Do you want to request a feature or report a bug? Feature

When compiling a delay_span_bug testcase (e.g. the last snippet in rust-lang/rust#63154 (comment) on the current nightly, or any of the other examples in that issue on older compilers) with -Cincremental (or Cargo which does that by default), only the initial run will ICE.

Subsequent runs will succeed without an ICE, in the absence of any regular error

Renovate is fantastic but because of our config at the moment, we get several updates in a day.

We should probably do two things

• batch all minor updates in a single PR
• schedule them at a reasonable time (twice a week, perhaps at 10 am GMT)

This should go in config at https://github.com/gatsbyjs/gatsby/blob/master/renovate.json

Documentation for options is at https://docs.renovat

If the flow plugin runs before the class properties one (or if the class properties plugin isn't used), we remove uninitialize class fields:

class Foo { x: string }

currently becomes

class Foo {}

while it should be

class Foo { x }

If someone still wants the old behavior, they can either use flow comments (as [offically recommended](https://github.c

Create the @parcel/transformer-markdown package in packages/transforms/markdown. This package is a Parcel 2 transform plugin for markdown. It should be based on the MarkdownAsset from Parcel 1.

The @parcel/transformer-markdown package is responsible for the following things:

• Compiling markdown to HTML using marked

Is your feature request related to a problem? Please describe.
In attempting to move logic from components to plain javascript files, I believe I've discovered that I can't use the $ syntax to autosubscribe to stores except in svelte files. Converting my code from a js file to an empty component on the page fixed the errors I was getting.

Describe the solution you'd like
I'd either li

Marked version: 0.7.0

• December 2017: Should we perform a miracle?
• February 2018: MarkedJS Organization is opened. Core team and roadmap established.

This is a community effort. The core team is primarily here for c

Hi,

I get the following warning when compiling my code:

 Warning Cannot resolve 'config.getServerMode() ? '../libserver/server_code.js' : '../liblocal/local_code.js''
  Use a string literal as argument for 'require', or leave it
  as is and specify the resolved file name in 'scripts' option.

I understand the reason for this and what I need to do. As it says in the manual: ```

In the past, when GopherJS was in early development stages, many packages were listed in the compatibility table as partially supported, simply because it was a work in progress, and they were rapidly progressing towards fully supported.

By now, GopherJS is a lot more mature, with only 12 packages still listed as partially supp

When you run finish.exe twice you get (I decided I wanted the start menu entry after the fact)
C:\Users\<USERNAME>\.nimble\bin is not in your PATH environment variable.

This was an installation without admin rights. The variable is placed correctly in the user path without requiring (unnecessary) admin permissions.

Example

Run finish.exe twice without admin privileges on Windows 10

class Foo {
  field: i32;
  field: i64;
}
const foo = new Foo();

But this should be valid:

class Foo {
  field: i32;
  static field: i64;
}

Expected compiler error:

Dublicate identifier 'field'

Actual:

ERROR: AssertionError: assertion failed

Is your feature request related to a problem? Please describe.
Catch2 is advertised as a 'header-only' library - which leads one to expect it to be usable within executable code on Compiler Explorer, using the 'a.out' option.

However, Compiler Explorer incorporates Catch2 using its include directory, which contains all the component header/source fil

In the pic, B(0,0,A 2, A 1) should have some whitespace, the fix should be in the module caml_chrome_debugger

Problem

walt-cli package when linking multiple .walt files together can "wrap" the modules in a stand-alone JS module. The resulting module is too large because it serialized the dependency tree with the AST information encoded directly into the output, resulting in a massive amount of js.

Encoding the dependencies into the file is done to ensure the module can be used stand-alone in br

foo = []
"/event/:id".scan(/:(\w+)/) { |name,| foo << name }
puts foo.inspect

MRI:

["id"]

Opal:

[["id"]]

Currently function doc comments and error set doc comments are integrated, but not these:

• doc comments of struct fields
• doc comments of enum fields
• doc comments of union fields
• doc comments of structs
• doc comments of enums
• doc comments of unions
• doc comments of generic instantiations (related: #3407)
• doc comments of global variables

Just giving 0.6.28 a whirl. Now that underlying RPCCore exceptions are being reported (awesome) we might finally be able to track down the root causes of stuff that's been around for a while. Exciting! First cab off the rank:

[error] org.scalajs.testcommon.RPCCore$RPCException: java.io.UTFDataFormatException: encoded string too long: 93549 bytes
[error] 	at org.scalajs.testcommon.RPCCore.

The Contributor's Guide has the excellent paragraph

All code contributed to the user-facing standard library (the nativelib/ directory) should
come accompanied with documentation. Pull requests containing undocumented code will not 
be accepted.

I applaud the bold stand and would like to make it easier for contributors to achieve it.

The paragraph is good but raises a number of qu

We currently have zero documentation on the new features of tuples (everything in https://github.com/lampepfl/dotty/blob/master/library/src/scala/Tuple.scala)