[swift-evolution] Swift null safety questions

[Elijah Johnson]

I’ve skimmed through it just now. What we are talking about is “Logic
Failures” and that is handled in three small paragraphs of the
ErrorHandling document the rest (universal, recoverable, simple domain
errors) are separate. Basically, to summarize the Logic Failures section,
there is no rule yet set down “yet".

Quote:

The correct handling of these error conditions is an open question and is
not a focus of this proposal. Should we decide to make them recoverable,
they will likely follow the same implementation mechanism as universal
errors, if not necessarily the same language rules."


The “universal errors” section referred to is equally open-ended. To sum up
the two sections, there seems to be an implicit proposal to add universal
exceptions in the future (“logic failures” and Obj-C/C++ exception support)
in the future, with a few concerns about discouraging its use. My concern
about webservers is mentioned directly in the “Logic Failures” section,
"High-reliability systems may need some way to limp on even after an
assertion failure.”.

So, granted that the proposal is more or less already there, I would
propose typed exceptions for "Logic Failures", Obj-C, and C++ exceptions.

If someone needs more discouragement than that, then all that needs to be
done is to make the syntax cumbersome. Like having to pass a closure to a
function instead of the clean do{} catch{} syntax, which btw is the way
that it is done now for those who need to catch Obj-C exceptions from
Swift. See
http://stackoverflow.com/questions/35119531/catch-objective-c-exception-in-swift
for
that. I can say that I’ve used it myself and that I am extremely
discouraged from using it to throw an NSException from Swift, or even to
catch the Obj-C exceptions for that matter.



On March 12, 2017 at 3:20:41 PM, Austin Zheng (austinzheng at gmail.com) wrote:

If you want improved error handling, feel free to put together a proposal
once Swift 4 phase 2 is over. A good proposal should take into account the
reasons Swift's error handling was designed as it currently is (
https://github.com/apple/swift/blob/master/docs/ErrorHandlingRationale.rst)

Otherwise, if you really need to build a web server framework that can
swallow any sort of application logic without choking, perhaps Swift as it
currently stands isn't the best tool for the job.

Best,
Austin

On Mar 12, 2017, at 12:06 PM, Elijah Johnson <ejrx7753 at gmail.com> wrote:

The problem for using Swift on the server is one of these optionals (or
array index out of bounds error, or cast from Int to UInt error) would
crash the entire server directing many requests. Granted, thats not Apple’s
busininess, but it is pretty much standad in a modern webserver never to
crash, and whoever is developing servers will find they don’t have the
necessary power to do that. Some languages like PHP have a unique process
per-request style, but that phrohibits shared data. One could do the same
with Swift, but it would be a major loss.

Just because Swift is not garbage collected doesn’t mean that you can’t
safely unwind the stack and continue execution somewhere. If “bad”
developers would start using this as unwanted exceptions, then it can be
made inconvient somehow.

Think about a server that handles 1,000 requests per minute - its going to
happen evenutally, even despite testing, and chances are that once it
happens it will re-occur - and now the server goes down every minute or
even every few seconds and needs an auto-restart loop.

If we’re going to be talking about “bad developers” as in ones who might
once code a bug with an optional or index out of bounds exception, then
that is going to be a very high standard and result in a very mild adoption
of server side Swift. Maybe it would take off for a while, and then when
people realize these things, they will say “thats why everyone writes their
server-side code in Java/C#/PHP, etc” and Swift will be a thing for people
who need ultra-fast performance on a server.

Its not really Apple’s business, and clearly you are trying to replace only
C++ and not Java/C#, but if you can add a compiler mode with exceptions
instead of fatalError(mesage:) calls, that would contribute a lot to making
web servers with Swift. Even if you leak the entire call stack, it would
still be better than crashing an entire web serer.


On March 6, 2017 at 9:47:03 PM, Austin Zheng (austinzheng at gmail.com) wrote:

Swift already is an unmanaged language, and is far closer to C++ than Java.
There is no equivalent to the CLR or JVM and it is trivially easy to call
into C code or the unsafe APIs and do nearly anything that C can do with
regard to direct access to memory.

The force-unwrap operator was never intended to model a Java-style runtime
exception. Instead, it is an assertion which indicates a programmer logic
error. The preferred way to handle an optional at runtime has always been
to use the conditional-unwrapping functionality the language offers. The
correct way to report errors at runtime that aren't due to programmer error
is through the do-try-catch mechanism. Pervasive, unnecessary use of '!' is
a project management and/or developer competence issue, and not something
the language should go out of its way to fix.

Sugaring 'if let' for programmer convenience is something on the commonly
rejected proposals list:
https://github.com/apple/swift-evolution/blob/master/commonly_proposed.md

Regards,
Austin


On Mon, Mar 6, 2017 at 6:27 PM, Elijah Johnson via swift-evolution <
swift-evolution at swift.org> wrote:

> Additionally, not discarding previous comments, one way the language could
> be improved is by replacing (or rather augmenting) optional binding with a
> faster syntax.
>
> ex.
>
> func test( name : String?, name2 :String? ){
>      name! , name2! {
> // inside this block, “name” and “name2” are non-optional
>     } else {
>       // no forced unwrapping, just optional else case
>    }
> }
>
> Whereas the alternative (current optional binding), requires a
> re-definition of variables and an additional “if” operator. Besides the
> extra characters involved in typing and reading it, usually variables have
> specific names that the user does not want to augument.
>
> func test( name : String?, name2 :String? ){
>     if let nameUnwrapped = name , let name2Unwrapped = name2 {
>
>     } else {
>    }
> }
>
> func test( name : String?, name2 :String? ){
>     guard let nameUnwrapped = name , let name2Unwrapped = name2 else {
>
>     }
>
> }
>
> The fact that we don’t need a “nameUnwrapped” variable (which is unsightly
> and requires refractoring) is one thing that still makes forced unwrapping
> still very much a necessary or common part of Swift coding, even where it
> could be replaced with optional binding.
>
>
>
>
> On March 6, 2017 at 7:20:40 PM, Elijah Johnson (ejrx7753 at gmail.com) wrote:
>
> Hi,
>
> I’ve been recently considering Swift on the server-side, and there came up
> the question, “what happens when a null optional is forcibly unwrapped?”
> and the answer was clearly that not only would the request crash, but the
> entire server would also crash, since the server itself is/would be also
> written in Swift.
>
> I think that this highlights serveral weaknesses in Swift’s “null safety”
> attempts. The first is that there needs to be a way to isolate blocks of
> code so that a “null pointer exception” does not crash the system. I think
> it is fair to say, in the language of Java, that these are really “runtime
> exceptions” and for some reason are being treated as segmentation faults
> would be in C/C++. In my opinion, the existence of these things has the
> ability to bring Swift down below Java and closer to an unamanged language.
> Not really sure why it would ever be desireable, but in terms of
> server-side programming, it is definitely a serious issue.
>
> Secondly, Swift’s “null safety” is really completely undermined by these
> “force-unwrap” errors. I agree with the usage of new language features like
> guard, optional binding, etc to remove a null value, but I see this
> force-unwrap as extremely pervasive for a lot of reasons:
>
> 1. When porting code from a C style language to Swift, force-unwrap is
> needed to make the code work without refractoring.
> 2. XCode itself reccomends this operator and when it is used to satisfy a
> requirement, then it can be left in the code
> 3. Some styles of coding just can’t work without force-unwrap.
>
> I don’t think the logic of making the consequences extreme for making a
> mistake are a rational for making better coders. In fact, I personally see
> the “force-unwrap” operator having very great potential usage as a
> deliberate assertion exception - the programmer demands that a certain
> value be non-null for execution to continue, only there should be a proper
> scope and context where these exceptions can propogate to. On debug modes,
> one might want it to pause on the line, but for other uses, it should be
> caught IMO - on regular code by the user, and inside dispatch blocks by the
> queue itself. For a multithreaded app or server to exit, the user should
> have to explicitly write exit(0), isn’t this the goal of a managed
> language? Maybe in some cases, Apple will want the program to crash, but if
> Swift is given an audience not just with Apple hardware, then it should
> have more flexibility IMO.
>
> Elijah
>
>
> _______________________________________________
> swift-evolution mailing list
> swift-evolution at swift.org
> https://lists.swift.org/mailman/listinfo/swift-evolution
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.swift.org/pipermail/swift-evolution/attachments/20170312/f6e58035/attachment.html>