algesten

set = (v) -> (f) -> f(v)
get = (f) -> f (v) -> v

t = set(42)
get(t) // 42

algesten

install logstash with lusis/chef-logstash 0.11.2

My json config file

Notice that java, elasticsearch and rabbitmq are already set up in my system.

{
    "name": "logstash-server",

algesten

module.exports = class Bundlr

    FORMAT_TTNITF:       'ttnitf'
    FORMAT_PICTURE:      'bild'     # deliberate to match OFP
    FORMAT_TTNINJS:      'ttninjs'
    FORMAT_INDESIGN:     'indesign'
    FORMAT_CUTPASTE:     'cutpaste' # txt-format in media-moj
    FORMAT_TTNITFXML:    'ttnitfxml'

    # ofp spec is "ttninf", "ttnintf+bild", "ttinjs" etc

algesten

1970-01-01 behandlas som null

versioncreated = datum dokumentet skapas. (sdl fotodatum med tidsdel 00:00)
versionstored  = indexeringsdatum. om vi är master, eget annars ta källan (sdl typ)

date = datum dokumentet gäller (t.ex. framtid för planering
    eller framtid grafik, men fotodatum för sdl)

if tidsdel för date finns
    datetime = samma logik som date men med tidsdel. annars null.

algesten

# Goals
#
# 1. Mix promise and non-promise arguments 
# 2. Behaves like a normal function if all arguments are non-promise
# 3. Returns a promise if any non-promise argument
# 4. Use no promise library, rely on the fact that:
#      promise.then(-> a1).then((a2) -> a1 === a2)
#      promise.then(-> Promise(a1)).then((a2) -> a1 === a2)
#

algesten

body_html5 spec 1.2

This is a "loose" spec for the body_html5 part of [ttninjs][ttninjs].

Revision history

• 1.3 Clarify html-doc structure
• 1.2 alt-attribute in ``, div.byline before `figcaption` (to be html5-valid)

algesten

Implemented in https://github.com/algesten/fnuc/blob/master/src/fnuc.coffee#L202

Case for plift

This example shows how plift can help to treat functions dealing with promises as regular functions.

N.B. This code is a "handler" it is not intended to be side effect free.

• userToDevices is a function looking up devices from a user id in a database. It returns a promise for the result.

algesten

web: node web.js

algesten

alphabet = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789".split("")
base = alphabet.length

exports.encode = (i) ->
  return alphabet[0] if i is 0
  s = ""
  while i > 0
    s += alphabet[i % base]
    i = parseInt(i / base, 10)

algesten

    private final static Similarity NO_LENGTH_SIMILARITY = new DefaultSimilarity() {

        // we are interested in length norm for the street number field.
        // it's good if 10 is sorted before 10a and 10b.
        //
        // the "country" field is very important it has no document boost since that means
        // simply turning Query.setCountry() on will make a different score. This is because
        // a document boost will add to every field and country is matched in a MUST.
        @Override
        public float computeNorm(String field, FieldInvertState state) {