syntax request: from std/[foo1,foo2] import nil ; from "." / [foo1,foo2] import bar1, bar2, bar3

[timotheecour]

proposal 1

accept this syntax:

from std/[foo1,foo2] import nil # equivalent to: from std/foo1 import nil; from std/foo2 import nil; 
from "." / [foo1,foo2] import nil # ditto

proposal 2

furthermore, also accept this syntax

from "." / [foo1,foo2] import bar1, bar2, bar3

with meaning:
find each symbol bar1,bar2,bar3 somewhere in import list
The semantics are clear:
CT error if one of these symbols can't be found in import list
no error if a symbol is found in more than 1 module in import list (the usual overloading rules shall apply)

benefits

• more concise syntax
• in particular, proposal 1 promotes better code for users of that new feature: less symbol pollution (since it requires fully qualified names) compared to import std/[foo1,foo2]

[krux02]

I don't like it. Not importing everything from a module is rarely used feature. I don't see the need to improve the syntax for it.

[timotheecour]

is rarely used feature

I use it when I have a lot of imports in a module and I want to reduce likelihood of symbol clashes.
The way I intend to use this feature (from proposal 1) is:

from std/[foo1,foo2] import nil # I use only a few from these; I'll need to fully qualify them
import std/[foo3,foo4] # I use a lot of symbols from these

[awr1]

I think deciding to use fully qualified imports is a matter of personal style rather than something that should be outright discouraged. I prefer to just use import too, but sometimes I need to more careful about avoiding name collisions. I like the 2nd proposal too.

Why the "." / syntax? Since we can import multiple modules in the same directory with import x, y, z, should not from a, b, c import nil suffice?

Also I feel the need to restate here that this does instigate a grammar change, albeit a small one. I'm iffy on grammar changes at this point but for this I don't really think it's too much of a big deal.

[timotheecour]

Why the "." / syntax?

that RFC doesn't mandate using "."/, it's an orthogonal issue: from a, b, c import nil would work as well. The reason I used it here is because import foo is ambiguous: it can either import foo.nim in path or foo in currentModule.parentDir/foo.nim

implementing this (proposal 1 in particular) is easy and makes nim's features more orthogonal

[github-actions]

This RFC is stale because it has been open for 1095 days with no activity. Contribute a fix or comment on the issue, or it will be closed in 7 days.