

Gem: KindleFS

Mon, 21 May 2012 00:00:00 GMT

After I had a functioning version of Rindle finished, I began working on a FUSE Kindle Management FS. You can mount the Kindles Collection data into you filesystem and move files from collection to collection, rename them or delete them.

Check it out its Github Page.

Gem: Rindle ... finally

Thu, 17 May 2012 00:00:00 GMT

Last year I began developing a simple gem for managing the Kindles collection data. Somehow I lost track of it over the past months. As I got some new eBooks for my Kindle I reminded myself of that small project and made it usable again.

Check it out its Github Page.

Here is a basic example:

require 'rindle'

Rindle.load '/path/to/kindle/mount/'

# Then just access collections and documents via their object representation
coll = Rindle::Collection.find_by_name 'my collection'
doc = Rindle::Document.find_by_name 'A book.pdf'
coll.add doc
coll.rename! 'a collection'

Rindle.save!

Gem: EM::SystemCommand

Fri, 04 May 2012 00:00:00 GMT

I just published my first ruby gem. It's called EM::SystemCommand and serves as an abstration to popen3 calls. You can easily execute system commands and handle the data that's been written to their stdout or stderr.

Check it out its Github Page.

Here is a basic example:

EM.run do
  ps = EM::SystemCommand.new 'my_command' do |on|
    on.success do |ps|
      puts 'Success!'
    end

    on.failure do |ps|
      puts 'Failure!'
    end

    on.stdout.line do |line|
      puts "Line received: #{line}"
    end
  end

  ps.stdin << "some data"
end

SonicWALL VPN with Ubuntu

Wed, 28 Sep 2011 00:00:00 GMT

The company I work for has a great new Virtual Private Network. Before it was powered by OpenVPN and a linux box. But because noone wanted to hire administrators who know how to work with linux systems, the boss got mad and bought a all-in-one enterprise firewall and VPN solution. That was the SonicWALL.

Unfortunately it is hard or simply impossible to connect a linux system to those SonicWALLs. if someone has a clue how to do it without the help of a Windows Virtual Machine, please enlighten me!

Anyway, to solve my lost of connection I used a virtual machine with Windows XP. And routed all traffic to the VPN client running on that machine.

To use that method go through the following steps:

Configure a Windows XP VBox with two Network Interfaces (one Bridged, one Host-Only)
Install the Sonicwall VPN Client software
Click on the newly created "SonicWALL VPN Connection" in your System configurations Network Connections Overview and choose Properties. On The 'Advanced' Tab enable 'Allow network users to connect through this computer's Internet connection' and use the Host-Only Adapter for 'Home network connection'
The TCP/IP settings for the host-only adapter should change to something like 192.168.0.1/255.255.55.0. If not, configure this manually.
In your Ubuntu system, configure the Host-Only vboxnet to use a static IP 192.168.0.2/255.255.255.0. The 192.168.0.1 should be "pingable" from within your Ubuntu host.
Now set up the routes: sudo route add -net 10.0.0.0 gw 192.168.0.1 netmask 255.0.0.0 dev vboxnet7 (where vboxnet7 is my Host-Only Adapter)
Reconnect the SonicWALL VPN Client

Now you should be able to access the VPNs machines.

To make the routes permanent, you may create a file static-route in /etc/network/if-up.d which may look like this:

#!/bin/sh
/sbin/route add -net 10.0.0.0 gw 192.168.0.1 netmask 255.0.0.0 dev vboxnet7

Rails 3 and SqlServer 2008

Tue, 08 Mar 2011 00:00:00 GMT

For the next project at work I have to interface Microsoft SqlServer 2008 from my rails applications. The setup is quite simple. First we need unixODBC. Normally Mac OS X ships with iODBC, but since I use linux servers, I wanted to use the layer that's going to be used on my production systems aswell.

Lets start!

brew install unixodbc

Next, we need FreeTDS, which is a free implementation of the Tabular Data Stream protocol.

brew edit freetds

Add --with-unixodbc=/usr/local to the ./configure line. Then:

brew install freetds

Then I created ~/.freetds.conf and added:

[development_db]
host = <Database IP>
port = 1433
tds version = 8.0
client charset = UTF-8

Edited /usr/local/etc/odbcinst.ini and added:

[FreeTDS]
Description = TD Driver (MSSQL)
Driver = /usr/local/lib/libtdsodbc.so
Setup = /usr/local/lib/libtdsodbc.so
FileUsage = 1

And ~/.odbc.ini

[development_db]
Driver= /usr/local/lib/libtdsodbc.so
Servername=development_db
Server=<Database IP>
Port=1433
Database=<database>
TDS_Version=8.0

To use the odbc connection, I just needed to install two ruby gems. Just add the following to your Gemfile.

gem "activerecord-sqlserver-adapter"
gem "ruby-odbc"

I ran bundle install and configured my config/database.yml

adapter: sqlserver
mode: odbc
username: <login>
password: <password>
database: <database>
dsn:  development_db

RSpec and Autotest for Non-Rails Development

Tue, 01 Mar 2011 00:00:00 GMT

I recently started a little new project. But since I were very spoiled by all the glitter of rails development, I struggled some time setting up rspec and autotest for my new project. And ... Mea Culpa! It really is that simple.

Simply create a spec/spec_helper.rb file, to require rspec and the libs that you want to test. Maybe it's gonna be a gem, so you have to adjust the load path.

$LOAD_PATH.unshift(File.join(File.dirname(__FILE__), '..', 'lib'))
$LOAD_PATH.unshift(File.dirname(__FILE__))

require 'rubygems'
require 'rspec'

require 'my_lib'

Rspec.configure do |config|
  # give me something to do!
end

And configure autotest to discover RSpec2 in autotest/discover.rb

Autotest.add_discovery { "rspec2" }

That's it. Just run autotest in the root of you project.

I added some more like rspec parameters in .rspec and the autotest/fs-event inside of the .autotest file.

Git - Lost Commits From No Branch

Fri, 03 Dec 2010 00:00:00 GMT

I did a mistake when working on submodule clone instead of a manual cloned repository.

That means as I checked out the master branch, my commits from the previous 'no-branch' state where reverted. To save them just do it like that:

cd cloned-repo
git reflog
git checkout master
git cherry-pick sha_of_the_missing_commit

Note to myself: Always check out the master branch before modifying submodule clones.

Emacs ... Finally?

Fri, 23 Jul 2010 00:00:00 GMT

Over the past few years I have been changing my development environments regularly. I began using a simple editor like nano or the windows editor (shame on me, but it's the only good software you get with windows)

Lately I have settled down using TextMate on the Mac for all of my ruby development. It's a fairly neat editor but I'm running linux on my workstation so I always have to run against my habits when changing the environments. That's kind of annoying.

A friend suggested emacs and what I can tell for now is that it's a pretty neat thing!

Switching To GitAuth

Thu, 15 Jul 2010 00:00:00 GMT

After some struggling with Gitosis I decided to go with GitAuth to administrate my private git repositories. Not only it is superior because it is written in Ruby. No, it provides a simple commandline interface to handle all the users, groups, repositories and permissions. I also appreciate the data stored in yml files for easy access from whatever software you may write in the future.

It was written by Brown Beagle Software. There is an easy guide at GitHub.

I went the `git clone` way and had the setup running in five minutes. Since I use Debian with Ruby Enterprise Edition I had to fix some issues that occured to me when ever I tried to access the created repositories. I don't know whether it works for your out of the box or not.

My Tiny DNS Helper

Thu, 01 Jul 2010 00:00:00 GMT

Until now I used that nice feature offered by hosteurope to use an auto-dns service, but somehow I'm missing the freedom right now so I had to setup a DNS server.

I decided to use TinyDNS and djbDNS. A neat alternative to BIND & Co.

sudo apt-get install djbdns

Afterwards we need to create some users and to configure the tinydns with our ip address.

adduser --no-create-home --disabled-password --force-badname Gtinydns
adduser --no-create-home --disabled-password --force-badname Gaxfrdns
adduser --no-create-home --disabled-password --force-badname Gdnslog

sudo axfrdns-conf Gaxfrdns Gdnslog  /etc/axfrdns /etc/tinydns 83.169.43.97
sudo tinydns-conf Gtinydns Gdnslog /etc/tinydns 83.169.43.97

Now we are going to put tinydns into our service directory to start it.

ln -s /etc/tinydns /etc/service/tinydns
ln -s /etc/axfrdns /etc/service/axfrdns
sleep 5
svstat /etc/service/tinydns
svstat /etc/service/axfrdns

And there you are. To customize the configuration you can switch to the /etc/tinydns/root directory and use the scripts inside. But remember to run make afterwards so that tinydns-data gets the changes.

If your secondary nameserver does not pull changes automatically you will need a [little helper script](/downloads/dnsnotify) that notifies the nameserver to initiate a zone transfer. I added that script to my /etc/tinydns/root/Makefile to run it automatically when I compile the new tinydns data.

Ever Wondered What's The Steam?

Mon, 26 Apr 2010 00:00:00 GMT

I don't know. But I hate to work with Java for all that university stuff.

Finally A Jekyll Blog

Tue, 20 Apr 2010 00:00:00 GMT

Alright. There I am. Another cool guy with a blog powered by Jekyll.

Now I just need some content. A typical problem.

Anyway. The setup is fairly simple. I am using two directories for the blog.

One /srv/blog/source, in which I export the current repository data. One /srv/blog/public, where Jekyll puts the generated page files.

Then I just had to install the following script as post-update hook for my blog repository which executes this rule on every push.

#!/bin/sh
rm -rf /srv/blog/source/*
git --bare --git-dir=/home/git/repositories/blog.git archive master | tar xC /srv/blog/source
jekyll /srv/blog/source /srv/blog/public

Do not forget to make the hook executable:

$ chown +x /home/git/repositories/blog.git/hooks/post-update

Afterwards I configured my web server to serve my files from /srv/blog/public

Now with every push to my blogs repository it will automatically get deployed. What a wonderful world.