Delyan Angelov

@enghitalo please avoid refactoring/renaming in the same PR, as functional changes. I know that it is sometimes unavoidable, but here, more than half of the changes are renames. Intermixing both...

Think of it as wading through a pile of straw (the renames), to search for a pearl, or a needle potentially hidden in it.

Use `typeof[T]()`, not `typeof(T)` .

I can not reproduce it on linux and macos.

Try disabling your antivirus service temporarily, while doing `v self`/`v up`, to see if that will affect the speed.

Try with `-d net_blocking_sockets -d use_openssl` . For me, this: ```v import net.http import sync.pool fn get_news(mut pp pool.PoolProcessor, idx int, wid int) voidptr { ticker := pp.get_item[string](idx) resp :=...

@felipensp  it works only with tcc, and even there, it is by implementation detail

Such a function could be simulated, by passing an additional parameter - the target address, and then doing the memcpy *inside* the function, when the fixed array is still valid...

```v fn f() [4]int { return [1,2,3,4]! } zzz := f() dump(zzz) dump(zzz[0]) ``` leads to a cgen error: ``` /tmp/v_1000/z.11365486265824014296.tmp.c:12604:57: error: subscripted value is neither array nor pointer nor...

imho that new structure, that has a `.ret_arr` field, has to be used *only* for returning fixed arrays, but it *should not* be used for their normal representation, *or* all...