Tomasz Wróbel

Tweets about: Ruby, Event-Driven Architecture, CLI, Async Remote

Professionally, former VP of Small Talk at @arkency

Personally, a sheep 🐑

[Tomasz Wróbel @tomasz_wro]

@happyboygetsjob It gets printed to stdout. My workflow was to copy it from the test output and manually paste it to the test file.

[Tomasz Wróbel @tomasz_wro]

Blogged: https://t.co/72w4dw8xOW

i.e. how to get an almost-like-VCR experience, but without VCR

[Tomasz Wróbel @tomasz_wro]

L3: `zM zr zr zm zm zR`, repeat

It actually feels like real KATA 🥋

[Tomasz Wróbel @tomasz_wro]

Learn code folding in vim by practicing simple kata

(set foldmethod=indent)

L1:
- go to an indented line
- `za`
- `za`
- repeat

[Tomasz Wróbel @tomasz_wro]

L2:
- (with all folds open)
- `zM`
- `zr`
- `zm`
- `zR`
- repeat

It's actually more fun when you don't know what these commands do :)

[Tomasz Wróbel @tomasz_wro]

@_m_b_j_ Interesting. I agree, you should write a blogpost :)

[Tomasz Wróbel @tomasz_wro]

@_m_b_j_ > domain logic takes magnitudes more DB state as input than request state.

Interesting take. But still it deals with "where the logic happens" instead of "what language it's expressed in".

Would you still hold your point if Ruby was the lang in the DB instead of plpgsql?

[Tomasz Wróbel @tomasz_wro]

@andrzejkrzywda most probably @mlomnicki :)

[Tomasz Wróbel @tomasz_wro]

@_m_b_j_ If it's the latter, then how feasible it is to have Ruby running inside the DB runtime? (or any other language that _sparks joy_)

[Tomasz Wróbel @tomasz_wro]

@_m_b_j_ Do you mean SQL is a better language to express domain logic? (in which case writing more SQL in a ruby appserver would also work)

Or do you want the the domain logic to run inside the DB runtime to avoid crossing the process/network boundaries?

[Tomasz Wróbel @tomasz_wro]

There's 260 chapters in the New Testament.
There's 260 weekdays in a year.

If you think this makes a nice occasion to read through the whole of it, here's a reading plan (standard book order with two helpful twists).

#NTin52weeks

[Tomasz Wróbel @tomasz_wro]

@yanoo_ Congratulations on making it to Ruby Weekly!

[Tomasz Wróbel @tomasz_wro]

@purinkle I think the difference lies in how they integrate with the rest of the app. Views are built from the SQL "state" (often private/internal), whereas read models are built from events, which aim to facilitate integration with explicit contract and loose coupling.

[Tomasz Wróbel @tomasz_wro]

New blogpost:

How does a read model differ from a cache?

https://t.co/ewj65t3bzC

[Tomasz Wróbel @tomasz_wro]

@purinkle Thanks. Yes, views (especially "materialized") also bear some resemblance to read models. Maybe it's a topic for another article.

[Tomasz Wróbel @tomasz_wro]

@pankowecki You can definitely look at it from an angle, where they don't differ — but this angle doesn't reveal too much :)

[Tomasz Wróbel @tomasz_wro]

@_Paul_Abrams_ Interesting perspective. I think I get your point. The thing is that apart from the reviewers attention cost, a PR-based process itself comes at a huge cost of impeding the development.

[Tomasz Wróbel @tomasz_wro]

Disadvantages of Pull Requests

https://t.co/ntcE7jaFI9

There's been some discussions about PRs recently. Here's my try to systematize the factors to help make better informed decisions

1. PRs promote long living branches, which means more merge conflicts.

[Tomasz Wróbel @tomasz_wro]

Super useful for pair programmers

- Github's `Co-authored-by` annotation
- Git's `--trailer` option to automatically add text to commit messages
- Git's snippets for trailer text

👇

[Szymon Fiedler 🇵🇱 @szymonfiedler]

When you’re pair programming a lot, you might want to mention co-authors of the commits by adding line below to commit messages

Co-authored-by: Jane Doe <[email protected]>

It’s supported by github and others: https://t.co/IlgUqRiHwp

Thanks for the tip @pawelpacana

[Tomasz Wróbel @tomasz_wro]

Good thread on what `null` can mean:
- value not initialized
- value not applicable
- value not specified
- value not found

[Tagir Valeev @tagir_valeev]

Some people think that the problem with 'null' is that doing null-checks is annoying and you can forget about them. In Java, this could be solved via annotations and IDE quick fixes. This problem is also well solved in Kotlin. But in fact, the problem with 'null' is different.

[Tomasz Wróbel @tomasz_wro]

@johnykov What we can do is
1. raise awareness of the hidden costs and help make informed decisions
2. suggest small steps for change, evolution instead of revolution, which isn't likely to happen

Your idea is a nice example for (2). Good luck on suggesting it to your manager.

[Tomasz Wróbel @tomasz_wro]

@johnykov Thanks for sharing your story.

I try to avoid thinking in absolute terms of what devs should or should not do in general. After all every company is free to impose any rules it wants.