Alexey Shamrin

Let's turn this issue into a question :) What is broken - documentation or implementation? I could try to fix it myself, but I don't know the original intent.

I think PR #103 is related to this issue.

> As a final feedback: I really struggle to understand _what Trio is for_? It goes to great length to explain how to do everything, but I just don't get...

@wwwtyro: > I _think_ that any untracked primitives and datasets should be automatically garbage collected (ignoring for now any associated draw commands that may be cached). ... Can you let...

@wwwtyro Thank you! Indeed, I forgot to check the parent class of `Scissor`. Good idea on keeping the y-axis!

Thank you for improving visual regression features, @aslushnikov! You may find the implementation experience of gemini-testing project useful. Some pointers: * [CIEDE2000 tolerance measure](https://github.com/gemini-testing/hermione#tolerance) * [ignoring (font?) anti-aliasing](https://github.com/gemini-testing/looks-same#comparing-images-with-ignoring-antialiasing), based on...

Creating a new library makes total sense! The `ctypes` approach is very interesting, I didn't know it was possible (or perhaps some of the relevant `ctypes` features were not yet...

Linking to Wikipedia may help: https://en.wikipedia.org/wiki/Shim_(computing)

@elprans Is it possible to apply backpressure on the socket level somehow? For example, through [`.pause_reading()`][1] and [`.resume_reading()`][2] methods? Are they available through `asynpg` API? For context, I'm [trying][3] to...

For comparison, `psycopg2` does support applying `LISTEN` backpressure, when used [in asynchronous mode][1]. It exposes the socket file descriptor via [`conn.fileno()`][2]. The descriptor can then be used by "is there...