It seems there is a different raises behaviour for newTerminal()

Example

import terminal

proc test() {.raises:[IOError, ValueError].} =
  setBackgroundColor(stdout, bgRed)

test()

This will compile on Linux, but not on Windows. Error:

C:\Users\deme\repos\getterminalraises.nim(3, 22) templa

I've been using Sourcery to generate documentation automatically for GitHub wiki (i.e. generate markdown files based on types & annotations), as it's pretty user friendly on there (as in "easy to read") and also very easy to do in templates thanks to Sourcery's capability to generate files on the fly.

However, using annotations is not very "coder-friendly" though, as it gets hard to maintain be

If you make a new class C that inherits from a class B with a registered hy-repr function f-B, but don't register a function for C, then hy-repr on an instance of C will call f-B, which can be misleading: it looks like you have an instance of B when actually you have an instance of C. We should probably fall back on repr instead.

For example, pandas's Timestamp class inh

curl --location --request POST 'http://example.com/api/test' \
--header 'Content-Type: multipart/form-data' \
--form 'file=@/C:/Users/Example/Pictures/index.png' \
--form 'name=test'

This is how you do multipart with http. This is for dio.

Justification
As a convenience to aid support issues, the "friendly" name of the Windows installation should be looked up from the raw version number.

Description
Currently, the RD logs OS version shows as a raw number, e.g.
Operating System: Microsoft Windows NT 6.2.9200.0 x64

Whilst this is technically correct, it's not particular

inline templates:

• with anonymous classes
• with lambda

Currently, the website is hosted in the repo https://github.com/scalameta/tutorial. We might want to consider having the website in the main repository. The benefit is that documentation would be closer to the code.

Add API nameof::cstring to https://github.com/Neargye/nameof/blob/master/doc/reference.md

Expected Behavior

example.tutorial_basic_first_steps_with_dynamic_bindings should compile.

Actual Behavior

gcc 9 may be released as early as tomorrow. Using its current git build, or using the new gcc 10 git build, fails with the error below.

No pull request coming from me on this. :-) I'm just temporarily commenting out the is_same<>.

[ 40%] Building CXX object test/

Currently in the documentation it shows that that you should install scalafix in the following way:

addSbtPlugin("ch.epfl.scala" % "sbt-scalafix" % "0.9.13")

However, 0.9.13 is the current version of core, not the plugin. However, the version of core gets put into the plugin here. This may not be common as they may always be in sync, but this might confuse people as they try t

Weaver's documentation lacks guidance about how to solve dependency cycles.

The most common case is for delegate patterns, which can be solved by using a self reference.

Other types of cycles (ones described in #140) aren't allowed by Weaver, and the documentation should show how to work around them.

Would you consider putting Boost.SML into vcpkg? I know it's only a header, and trivial to install, but vcpkg is just so nice. At the very least, it would cause some people to find out about Boost.SML for the first time.

An example: https://github.com/microsoft/vcpkg/blob/master/docs/examples/packaging-github-repos.md

If you issue :help on the repl you get a list of available commands, among others. But the list is incomplete, e.g. :modules, :undeclare or :unimport are missing.

Add a section to the documentation describing some of the techniques used under the hood. It should serve as a reference for contributors, as well as a gentle introduction to template metaprogramming for the uninitiated.

@lberti could you commit the falling sphere case (geo+cfg+json) ?
please add a README.adoc to explain the expected results

@vincentchabannes needs it to test new developments