Mathias LANG

> `in` is defined as `const`. Except for a brief period (< 5 months) in 2018, `in` has **always** be defined as `const scope`, as @pbackus [pointed out](https://issues.dlang.org/show_bug.cgi?id=23175#c3). > With...

@schveiguy : No, you said that before I changed the meaning of `in`, it wasn't legal for it do do this (this being for the compiler to consider the parameter...

@schveiguy > In means const. Is it legal to make this optimization with const instead of in? If not then it’s not legal for in. “ **When not in use**,...

@dkorpel : > Possibly introducing memory corruption without warning is a bit more severe than that, so a transition helper is in its place. The change that introduced the memory...

> One of his latest comments suggested otherwise: > https://forum.dlang.org/post/[email protected] Haven't seen that one, great :) > I'm not, I'm trying to recall what the problem with adding `ref` was....

Ergh, `--retry-all-errors` was introduced in cURL 7.71.0, which was released [2020-06](https://curl.haxx.se/mail/archive-2020-06/0055.html), a tad too recent... Will switch to a manual retry then...

Why do you want to remove it ? Is it blocking something ? In another PR, you mentioned: > The replacement for `T[]...` is `scope T[]`. But it's not quite...

> 1. it is rarely used I've seen it used at Sociomantic quite a bit. A quick grep on Ocean returns over 25 occurences. Haven't looked at other libraries, but...

@WalterBright : I believe this gives us enough evidence that they are, in fact, much more used than originally thought.

CI points to a lot of things not being nothrow yet.