Re: [swift-evolution] [Review] SE-0187: Introduce Sequence.filterMap(_:)
> • What is your evaluation of the proposal? -1 Code churn, new function name has the same issues as the old one. Benefits small. It is also likely that those suffering with the String collectionifying will have already solved their issues before this update could be introduced. > • Is the problem being addressed significant enough to warrant a change to Swift? No. > • Does this proposal fit well with the feel and direction of Swift? Distraction from major directions, little gain. > • How much effort did you put into your review? A glance, a quick reading, or an in-depth study? Read proposal and scanned the discussion so far in this review. We did actually hit the bug where a String was flatMapped and behaviour changed in Swift 4 migration a couple of weeks ago which caused some discussion to find out what the issue was but we solved it and there is a fair chance the renamed version could equally have had exactly the same issue. Joseph ___ swift-evolution mailing list swift-evolution@swift.org https://lists.swift.org/mailman/listinfo/swift-evolution
Re: [swift-evolution] [Review] SE-0187: Introduce Sequence.filterMap(_:)
> • What is your evaluation of the proposal? -1 Code churn, new function name has the same issues as the old one. Benefits small. It is also likely that those suffering with the String collectionifying will have already solved their issues before this update could be introduced. > • Is the problem being addressed significant enough to warrant a change to Swift? No. > • Does this proposal fit well with the feel and direction of Swift? Distraction from major directions, little gain. > • How much effort did you put into your review? A glance, a quick reading, or an in-depth study? Read proposal and scanned the discussion so far in this review. We did actually hit the bug where a String was flatMapped and behaviour changed in Swift 4 migration a couple of weeks ago which caused some discussion to find out what the issue was but we solved it and there is a fair chance the renamed version could equally have had exactly the same issue. Joseph ___ swift-evolution mailing list swift-evolution@swift.org https://lists.swift.org/mailman/listinfo/swift-evolution
Re: [swift-evolution] [Review #2] SE-0117: Default classes to be non-subclassable publicly
+1 I'm sold by Brent's reasoning that sealed is the only way to keep options open, both for any given library and for the whole Swift language. I've stayed quiet until now because while philosophically inclined to final by default I was concerned about potential complexity and confusion. I haven't read everything on the mailing list but I've read several blog posts on both sides. In terms of the necessity of subclassing it feels to me that subclassing where not designed for it is a similar in principle to using private APIs which may at times be the only way to achieve certain things but at least on iOS the App Store policies prevent that. I don't have very strong views about the syntax I'm afraid but if this is something that could be reversed there should be an explicit version of the default sealed case (like internal is with for access controls). Joseph On Jul 19, 2016, at 8:43 AM, Brent Royal-Gordon via swift-evolutionwrote: >> On Jul 18, 2016, at 11:04 PM, Gwynne Raskind via swift-evolution >> wrote: >> >> Denial of subclassing has always been opt-in in ever other language I’ve >> used (C++ and Java, to name two, and Objective-C (and older C++) don’t even >> have that much). Sealing a class against subclassing is one thing, but not >> providing any kind of escape hatch, any kind of >> IUnderstandThatSubclassingMayCauseSunsToGoNovaOrGalaxiesToExplode marker, >> hamstrings all users of the code. Opt-in sealing at least constrains this >> scenario to places where the framework writer thought it was worth adding >> the extra protection against horrible horribleness. > > You know, one thing I haven't seen mentioned is that, just as > sealed-by-default preserves the options of library programmers, it also > preserves the options of the language itself. > > Suppose the people who think this is a huge mistake are correct, and we > ultimately conclude that sealed-by-default is a disaster. What can we do > about it? Well, we change Swift 3+n to open all classes by default. This > would be source- and binary-compatible with all existing code; the only > wrinkle is that classes compiled with a sealed-by-default compiler would > still be sealed. (And that's not even a problem yet, since stable ABIs are > not a thing yet.) > > The reverse, however, is *not* true. Going from open-by-default to > sealed-by-default is source- and binary-incompatible. If we don't do it now, > we may never be able to do it. > > That means this is our one chance to try this. The outcome is uncertain; > there are good arguments that it will improve things, but there are also good > arguments that it will make things worse. But if we're afraid to try this > now, we'll never be able to try it again, and we won't know if it would have > worked. Whereas if we *do* try it now, we can always roll it back later. > > Software quality is one of the biggest problems our profession faces. We > handle crushing amounts of complexity, teetering towers of abstraction, > intertwined code that's at the ragged edge of our ability to comprehend it. > Quite possibly the most urgent need in our industry is tools to help us > manage it. > > Sealed-by-default might turn out to be a powerful tool for managing > complexity, helping us prevent unexpected interactions between the > implementation details of separate modules. Or it might not. But we ought to > find out. If we always take the conservative option, if we always stick to > the tried and true, we will never advance the state of the art, never find > solutions to the problems that make "all software sucks" a truism. > > We need to be bold and take a chance. If it doesn't work out, we can undo it. > But if it does work out, we'll be better for it. > > -- > Brent Royal-Gordon > Architechies > > ___ > swift-evolution mailing list > swift-evolution@swift.org > https://lists.swift.org/mailman/listinfo/swift-evolution ___ swift-evolution mailing list swift-evolution@swift.org https://lists.swift.org/mailman/listinfo/swift-evolution
Re: [swift-evolution] [Review] SE-0121: Remove `Optional` Comparison Operators
On 12/07/2016 19:26, Chris Lattner via swift-evolution wrote: The review of "SE-0121: Remove `Optional` Comparison Operators" begins now and runs through July 19. The proposal is available here: https://github.com/apple/swift-evolution/blob/master/proposals/0121-remove-optional-comparison-operators.md * What is your evaluation of the proposal? I approve. I actually filed a (rdar://19366632) in Jan 2015 about this behaviour. * Is the problem being addressed significant enough to warrant a change to Swift? Yes. It can catch out the unwary in several ways and is objectively surprising behaviour. * Does this proposal fit well with the feel and direction of Swift? Yes. * If you have used other languages or libraries with a similar feature, how do you feel that this proposal compares to those? Well Swift is the main language with explicit optionals that I have used but comparisions with nil/NULL vary in behaviour between languages. In SQL for example NULL is taken as unknown and even an equality comparision between NULL values fails. It could be worth going as far as removing the equality between optionals for this reason (although we could still keep the NilLiteralConvertible (or Swift 3 named version) for equality checks (I would also be fine with reusing the `is` keyword but haven't thought about that deeply. i.e. `if foo is nil {` * How much effort did you put into your review? A glance, a quick reading, or an in-depth study? Read proposal and review discussion so far but have also considered it in the past and have warned about the behaviour in talks in the past. Joseph -- Human Friendly Ltd. ___ swift-evolution mailing list swift-evolution@swift.org https://lists.swift.org/mailman/listinfo/swift-evolution
Re: [swift-evolution] [Review] SE-0113: Add integral rounding functions to FloatingPoint
> The review of "SE-0113: Add integral rounding functions to FloatingPoint" > begins now and runs through July 5. The proposal is available here: > > > https://github.com/apple/swift-evolution/blob/master/proposals/0113-rounding-functions-on-floatingpoint.md > >* What is your evaluation of the proposal? It is an improvement but would it be even better if the return type was Int (or possibly type inferred to Int16 etc.). If a float is really desired it can be initialised from the Int (or it could even infer float result is required). >* Is the problem being addressed significant enough to warrant a change to > Swift? Yes, reduces reliance on C APIs. API also clearer about semantics. >* Does this proposal fit well with the feel and direction of Swift? Yes >* If you have used other languages or libraries with a similar feature, > how do you feel that this proposal compares to those? >* How much effort did you put into your review? A glance, a quick reading, > or an in-depth study? Quick read and a quick look at earlier discussion linked from proposal. Joseph ___ swift-evolution mailing list swift-evolution@swift.org https://lists.swift.org/mailman/listinfo/swift-evolution
Re: [swift-evolution] [Review] SE-0109: Remove the Boolean protocol
https://github.com/apple/swift-evolution/blob/master/proposals/0109-remove-boolean.md > * What is your evaluation of the proposal? Agree. In extensive Swift use I've never had cause to actively use Boolean rather than Bool. If there is a need to keep it (which I don't see) I think it would be better renamed to something less succinct and confusing. * Is the problem being addressed significant enough to warrant a change to Swift? Yes. Removing things, especially confusing things is good when there is little impact. * Does this proposal fit well with the feel and direction of Swift? Yes. * How much effort did you put into your review? A glance, a quick reading, or an in-depth study? Quick read. Joseph ___ swift-evolution mailing list swift-evolution@swift.org https://lists.swift.org/mailman/listinfo/swift-evolution
[swift-evolution] [Draft Proposal for pre discussion]NS_ENUM's should be considered Open
Can I get some thoughts about this idea to remove undefined behaviour from switch statements when an enum imported from Objective-C has a raw value outside the expected set? The proposal is on my fork and below: https://github.com/josephlord/swift-evolution/blob/master/proposals/-consider_ns_enums_open.md Thanks, Joseph # NS_ENUMs should be considered Open * Proposal: [SE-](https://github.com/josephlord/swift-evolution/blob/master/proposals/-consider_ns_enums_open.md) * Author(s): [Joseph Lord](https://github.com/josephlord) * Status: [Early Draft] * Review manager: TBD ## Introduction Enums imported from C/Objective C could (as a result of library versioning or bugs) actually have raw values outside of the defined values. A default case should be required on switch statements involving them and the current undefined behaviour should be removed. Swift-evolution thread: [To be filled in after discussion] ## Motivation In cases where an enum value is outside of the expected range and there is no default in a switch the behaviour is currently undefined (it seems to take the first case in practice) potentially leading to runtime crashes. Undefined behaviour and crashes at runtime due to library changes outside the developer's control are highly undesirable and Swift should be helping us to make safe code to prevent them. At the moment if you are handling all cases it is not even possible to add a default case. I encountered the issue when testing on iOS 8 and hitting [an issue with CoreData](http://stackoverflow.com/a/32978387/1476206) where it was passing a zero value the NSFetchedResultsChangeType to my code. That thread clearly shows that the issue is affecting people and while it is clearly a bug in the unsafe code of Core Data that is something which Swift should be defending against. There is a simple test example showing the issue in [SR-1258](https://bugs.swift.org/browse/SR-1258) ## Proposed solution Redefine NS_ENUMs from being a closed type to an open one. This would then require a default case on all switch statements over the type. With the default case explicit situation must be handled by the developer. They can choose to raise errors call assertionFailure or preconditionFailure as the wish or handle the case safely. ## Detailed design - Help required here ## Impact on existing code This will break all switch statements over an enum imported from Objective-C or C where there is no default case. There should not be any automatic migration because it is appropriate for the developer to decide on the course of action in each case. ## Alternatives considered 1) No change This leaves the potential for undefined behaviour in all build modes if cases of bugs in code outside of Swift OR in the event that new cases are added to enums in future library versions. 2) Debug assertion This would have the benefit that developers might discover errors more quickly but could still leave risks where libraries (including OS frameworks) are updated and introduce bugs after application release or in untested configurations. ___ swift-evolution mailing list swift-evolution@swift.org https://lists.swift.org/mailman/listinfo/swift-evolution
Re: [swift-evolution] [Review] SE-0054: Abolish ImplicitlyUnwrappedOptional type
On 25/03/2016 15:24, Chris Lattner via swift-evolution wrote: Hello Swift community, The review of "Abolish ImplicitlyUnwrappedOptional type" begins now and runs through March 30th. The proposal is available here: https://github.com/apple/swift-evolution/blob/master/proposals/0054-abolish-iuo.md As I don't use IUO optionals I'm really fairly neutral on the proposal. I just wanted to mention with the powerful tools we have for dealing with optionals (optional, chaining, nil-coalescing, map, flatmap, guard, and let - thanks Chris and team) it really would be quite feasible to do away with IUOs completely. I don't think the community is ready for that yet so I'm not proposing that course of action at this time but such a proposal wouldn't get a universally hostile response. Even force unwrap is something that I only do test code since flatmap was added to the Standard Library (I used to have my own implementation of flatmap with a force unwrap after a filter which was the only production use I've put force unwrap to). Joseph ___ swift-evolution mailing list swift-evolution@swift.org https://lists.swift.org/mailman/listinfo/swift-evolution
Re: [swift-evolution] [swift-evolution-announce] [Review] SE-0053 Remove explicit use of let from Function Parameters
> On Mar 24, 2016, at 6:00 PM, Chris Lattnerwrote: > > The review of "Remove explicit use of let from Function Parameters" begins > now and runs through March 27th. The proposal is available here: > > > https://github.com/apple/swift-evolution/blob/master/proposals/0053-remove-let-from-function-parameters.md +1 seems obvious to me. I suspect the syntax is rarely used anyway. Joseph ___ swift-evolution mailing list swift-evolution@swift.org https://lists.swift.org/mailman/listinfo/swift-evolution
Re: [swift-evolution] [Review] SE-0016 - Adding initializers to Int and UInt to convert from UnsafePointer and UnsafeMutablePointer
> On Mar 22, 2016, at 9:35 PM, Chris Lattner via swift-evolution >wrote: > > Hello Swift community, > > The review of "Adding initializers to Int and UInt to convert from > UnsafePointer and UnsafeMutablePointer" begins now and runs through March > 25th. The proposal is available here: > > > https://github.com/apple/swift-evolution/blob/master/proposals/0016-initializers-for-converting-unsafe-pointers-to-ints.md I give the concept a +1. Swift should support dangerous actions but should make them clear and obvious. I do think that there is room to improve the argument label. There are two issues with "bitPattern" that concern me: 1) it should be clear that it will be the bitPattern of the pointer itself not whatever the pointer refers to. 2) it doesn't indicate the danger of using the results of this I don't have a great alternative to bitPattern but maybe something like: "rawMemoryAddress" I've read the proposal and given it a little bit of thought but have not yet read preceding or recent comments. Joseph ___ swift-evolution mailing list swift-evolution@swift.org https://lists.swift.org/mailman/listinfo/swift-evolution
Re: [swift-evolution] Asserts should not cause undefined behaviour
I gave a quick reply to an email later in the chain last night but I think these points are worth addressing. Apologies for the slow response, I wanted to ponder and consider the points rather than rush the response. >> On Dec 31, 2015, at 8:27 PM, Chris Lattner <clatt...@apple.com> wrote: >> >> On Dec 28, 2015, at 5:48 AM, Joseph Lord via swift-evolution >> <swift-evolution@swift.org> wrote: >> The documented behaviour of assert and assertionFailure in "disable safety >> checks" builds (still documented as -Ounchecked) is that the compiler "may >> assume that it would evaluate to true" or in the assertionFailure case never >> be called. > > Right. > >> This documented behaviour would allow the compiler to completely eliminate >> tests and branches before or after the assertion and take the operation deep >> into undefined behaviour. > > Only in cases where the assertion would have failed, right? The point of > -Ounchecked is that - if your code was correct with the checks - that it will > still be correct. Disabling overflow and array bounds checks is far more > dangerous than the assertion behavior you cite here. Yes only when it would have failed but that may still make correct code incorrect where whole branches of runtime belt and braces code could be a eliminated. >> It appears from the code as if the assumption is not currently applied on >> the assert method although it is on the assertionFailure case by means of >> the _conditionallyUnreachable() call. assert seems to be a no-op in both >> normal release and disable safety checks build modes. > > Right. > >> [Proposed Change] >> >> Change the documentation for assert and assertionFailure so that behaviour >> in unchecked mode is the same as in normal release - no evaluation and no >> effect. > > Why? : > >> 1) Expected behaviour based on other languages is for assert to have no >> effect in release. (If current behaviour is highly desired another function >> name should be used). Having potential dangerous behaviour from a function >> that is familiar across languages and is regarded as a safety feature is >> undesirable. > > This is the C model, but as you know, there is a whole field of custom > assertions libraries that people have developed. I don’t think there is > anything like consensus on this topic. C, Python, Erlang, Ocaml, Java (although it can enabled at runtime) and probably more. I recognise that in house assertion policies and systems may have different behaviours and agreed modes but I think affecting surrounding code (before and after the assertion) is surprising based on people experience with these other languages. I have not seen elsewhere any instances of compilers removing code outside of the branch itself which could be possible if the compiler is allowed to assume the truth of the assertion condition. if let names = jsonObject as? [String] where names.count > 0 { print(names.first) } else { assertionFailure("Invalid JSON") return nil } Could be compiled to pretty much an unconditional print of an address in unchecked. The nil return could be unreachable, the if let reduced to an unsafe bitcast (maybe not given array bridging but possibly in other cases) and the .first converted to [0] given the assumed passing of the where clause. >> 2) Adding asserts to code should not make the code more dangerous whatever >> the build. Assuming the truth of the assert may lead to runtime safety >> checks being skipped and undefined behaviour when a no-op would be a safe >> behaviour. > > This only affects code built with -Ounchecked, which is definitely not a safe > mode to build your code. The intention of this mode is that you can use it > to get a performance boost, if you believe your code to be sufficiently > tested. This mode, which isn’t the default in any way, intentionally takes > the guard rails off to get better performance. > > If you don’t like that model, don’t use this mode One fear is that libraries will be compiled in this way by people who aren't the original authors and that people won't realise the consequences. >> 3) "For highly robust code assert and then handle the error anyway" [Code >> Complete 2nd Edition section 8.2] > > Highly robust code shouldn’t build with -Ounchecked, so I don’t see how this > point is pertinent. The quote was external validation of the concept of having asserts where the case is also handled. I would not limit the use to "highly robust code". However as I said in my other email if no one else is concerned I will just go away and use my custom assertions and miss out on the compiler hinting. Joseph ___ swift-evolution mailing list swift-evolution@swift.org https://lists.swift.org/mailman/listinfo/swift-evolution
Re: [swift-evolution] Asserts should not cause undefined behaviour
> On Jan 2, 2016, at 5:39 PM, Dave Abrahams via swift-evolution > <swift-evolution@swift.org> wrote: > > >>> On Dec 31, 2015, at 1:56 PM, Dave Abrahams via swift-evolution >>> <swift-evolution@swift.org> wrote: >>> >>> >>> On Dec 31, 2015, at 12:27 PM, Chris Lattner via swift-evolution >>> <swift-evolution@swift.org> wrote: >>> >>> On Dec 28, 2015, at 5:48 AM, Joseph Lord via swift-evolution >>> <swift-evolution@swift.org> wrote: >>>> The documented behaviour of assert and assertionFailure in "disable safety >>>> checks" builds (still documented as -Ounchecked) is that the compiler "may >>>> assume that it would evaluate to true" or in the assertionFailure case >>>> never be called. >>> >>> Right. >>> >>>> This documented behaviour would allow the compiler to completely eliminate >>>> tests and branches before or after the assertion and take the operation >>>> deep into undefined behaviour. >>> >>> Only in cases where the assertion would have failed, right? The point of >>> -Ounchecked is that - if your code was correct with the checks - that it >>> will still be correct. Disabling overflow and array bounds checks is far >>> more dangerous than the assertion behavior you cite here. >>> >>>> It appears from the code as if the assumption is not currently applied on >>>> the assert method although it is on the assertionFailure case by means of >>>> the _conditionallyUnreachable() call. assert seems to be a no-op in both >>>> normal release and disable safety checks build modes. >>> >>> Right. >>> >>>> [Proposed Change] >>>> >>>> Change the documentation for assert and assertionFailure so that behaviour >>>> in unchecked mode is the same as in normal release - no evaluation and no >>>> effect. >>> >>> Why? : >>> >>>> 1) Expected behaviour based on other languages is for assert to have no >>>> effect in release. (If current behaviour is highly desired another >>>> function name should be used). Having potential dangerous behaviour from a >>>> function that is familiar across languages and is regarded as a safety >>>> feature is undesirable. >>> >>> This is the C model, but as you know, there is a whole field of custom >>> assertions libraries that people have developed. I don’t think there is >>> anything like consensus on this topic. >>> >>>> 2) Adding asserts to code should not make the code more dangerous whatever >>>> the build. Assuming the truth of the assert may lead to runtime safety >>>> checks being skipped and undefined behaviour when a no-op would be a safe >>>> behaviour. >>> >>> This only affects code built with -Ounchecked, which is definitely not a >>> safe mode to build your code. The intention of this mode is that you can >>> use it to get a performance boost, if you believe your code to be >>> sufficiently tested. This mode, which isn’t the default in any way, >>> intentionally takes the guard rails off to get better performance. >>> >>> If you don’t like that model, don’t use this mode. >> >> Let’s just consider -O; I think I understand Joseph’s objection here, and it >> seems valid. >> >> Normally in -O, we say that if you stay in the “safe subset” of Swift code, >> you never get undefined behavior, even if there’s a bug in your code. You >> might get *unpredictable* behavior of course, but presumably guaranteeing no >> undefined behavior rules out large classes of problems, including many >> security holes. Now suppose you decide to be responsible and add some >> asserts to help you catch bugs during development. Hopefully they help you >> catch all the bugs, but what if they don’t? All of a sudden, if you still >> have a bug when you ship, you now have undefined behavior. As much as I’m a >> fan of assertions having optimization benefits, It seems a little perverse >> that using them could make shipping code less secure. > > Wait a sec; I just read the doc comments for assert over again. They don’t > say there’s undefined behavior in -O if the condition isn’t satisfied. So > now I don’t understand what Joseph is complaining about. assert in -O is > documented to act exactly as C’s assert would with NDEBUG #defined. > > -Dave Than
[swift-evolution] Asserts should not cause undefined behaviour
I propose that assert and assertionFailure should be no-ops (with branch hints) in unchecked builds as they are in normal release builds rather than resulting in undefined behaviour in the failure condition. I would like to kick off a discussion of this. I found the proposal template useful for setting out my thoughts clearly to trigger the discussion but I'm not trying jump the discussion phase. I'm open to radical rewrites or to not proceed if there is opposition. Also if my understanding of the current code is wrong please let me know. [Current Situation] The documented behaviour of assert and assertionFailure in "disable safety checks" builds (still documented as -Ounchecked) is that the compiler "may assume that it would evaluate to true" or in the assertionFailure case never be called. This documented behaviour would allow the compiler to completely eliminate tests and branches before or after the assertion and take the operation deep into undefined behaviour. https://github.com/apple/swift/blob/cf8baedee2b09c9dd2d9c5519bf61629d1f6ebc8/stdlib/public/core/Assert.swift (latest commit to this file at time of writing) [NOTE - Actual current behaviour] It appears from the code as if the assumption is not currently applied on the assert method although it is on the assertionFailure case by means of the _conditionallyUnreachable() call. assert seems to be a no-op in both normal release and disable safety checks build modes. It also appears that precondition does not apply the permitted assumption when in _isFastAssertConfiguration mode. There also appears to be code in assert to hint the compiler that the assert will likely be true. This was something that I was planning to suggest and means that code containing asserts can be faster than the same code without the assert in release mode. [Proposed Change] Change the documentation for assert and assertionFailure so that behaviour in unchecked mode is the same as in normal release - no evaluation and no effect. Change the behaviour of assertionFailure to match the updated documentation. [Motivation] 1) Expected behaviour based on other languages is for assert to have no effect in release. (If current behaviour is highly desired another function name should be used). Having potential dangerous behaviour from a function that is familiar across languages and is regarded as a safety feature is undesirable. 2) Adding asserts to code should not make the code more dangerous whatever the build. Assuming the truth of the assert may lead to runtime safety checks being skipped and undefined behaviour when a no-op would be a safe behaviour. 3) "For highly robust code assert and then handle the error anyway" [Code Complete 2nd Edition section 8.2] While the designed expectation from the Swift team is to use them for checks against internal programming errors within a module in a number of cases I use assertions and assertionFailure calls while processing input data. The code without the assertion would simply fail to read the input gracefully but as the developer I want to know immediately if there are realistic cases that have not been handled so use assertions to flag it to myself that a case may need better handling. As such there are assertions in my code that are in error cases that I expect not to occur. I do not want the runtime safety checks being optimised out. [Being aware of this issue I use custom assertions but am therefore missing out on the branch hinting of the stdlib version and others may want similar behaviour and may not fully read the documentation]. [Impact on existing code] Loss of performance in cases where assertionFailure is used. Loss of potential performance improvement option in assert usage. In most cases this performance loss will be small but there is potential where the assumption of the value allows large code blocks (to evaluate the condition) to be eliminated for significant effects but I suspect that these cases are rare and preconditionFailure could be used instead to get the performance in unchecked builds (at the cost of release performance) or an additional method could be added. [Alternatives] Renaming assert so that the behaviour is not assumed by users familiar with other languages. Function name suggestion: assume/assumeUnreachable This could either behaviour as the current assert/assertionFailure documentation or possibly allow the assumption to be made in normal release mode not just the unchecked. It should be a fatal error in debug builds if assumption is incorrect. [Note about precondition] precondition/preconditionFailure also have undefined behaviour in unchecked mode but this is not a problem for me for a couple of reasons: 1) I don't have the same prior understanding about what they do. 2) It is clear from release build behaviour that hitting the condition is always an error.