I don't have much sympathy for people who were doing things like writing multithreaded programs in the days before C documented its memory model and then becoming unhappy because new optimisations that legitimately help single-threaded code broke their programs.

In my experience C compiler maintainers have generally been open to the idea of offering guarantees beyond a narrow reading of the standard, but they want to be able to clearly state what it is that they're guaranteeing. "Keep old programs running" isn't enough.

I think the "Prevailing interpretation" that Yodaiken complains about is coming from the same place as suspicion of the "be lenient in what you accept" IETF principle: that sort of thing doesn't lead to robustness in the long run.

The way forward at this point is surely to define more things that are currently undefined (whether in the standard or by widely-implemented extensions).

No. It's fine (arguably desireable, but reasonable people might disagree) for implementations that encounter undefined behaviour to terminate execution immediately, especially if it's with a error message. The problem is when implementations silently willfully misinterpret what they accept, particularly in ways that cause (not "expose"; reading from a null pointer and discarding the value (for example) isn't) security vulnerablities.

The inflamed backlash should tell you just how damaging it is to impose silent failure on meticulously written, previously fine programs.

With relatively few exceptions, if your program hits undefined behavior, then your program was already doing something pretty wrong to begin with. Signed overflow is a poignant example: in how many contexts is INT_MAX + 1 overflowing to INT_MIN actually sane semantics? Unless you're immediately attempting to check the result to see if it overflowed (which is extremely rare in the code I see), this overflow is almost certain to be unexpected, and a program which would have permitted this was not "previously fine" nor "meticulously written."

I feel compelled right now to point out that software development is a field where it is routine to tell users that it's their fault for expecting our products to work (that's what the big all-caps block of every software license and EULA says, translated into simple English).

The code was written in Pascal and assembly, though, so it was safe from a C compiler.

As such my vision of low level coding isn't tainted by the ways of C.

It's not that rare - I know that postgres got bit by particularly that issue, and several other other projects as well. Particularly painful because that obviously can cause security issues.

    int a;
    // lots of code
    int b = a + 1;
    // check for overflow
    if (b <= a)
        abort();

In that case, something like: [...] if (a > INT_MAX - 1) abort(); int b = a + 1; [...]

Much easier said than done:

https://github.com/postgres/postgres/blob/master/src/include...

Yeah, a standard library function that does this would be good. But many people would just use * instead of this function, and so the problem would partially remain.

In c++ it would be a little simpler due to templates, so the types of the arguments to the function can be derived. But the type of the result can still confuse programmers. Although maybe it's not so bad, because an overflow that happens during multiplication (or other math operations) is undefined-behavior, but an overflow that happens during assignment of the multiplication result to a variable that can't hold it is only implementation-defined behavior, not undefined behavior.

The author claims: “We have the absurd situation that C, specifically constructed to write the UNIX kernel, cannot be used to write operating systems. In fact, Linux and other operating systems are written in an unstable dialect of C that is produced by using a number of special flags that turn off compiler transformations based on undefined behavior (with no guarantees about future “optimizations”). The Postgres database also needs some of these flags as does the libsodium encryption library and even the machine learning tensor-flow package.”

So basically the programers of the most used C programs consider the C standard so broken that they force the compiler to deviate from it. (Or they are not able to do things right?)

Standard is just common denominator agreed by actual competitors. (Standard takes into account the chipset where incrementing max_int produces a beep instead of min_int). If user wants to use more advantages of the product (dbms/compiler/hardware), one must sacrifice portability and use cnonstandard extensions.

Well, I learned of the change in compiler behavior some years back because I had written loop code with a sanity check which depended on signed integer overflow wrapping, along with a test case to prove that the sanity check worked, and that test case started failing:

   not ok 2 - catch overflow in token position calculation

   Failed test 'catch overflow in token position calculation'
   at t/152-inversion.t line 70.

To be fair, as several people and TFA have pointed out, this isn't a problem with C, but with defective/malicous C compilers. Admittedly, that's not much help if you can't find a compiler that isn't defective/malicous, though, so I can only wish you the best of luck.

Do you want compilers to stop adding optimizations while staying within the bounds defined by the spec? That they somehow guess that a given piece of code that may trigger UB is too important for them to optimize it based on the assumption that the developer knew what she was doing and ensured that it wouldn't?

The case of a compiler update breaking the test sounds like desirable to me. It pinpoints a critical piece of code that was relying on a specific implementation's behavior that is not specified. This could have been triggered by a compiler or architecture change. If this is something that is a hassle to fix immediately, you can temporarily downgrade to the version of the compiler used earlier or disable some optimizations.

I fail to see why one would consider a compiler evolving while conforming to language specification defective or malicious (however, with that definition I fear that finding one that isn't may indeed be difficult).

It would be fine if I could opt into new semantics — something like Rust's "editions" would resolve my objections about these compiler optimizations. But that doesn't seem to be on offer in the C ecosystem.

C culture and to certain extent Objective-C and C++ ones are tainted by microptimizaitons while typing, where the compilers are the worst examples.

Unfortunely UNIX and C go together, so those that want to keep UNIX like platforms around bettter fix C somehow.

Yet Azure Sphere, despite its security message, uses C only SDK.

Meanwhile Visual Studio now supports C11 and C17.

Market pressure, their customers weren't willing to buy into it, and the new Microsoft also wants all those POSIX FOSS packages written in C running on Windows.

Hopefully the newer compiled languages will kill C++.

Quoting from another of my comments:

> > [What are you objecting to?]

> Inferring any propositional statement about the program (eg "this pointer is not null") from the fact that its negation would imply undefined behaviour.

That is what the problem is. Undefined behaviour is a licence to implement operations without regard for unusual corner cases, not to infer the absence of said corner cases from those operations and then apply that 'knowledge' elsewhere.

> I fail to see why one would consider a compiler evolving while conforming to language specification defective or malicious

It's https://en.wikipedia.org/wiki/Malicious_compliance in near-textbook form.

I'm not completely sure I understand that correctly, but do you mean that statements that are constant unless considering a possible (and "credible"?) implementation of UB shouldn't be fair-game for the compiler to optimize out? EDIT: I think I see a more tricky case that may be one of those you're referring to. Dereferencing a pointer further in the code shouldn't be a valid justification for optimizing out previous tests of it being null. I can relate with that but I suspect that it would prevent many classes of branch pruning.

I get what you suggest while describing malicious compliance but I can imagine that it could be a false impression resulting from trade-offs that favor optimization opportunities to "out-of-spec but canonical/natural" implementations.

Not further. Anywhere. If the implementation wishes to rewrite pointer dereferences from `use(*p)` to:

  if(!p) abort();
  use(*p);

  use(*p);
  if(!p) return NOPE;

  if(!p) abort();
  use(*p);
  //if(!p) return NOPE; // unreachable because of if, not because of use

  use(*p);
  //if(!p) return NOPE; // CVE-20XX-XXXXX

> it would prevent many classes of branch pruning.

Yes, that's the general idea.

It dereferences a null pointer, invoking undefined behaviour, then calls the function `use` with the resulting value.

> what should the compiler emit?

Probably something to the effect of:

  ld r0 [sp+.p]  # if p is not already in a register
  ld r0 [r0]  # *p
  jsr use

  ld r0 [sp+.p]  # if p is not already in a register
  jz r0 .panic
  ld r0 [r0]  # *p
  jsr use

Conforming to the spec is not a virtue. We want the compiler to be reasonable, regardless of whether the spec is. When the spec is malicious, conforming to the spec is malicious behavior.

For example, the Java spec says that the << operator (bit shift left) will accept any right operand, but performs a modulo-32 operation on the right operand before doing the shift. So `a << 4` is `a` shifted 4 places left, `a << 20` is `a` shifted 20 places left, and `a << 36` is `a` shifted 4 places left again.

This is absurd, and I'm comfortable calling it a bug in the spec. `a << 40` needs to have 0 in the lowest 40 bits. It does not need to have random values in bits 8-31.

This behavior is documented, but that doesn't make things better, it makes them worse.

But the philosophy that says "if it's documented, then it's OK" doesn't even allow for the concept of a bug in the spec.

It had code like:

    int *p;
    // lots of code
    
    if (p != NULL)
        return 1;
    // use p

    int *p;
    // lots of code
    
    int a = *p;
    if (p != NULL)
        return 1;
    // use p

I think in general the "sanity check" cases are the worst. It is hard to determine whether an expression causes undefined behavior if you cannot try and evaluate it. Perhaps a compiler intrinsic that checks (at runtime) whether an expression causes undefined behavior could be useful here. Though I can imagine such an intrinsic being essentially impossible to implement.

In user code GCC will still happily remove the null check because in user code this is an actual bug in the user's code.

  struct foo { ...; bar_t bar[NBAR]; };
  struct foo* p = ...;
  bar_t* q = &p->bar[0]; // add rq, rp, #foo_bar_offs
  // other declarations
  
  if(!p) return NOPE; // optimized out
  // use p and q

https://lwn.net/Articles/342420/

There's some disagreement on whether you can call a program "fine" that breaks after switching to a newer version, or a different compiler.

I see a lot of programmers out there that unfortunately use the behavior of their code on whatever compiler they're using at the moment as a proxy for what the language actually guarantees.

a = 5 //* junk here */ 2 ;

I'm sure there are ioccc or underhanded C contest entries doing this to make code work differently on compilers based on if // is starting a comment line or not.

Sure it is an ugly way of writing stuff and you'd be hard pressed to find lots of real world traps like this, but when/if you did have code that "suddenly" miscompiles you might actually think your old code with an old compiler did work, and a new compiler for "the same" language breaks your program. I don't think everyone code base should need full rewrites ever time a new compiler comes out.

Yes, that means C without optimizing without UB is slower than Java. Optimizations need some form of reasoning about what's happening. For Java it's optimizing based on guarantees provided by the language (there's no raw pointers that could mess with the things listed above). But C doesn't provide any hard guarantees, so instead it needs to blindly assume that the code will behave sanely.

Also note that for many of the more manageable sources of UB, most compilers provide a choice (-fwrapv, -fno-strict-aliasing, ...). Yet few projects use these options, even when they use other gcc/clang-specific features. Doesn't that indicate that C programmers indeed want to sacrifice safety for optimizations?

For example there were programs 30-40 years ago that relied on exact stack layouts. These days everybody would agree they are completely broken.

The issue of course is that it is extremely hard to write programs that have no UB. It would be nice for compilers to have an option to automatically introduce assetions whenever they rely on some UB-derived axiom, basically as a sort of lightweight sanitizer.

In fact if we had sanitizers 30-40 years ago probably things would be better today.

Modifying a value from a different thread without synchronization is UB. The compiler assumes this does not happen in order to e.g. move things into registers. Could you elaborate how (and how often) you would like to have this kind of UB-derived axiom ("this value remains the same from here to there") checked with assertions?

https://www.parasoft.com/products/parasoft-insure/

21 years later it is still an uphill battle to adopt such technology in C and C++ projects.

"Although the first edition of K&R described most of the rules that brought C's type structure to its present form, many programs written in the older, more relaxed style persisted, and so did compilers that tolerated it. To encourage people to pay more attention to the official language rules, to detect legal but suspicious constructions, and to help find interface mismatches undetectable with simple mechanisms for separate compilation, Steve Johnson adapted his pcc compiler to produce lint [Johnson 79b], which scanned a set of files and remarked on dubious constructions. "

-- https://www.bell-labs.com/usr/dmr/www/chist.html

That seems to be a static analysis tool though (which generally have not been great). Did it also inject runtime checks?

And a famous commercial variant of it has been PC-lint from https://www.gimpel.com/.

Being available on open source compilers does little to change the culture, as per latest surveys only 11% of developers care to use any kind of tooling for improving their code quality in C and C++.

At CppCon a couple of years ago, only about 1% of the audience answered positively to Herb Sutters' question.

https://isocpp.org/files/papers/CppDevSurvey-2021-04-summary...

Of course compilers have to make assumptions. The debate is (a) over what assumptions it is proper to make and (b) what are the permissible behaviors. The false dichotomy: either do without any optimizations at all or accept whatever UB gives you, is not a useful approach.

Inferring any propositional statement about the program (eg "this pointer is not null") from the fact that its negation would imply undefined behaviour.

Even if you opt into those (and the projects I've been involved with have), the precedent is established: when new compiler optimizations are introduced with disruptive semantics, they are opt-out, not opt-in — a fail-dangerous failure mode.

> Doesn't that indicate that C programmers indeed want to sacrifice safety for optimizations?

Maybe so. Which is one reason I wouldn't call myself a "C programmer" any more. The demands to program responsibly in C are absurdly high.

Actually, the default for gcc is -O0. You are opting in with -O2 etc.

BCPL was anyway designed to Bootstrap CPL, nothing else.

That’s just not true. Already examples in this thread: https://news.ycombinator.com/item?id=27223870

What about backwards compatibility in regards to a new compiler version deciding to issue errors on UB now? You don't have any guarantees about what happens with UB right now, so if you upgrade to a new version compiler that generates errors instead of "optimizations" everything would be still as before: no guarantees. And it's frankly a lot better to blow up the compilation with errors than to have the compiler accept the UB code and roll a dice on how the final binary will behave later. You can either fix the code to make it compile again, or use an older "known good" version of the compiler that you previously used as a stopgap measure.

I fail to see any reason whatsoever why compilers are still doing all kinds of stupid stuff with UB instead of doing the right thing and issuing errors when they encounter UB.

I also fail to see why the C language designers still insist on keeping so much of the legacy shit around.

The compiler assumes UB will never happen and it makes transformations that will be valid if there happens to be no UB. This doesn't require any explicit detection of UB, and in some cases UB or not is simply undecidable at compile time (as in no compiler could detect it without incorrect results).

Without these assumptions the resulting compiled code would be much slower, though some optimizations have different danger vs speed impact and there certainly can be a case that there are some optimizations that should be eschewed because they're a poor trade-off.

There are many cases where current compilers will warn you when you've done something that is UB. It's probably not the case that they warn for every such detectable case and if so it would be reasonable to ask them to warn about more of them.

I think your irritation is just based on a misunderstanding of the situation.

Compiler authors are C(++) programmers too, they also don't like footguns. They're not trying to screw anyone over. They don't waste their time adding optimizations that don't make real performance improvements just to trip up invalid code.

The main reason to not spec these things is because people would be arguing "this makes compiled code on my esoteric 9-bit 1-complement chip slower" or "there was this chip in the 70s that did things differently" or "but a short int on Cray was 64-bit". Great, so now the spec has avoidable unnecessary undefined behavior all over the place, and the code other people wrote still does not run correctly on your 9-bit chip. Brought to you by the same people who decided "NULL is not necessarily (void*)0", and who define those integer types everybody uses (instead of stdint) with an "at least this big".

Yes, a lot of that is legacy stuff and was added to accommodate and model things that already existed (the wrong way to go about it, IMO, but hindsight is 20/20), but that's my argument: fix this stuff once and for all and for good in an upcoming spec iteration.

>Without these assumptions the resulting compiled code would be much slower

In some cases, this is true (for different levels of "much slower"), but the trade off here is still "running code that works, but a little slower" vs "running code that does not work and will launch a nuclear strike at Switzerland by accident, but really fast".

In a lot of cases, it will not be slower, or at least not much slower.

>I think your irritation is just based on a misunderstanding of the situation.

Frankly, not really. I started writing my first C (and C++) in the early 90s, and I think I do understand the situation pretty well by now. But I should have been more precise in my initial ranting comment, I give you that.

>They're not trying to screw anyone over.

I didn't say that they are.

But, to address the content of your comment: defined behavior at runtime is not necessarily good behavior at runtime. Defining signed integer overflow to wrap, for example, is probably a bad idea, because this is rarely the intent of the code. Having all such operations trap might be a good idea, but now you're going to get the same "stop breaking my working programs" people angry at you.

>Defining signed integer overflow to wrap, for example, is probably a bad idea

I wouldn't call it great behavior, but it's at least what most people expect will happen, and most people will be able to understand what's going on, and it's fast on most systems that matter. However, it's still undefined behavior. Just codifying overflows to be wrapping, would therefore be an improvement in my opinion, at least over what we have today.

Also note that this is distinct from the memory representation of (void *)0 being all 0 bits, which is explicitly not mandated.

That's not how compwillrs work. In fact in the general case it is impossible to figure out at compile time that "there is an UB".

The compiler instead assumes as an axiom that no UB can ever happen and uses the axioms to prove properties of the code.

These days if you want to catch UB, compile with -fsanitize=undefined-behaviour. The program wll then trap if UB is actually detected at runtime.

So, let me get this straight, someone wants to make sure pointer p is not null (in the wrong way), and codes something like the examples in posts above like if (!p) ... and if that doesn't trigger calls use(*p), but compiler decides p can never be null because that would be UB and hence removes the check.

The C coder dumps the code and gets upset because the check is removed and gets the hint to catch UB by adding -fsanitize .. that "catches UB" in the above scenario so that the program will "trap if UB is detected".

I think we just came full circle there.

Sure, the -f will catch ALL detected bugs and so on, but I still found it a bit funny.

Ubsan will abort an invalid program if it detect ub. It doesn't let you handle it. So you shouldn't remove the erroneous check, but fix it so it is no longer erroneous, and ubsan will help you identify these errors.

Also ubsan adds significant overhead so it is not really appropriate for production builds unfortunately (hence my wish for a less powerful ubsan-lite but with lower overhead).

    if (!p) {
        use(*p);
    }

What people are complaining about is the opposite case:

    use(*p);

    /* The compiler reasons that if p == NULL, the program would have crashed by now,
       so if we got here, p != NULL must hold. */

    if (!p) {  // the compiler can remove this branch
        report_error();
    }

> The set of people doing so overlaps with the set of people complaining that the compiler doesn't optimize their code sufficiently to a significant degree.

There's no contradiction here, and the overlap is generally just "people who care". The optimizations that are not safe shouldn't exist, and the optimizations that are safe should be good.

> nearly all of the time are things the code author would agree with if they thought long and hard

I highly doubt this is the case for even one situation.

The whole point is that there isn't such adversarial thing like "we're going to find the UB right there, won't even print a warning about it, and mess up your crappy code, haha!"

Optimizers aren't reasoning about code like people do (start to finish with high-level understanding of the whole function), but rather as series of mostly dumb, mostly isolated small passes, each pass changing one little thing about the code.

It just happens that one pass marks certain instructions as "can't happen" (like the spec says), then another pass simplifies expressions, and then another pass that removes code that doesn't do anything, usually left over from the previous steps. They sometimes combine in an "adversarial" way, but individually each pass is justified and necessary.

Compilers already have lots of different passes. Splitting optimizations into passes is a way to keep complexity closer to O(n) rather than O(n^2), but this architecture makes interactions between passes very delicate and difficult to coordinate, so it's difficult to instrument the data to avoid only cases of annoying UB without pessimizing cases that users want optimized.

All code transformations have a compile time cost and runtime perf impact. We don't add transformations unless the runtime perf impact greatly outweighs the compile time cost.

These optimizations are added because they measurably improve the performance of real code. This comes up in every review for new or updated optimization passes. This claim that they aren't justified is actually rather insulting to the effort put in to improve perf without taking days to compile.