Cerber-Ursi

In the mentioned case, the date on spec generation is not parsed at all. I'm not sure how to deal properly with data objects in this case.

Got similar problem also with `$map` operation. The following code works (with the same base model) only without strict mapping: ```java AggregationOperation project = Aggregation.project() .and( VariableOperators.mapItemsOf("names") .as("name") .andApply(StringOperators.valueOf("$$name").lengthCP()) )...

If I understand correctly, the request might be better forwarded to the underlying HTTP implementation, i.e. https://github.com/tiny-http/tiny-http. Doesn't have the time to formulate it properly right now, though.

In fact, the fix seems to be easy, so sent the #96 with patch.

> You want to turn the following into a compiler error, right? (Problem is that a CreateWindow creates a window while FreePixmap expects a pixmap and will produce a runtime...

I did a little digging into the types in question, and here's the simplified reproduction code (minimized from emotion's code base): ```typescript import * as CSS from 'csstype'; type CssValue...

> Luckily there doesn't seem to be any performance impact from using Scope::new() in the hot path. That's expected, I guess? `const`s are essentially inlined into every usage place, so...

`const` items are not bit-copied (semantically - their creation can be _optimized_ to bitwise copy, but so can be `Scope::new` itself). Using `const` as a value is exactly the same...

The code works with `|&(ref a_item, ref b_item)| a_item == b_item`, however, so probably it's not a false positive, it's simply that the help message could be more clear?

The backtrace points on [`build.rs:80`](https://github.com/DDoSolitary/yasfw/blob/master/build.rs#L80), that is, on the following line: ```rust let vcpkg_toolchain = format!("{}/scripts/buildsystems/vcpkg.cmake", env::var("VCPKG_ROOT").unwrap()); ``` Seems that the `VCPKG_ROOT` environment variable is simply not set. Do you...