Disqus - Latest Comments for hubertlepicki

Re: Mock current time in Elixir - Adam Niedzielski

hubertlepicki — Thu, 19 Apr 2018 02:58:58 -0000

Hey, so if you are still looking on a way to do it concurrently, I think you can use https://github.com/platafor...

Re: Ketoza i dieta ketogeniczna: FAQ / Q&A

hubertlepicki — Mon, 05 Feb 2018 04:55:15 -0000

Bardzo fajny artykuł. Mam co linkować znajomym pytającym o dietę ketogeniczną.

Pierwszy problem który znalazłem to że:
- stewia może powodować wzrost poziomu insuliny. To jest błąd, który może wynikać z tego że w sprzedaży są słodziki które zawierają stewię *oraz inny słodzik* np. maltodekstrynę. Trzeba czytać skład. Czysta stewia nie ma prawa podwyższać poziomu insuliny, nie jest cukrem, nie jest trawiona jak cukry a nawet sama ilość stewii wystarczająca do słodzenia jest tak mała że gdyby nawet była cukrem to nie miała by znaczenia. Podobnie rzecz ma się z resztą z sukralozą (którą można z powodzeniem stweię zastąpić jeśli nie lubimy posmaku stewii).

Drugi problem to że wymieniasz że na diecie musisz jeść ekologiczne i zdrowe produkty. Wcale nie, możesz zjadać 8 hamburgerów z makdonalda i dopychać się wysoko przetworzonymi parówkami i mortadelą i ketoza będzie działać. Jedzenie zdrowszych produktów będzie miało pozytywne skutki oczywiście ale jeśli chodzi nam jedynie o spalanie tłuszczu to nie ma znaczenia czy jemy jagnięcinę wypasaną na Islandzkich łąkach, czy wieprzowinę z polskiej fabryki.

Re: Things I wish ActiveRecord had after using Ecto

hubertlepicki — Tue, 23 Jan 2018 14:13:10 -0000

You missed two fairly important and big features:

1. Ecto.Multi - a way to wrap up transactions in a monad-like thingy
2. Different way database connections are handled, i.e. are used only when needed. ActiveRecord / Rails wastes a lot of time keeping idle connections open, because of the way ithandles transactions and associations it needs to keep those open. https://www.amberbit.com/bl... database connections

Re: Things I wish ActiveRecord had after using Ecto

hubertlepicki — Tue, 23 Jan 2018 14:10:39 -0000

but it's not the kind of composition you might know from ActiveRecord...

Re: (Abusing) Ecto as Elixir data casting and validation library

hubertlepicki — Thu, 28 Dec 2017 04:45:45 -0000

in short: I have two extra layers. Dumb "schemas" that only have generic 'changeset' function allowing you to set any attributes, and they also declare associations. No business logic here. And then I have a layer of entities managers, that take data from business layer (services) and persist it using database access layer (schemas).

Re: (Abusing) Ecto as Elixir data casting and validation library

hubertlepicki — Thu, 28 Dec 2017 04:44:18 -0000

I am putting together a (mini) book that will get to that topic in details. The above is pretty much extract from one of the chapters. I will put together a readable beta version shortly.

Re: (Abusing) Ecto as Elixir data casting and validation library

hubertlepicki — Thu, 28 Dec 2017 03:19:16 -0000

good point!

Re: (Abusing) Ecto as Elixir data casting and validation library

hubertlepicki — Wed, 27 Dec 2017 13:58:56 -0000

yeah... that's probably the most confusing part. The missing `:action` got me more than once, so just had to write this stuff down, will better remember myself :D

Re: AmberCasts #1 Ruby on Rails development environment with Docker and docker-compose

hubertlepicki — Sun, 26 Nov 2017 02:50:27 -0000

hey, that's just Gnome on Arch Linux over there :)

Re: Time zones in PostgreSQL, Elixir and Phoenix

hubertlepicki — Fri, 10 Nov 2017 02:01:32 -0000

See this is some sort of compromise. Yes, it is unsafe, but if you have events that span across multiple time zones you will never have a silver bullet. In case of a time zone change, the event's start date would change for at least some of your users.

That was the exact choice I was given in fact and we decided that having a common denominator in form of UTC is not the worst decision ever.

Re: Elixir :applications vs :extra_applications - a guide

hubertlepicki — Tue, 10 Oct 2017 11:53:34 -0000

Glad I could be of a help

Re: A different approach to testing elixir with mocks/doubles

hubertlepicki — Fri, 29 Sep 2017 06:48:25 -0000

ah. I like the callbacks, it warns you if something is defined in mock but not in real implementation. Does the other thing do that too?

Re: A different approach to testing elixir with mocks/doubles

hubertlepicki — Fri, 29 Sep 2017 06:37:05 -0000

You should check out mox. https://github.com/platafor...

the current (unreleased) version of master has the support for Rspec-like mocks and stubs, i.e. function calls that are required to be called and function calls that are optional. It's pretty handy.

Re: Time zones in PostgreSQL, Elixir and Phoenix

hubertlepicki — Thu, 03 Aug 2017 07:26:32 -0000

@Michał Muskała thanks for the tips! I have corrected the article. I have also copied your example for reference (and attributed to you) hope that's fine!

Re: How learning Elixr made me better Ruby developer

hubertlepicki — Fri, 28 Jul 2017 05:41:12 -0000

I don't see how this is related. You don't modify the data in place, you return new objects from each step in the pipeline.

Re: How learning Elixr made me better Ruby developer

hubertlepicki — Thu, 27 Jul 2017 10:24:24 -0000

That's true! I'm missing true pipe operator *and* pattern matching in function heads a lot when writing Ruby code :(. Would not hurt if it checked compile-time function names for typos either. Those are the 3 things Elixir does better for sure :)

Re: How learning Elixr made me better Ruby developer

hubertlepicki — Thu, 27 Jul 2017 09:20:52 -0000

True that!

Re: AmberCasts #1 Ruby on Rails development environment with Docker and docker-compose

hubertlepicki — Wed, 14 Jun 2017 14:55:14 -0000

I know that docker-compose down will discard those volumes. But they persist after reboots for me... Hmm I'm pretty new to the whole thing so may not be understanding something and I may be doing something wrong.

Re: AmberCasts #1 Ruby on Rails development environment with Docker and docker-compose

hubertlepicki — Wed, 14 Jun 2017 14:38:11 -0000

ok, but somehow docker-compose always mounts the same volume for me. So unless I want to specify my own directory I don't really have to specify anything more than:

  db:
    image: postgres

Unless I am mistaken? :)

Re: Handling URL redirects in Phoenix with Plug

hubertlepicki — Fri, 21 Apr 2017 05:34:45 -0000

So your plug should look like something like this:

defmodule RedirectsPlug do
import Plug.Conn

def init(options), do: options

def call(conn, options) do
to = options[conn.request_path]
do_redirect(conn, to)
end

defp do_redirect(conn, nil), do: conn

defp do_redirect(conn, to) do
conn
|> Phoenix.Controller.redirect(to: to)
|> halt
end
end

and don't forget to also include it in endpoint.ex *before* call to router plug since putting it in router will no longer work either.

Re: Handling URL redirects in Phoenix with Plug

hubertlepicki — Fri, 21 Apr 2017 05:28:29 -0000

As the comment above suggests, you should use request_path instead.

Re: Elixir - different kind of promises

hubertlepicki — Thu, 09 Mar 2017 07:07:11 -0000

name a serious project where this wasn't the case eventually...

Re: Creating Elixir libraries as OTP applications

hubertlepicki — Tue, 31 Jan 2017 08:26:18 -0000

Than you so much @sasajuric on clarifying this. I'll update the post.

Re: How to deploy Rails App on a VPS

hubertlepicki — Mon, 30 Jan 2017 04:19:41 -0000

The easiest way to do it, is to create multiple users on the server. I usually do it this way:

/home/app1
/home/app2
/home/app3

etc. For each app, you can then use very similar capistrano deployment strategy, just replace the username with name of server user.

You also want to create multiple nginx server config files (we put them in /etc/nginx/sites-available and symlink to /etc/nginx-sites-enabled), where your server_name will be unique, and upstream name as well (it's a gotcha). You point each nginx config file to individual app.

This way you can enable/disable those individual apps as you please, by simply removing the symlink from sites-available directory and reloading nignx. Hope that helps!

Re: Getting ready for new concurrency in Ruby 3 with Guilds

hubertlepicki — Fri, 09 Dec 2016 04:45:18 -0000

Do you have any idea if Ruby 3 will have some sort of mechanism for dynamically controlling the guilds? I mean, what guilds are started, catching the exceptions, restarting them when needed etc?

I am mind blown with Eralng's handling of concurrency. Not only it's pretty easy to understand parallelism (thanks to immutable data structures), but the supervision / control mechanism is quite awesome.

Threads are a mess in most languages. Erlang's the only solution that seems elegant to me, one that we, programmers with limited brain capacity can actually understand, design and control.

Guilds so far look like another layer of abstraction, and APIs to learn, not sure if that will do more harm or good in the end.