Flow

Question 1

Hello everybody

Answer 1

not haskell,

side projects

Question 2

whole application

Answer 2

lots of functions, types, and code

Question 3

main goals

Answer 3

why bother
explicit
6 months
-> solution

Question 4

Gitlab

Answer 4

repo and code reviews

github alternative

Question 5

Merge requests

Answer 5

corelation
how long vs size
adventurous

Question 6

architecture

Answer 6

-

Question 7

Haskell & Elm

Answer 7

prerequisite

Question 8

sum3ints

Answer 8

-> factorial

Question 9

HTTP Client

Answer 9

understand some Haskell

Question 10

Haskell Servant

Answer 10

takes over HTTP
focus on APIs
exclusively in types

Question 11

Merge Request Fetcher

Answer 11

detailed architecture

Question 12

Gitlab API

Answer 12

-

Question 13

JSON type

Answer 13

-

Question 14

data type

Answer 14

-

Question 15

evil smiley face emoticon

Answer 15

-

Question 16

API type

Answer 16

one type

any change makes it a diff type

Question 17

apiQuery

Answer 17

• > first parameter
• > second parameter
• > third parameter
• > last input
• > return type

Question 18

return type

Answer 18

most comlicated type
Merge Requests List
Paged == HTTP Response header with URL

Question 19

IO

Answer 19

behind the curtains
IO guy grabs our leg
calculated value stays in IO

Question 20

IO Side Note

Answer 20

-> examples

your code?

Question 21

EitherT

Answer 21

-

Question 22

client

Answer 22

just type so far

Question 23

nicer signature

Answer 23

4 params

• > pagedMergeRequests
• > AppConfig

Question 24

pagedMergeRequests implementation

Answer 24

another small function

Question 25

allMergeRequests

Answer 25

remember EitherT return type

Question 26

allMergeRequests implementation

Answer 26

more complicated
fetchPage
-> the whole do
left arrow unpacks
nextPage

Question 27

the only side effect

Answer 27

get bored

Question 28

Merge Request Fetcher module summary

Answer 28

-

Question 29

allComments

Answer 29

merge request id

Question 30

MergeRequestStats

Answer 30

essential

only two values

Question 31

calculateStats

Answer 31

Maybe context
-> doesn’t have any IOs inside
-> nice thing
pure

Question 32

fetchMRsAndSaveStats

Answer 32

connecting the dots
do notation
-> appConfig
env variables
-> errorToString
any fail stops computation

Question 33

serve the stats

Answer 33

surprised that almost the same

Question 34

main

Answer 34

-

Question 35

we did it!

Answer 35

-

Question 36

automatically generate

Answer 36

one type to rule them all

Question 37

frontend

Answer 37

statically typed
compiled to JS
documentation, compiler errors, fast, streams, types and vdom
table with stats refreshed each 5 seconds

Question 38

Elm architecture

Answer 38

-> update
given the same model and action returns the same tuple
Action is defined by us
-> Elm Runtime
Actions can be spawned based on Effects like HTTP tasks
-> view

Question 39

Elm architecture types

Answer 39

3 puzzles

Question 40

model

Answer 40

lastRefreshTime surprise you

can’t get time

Question 41

Actions

Answer 41

-

Question 42

update

Answer 42

-

Question 43

mergeRequestsFetchEffect

Answer 43

-

Question 44

view

Answer 44

-

Question 45

effects are data

Answer 45

side effects and impurity happen outside

Question 46

our application

Answer 46

-

Question 47

time travel debugger

Answer 47

-

Question 48

Links

Answer 48

-

Question 49

Thank you

Answer 49

hope I got you a little bit excited