[swift-evolution] [Out of scope] Discussion on general Darwin/GlibC module

Alex Blewitt alblue at apple.com
Wed Nov 9 12:58:20 CST 2016

Although out of scope for phase 1, something that keeps cropping up in a variety of Linux/Darwin Swift scripts is the conditional inclusion of Darwin or GlibC per platform. The last point was an observation that creating a 'nice' wrapper for LibC or a cleaned up POSIX API is a non-goal:


> I think it makes sense to have a cross platform “libc” which is an alias for darwin, glibc, or whatever, and just leave it at that.
> Other proposals for a “POSIX” module have gotten bogged down because inevitably the idea comes up to make the resultant API nicer in various ways: rename creat, handle errno more nicely, make use of multiple return values, … etc.  The problem with this approach is that we don’t *want* people using these layer of APIs, we want higher level Foundation-like APIs to be used.
> ...
> I think we should formally decide that a “nice” wrapper for libc is a non-goal.  There is too much that doesn’t make sense to wrap at this level - the only Swift code that should be using this is the implementation of higher level API, and such extremely narrow cases that we can live with them having to handle the problems of dealing with the raw APIs directly.
> -Chris

I have created a draft for a proposal to create such a module. Comments are welcome.



# Libc module for Swift

* Proposal: [SE-NNNN](NNNN-filename.md)
* Authors: [Alex Blewitt](https://github.com/alblue)
* Review Manager: TBD
* Status: **Under discussion**

## Introduction

When running on Darwin, the base module is called `Darwin`. When running
on Linux or other operating systems, it's called `GlibC`. 

This repeatedly leads to code such as:

    #if os(Linux)
      import Glibc
      import Darwin

As the set of operating systems evolve, one of these conditional imports
needs to be updated. Instead of repeating this, make it available via a
standard `Libc` module in the base Swift library.

Swift-evolution thread: [Discussion thread topic for that proposal](https://lists.swift.org/pipermail/swift-evolution/Week-of-Mon-20161003/027621.html)

## Motivation

The [set of platforms](https://github.com/apple/swift/blob/fdf6ee20e4ca1fd32482f4b7b88a97ebdda52cd2/lib/Basic/LangOptions.cpp#L26-L36)
that Swift currently runs on can be divided into two; Darwin and XNU based systems
(macOS, iOS, watchOS, tvOS), Windows, and Unix based systems
(Linux, FreeBSD, Android, PS4). 

The base module on Darwin is called `Darwin`, while on Linux and
other Unix systems the base module is called `Glibc`. The base
module is typically conditionally included when working at a lower layer
than Foundation (which has the same detail involved in importing the
base module).

As a result, conditionally importing the right version typically uses
a conditional test based on the operating system, and the same code is
seen in a number of different modules, both internal to Swift and external:

* [Test for mmap in stdlib](https://github.com/apple/swift/blob/07b196d2f9a5facc490b35e3649e18937796239b/test/stdlib/mmap.swift#L4-L9)
* [Validation test for PassIfChildCrashedDuringTestExecution](https://github.com/apple/swift/blob/c3b7709a7c4789f1ad7249d357f69509fb8be731/validation-test/StdlibUnittest/ChildProcessShutdown/PassIfChildCrashedDuringTestExecution.swift#L4-L9)
* [Kitura's Socket definitions](https://github.com/IBM-Swift/BlueSocket/blob/49c5af8b6953cecc8674a7fcf746fa27a72c056a/Sources/Socket.swift#L21-L25)
* [Vapor's HTTP Server](https://github.com/vapor/engine/blob/1f95094ee470408309e98dd56b2251210d6a2a3d/Sources/HTTP/Models/Server/HTTP%2BServer.swift#L1-L5)

Some have already created a `Libc` module that effectively does what this
proposal suggests, such as [Vapor's Core Libc](https://github.com/vapor/core/blob/master/Sources/libc/libc.swift)

    #if os(Linux)
      @_exported import Glibc
      @_exported import Darwin.C

Each of these examples has subtly different behaviour; for example,
whether or not the os tests only include Linux (and then fail over to
Darwin), or whether they contain other Unices such as FreeBSD and Android.

## Proposed solution

The solution is to formalise these patterns in the base Swift library
and present a `Libc` module that conditionally imports `Glibc` or `Darwin`
based on the correct platform. Additional operating systems can be added
and kept up to date with the list of supported operating system conditionals
and including a failure message when an unknown operating system is detected.

## Detailed design

This will add a `Libc` module for the standard library that re-exports
the correct import depending on the operating system:

    #if os(macOS) || os(iOS) || os(tvOS) || os(watchOS)
      @_exported import Darwin
    #elseif os(Linux) || os(FreeBSD) || os(Android) || os(PS4)
      @_exported import Glibc
      fatalError("Libc not supported on operating system")

As new operating systems are added or become supported (such as Windows)
the standard imports can be added appropriately to this module.

## Source compatibility

There is no impact to source compatibility, since this proposal is additive.
Existing source code will work regardless of if this module is used or not.
However it improves source compatibility going forwards, since as new
operating systems are added this file will be updated, instead of the change
having to be made in multiple open-source projects.

## Effect on ABI stability

There is no impact to ABI compatibility, since this proposal is additive.
Existing source code will work regardless of if this module is used or not.

## Effect on API resilience

There is no impact to ABI resilience, since this proposal is additive.
Existing source code will work regardless of if this module is used or not.

## Alternatives considered

The first alternative is to do nothing. Existing Swift projects already
conditionally import these modules, or import a higher-level module (such
as `Foundation`) that performs the conditional import.

The second alternative is to export sub-modules of the modules. Clang
permits imports of sub-modules, so it could be possible to import only
`Darwin.POSIX` and `GlibC.POSIX`. However, in Swift, importing a sub-module
makes the whole module available anyway, so the difference between importing
a whole module versus a submodule is irrelevant.

The third alternative is to explore creating standard functions (in Swift)
corresponding to POSIX functionality, but where the format of the return
results are known. This would require a per-operating system binding to
expose operating-system details such as the byte ordering of structures
as used in the various `getaddrinfo` calls. These may evolve out of future
evolution proposals and this does not conflict with those goals at this
stage. There are additional clean-ups that this could address, such as the
use of the (thread-local) `errno` which may not be reliably read from within
Swift. However, the (swift-evolution thread)[https://lists.swift.org/pipermail/swift-evolution/Week-of-Mon-20161003/027602.html]
calls this "the perfect being the enemy of the good". Instead of trying to
solve all of these problems, they should be handled by subsequent
proposals (such as (Johannes' proposal)[https://lists.swift.org/pipermail/swift-evolution/Week-of-Mon-20161031/028627.html]
regarding errno handling sent to swift-evolution previously).
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.swift.org/pipermail/swift-evolution/attachments/20161109/7308e637/attachment.html>

More information about the swift-evolution mailing list