theory | Abstraction layer for cross platform JavaScript | Reactive Programming library

You shouldn't consider this a production environment. The guides are examples, often for lab environments, and testing scenarios with the application. I would not consider them production ready, and compose is often not considered a production grade tool since everything it does is to a single docker node, where in production you typically want multiple nodes spread across multiple availability zones.

To directly answer your question, without further elaboration about how to break ties, two different implementations of Shannon-Fano could produce different codes of different lengths for the same inputs.

As @MattTimmermans noted in the comments, Shannon-Fano does not always produce optimal prefix-free codings the way that, say, Huffman coding does. It might therefore be helpful to think of it less as an algorithm and more of a heuristic - something that likely will produce a good code but isn't guaranteed to give an optimal solution. Many heuristics suffer from similar issues, where minor tweaks in the input or how ties are broken could result in different results. A good example of this is the greedy coloring algorithm for finding vertex colorings of graphs. The linked Wikipedia article includes an example in which changing the order in which nodes are visited by the same basic algorithm yields wildly different results.

Even algorithms that produce optimal results, however, can sometimes produce different optimal results based on tiebreaks. Take Huffman coding, for example, which works by repeatedly finding the two lowest-weight trees assembled so far and merging them together. In the event that there are three or more trees at some intermediary step that are all tied for the same weight, different implementations of Huffman coding could produce different prefix-free codes based on which two they join together. The resulting trees would all be equally "good," though, in that they'd all produce outputs of the same length. (That's largely because, unlike Shannon-Fano, Huffman coding is guaranteed to produce an optimal encoding.)

That being said, it's easy to adjust Shannon-Fano so that it always produces a consistent result. For example, you could say "in the event of a tie, choose the partition that puts fewer items into the top group," at which point you would always consistently produce the same coding. It wouldn't necessarily be an optimal encoding, but, then again, since Shannon-Fano was never guaranteed to do so, this is probably not a major concern.

If, on the other hand, you're interested in the question of "when Shannon-Fano has to break a tie, how do I decide how to break the tie to produce the optimal solution?," then I'm not sure of a way to do this other than recursively trying both options and seeing which one is better, which in the worst case leads to exponentially-slow runtimes. But perhaps someone else here can find a way to do that>

There is no bug and Delphi returns proper value in PixelsPerInch.

PPI OS will return for the purpose of scaling your application is not the actual PPI value of the actual display device, but virtual pixel density.

For developing applications the PPI value you need is the one that OS gives you, not the actual PPI value of the display device.

Everything you need to know is baseline PPI for the OS and current PPI or the scale factor. Using those numbers you can then calculate number of scaled pixels from some baseline pixel value.

For instance, if your control baseline width is 100 pixels and screen scale is 150% from the baseline PPI then your runtime control size will be 150 pixels.

Different operating systems have different baseline PPI.

Calculation:

The original code can be re-written in the following way:

The logic of comparing files seems alright, I don't find any outstanding problem with it, it is ok to prepend the "/" to match what you need. Could be even better if you could use the System.IO.Path.DirectorySeparatorChar for the directory root path as well, so if you run on windows or Linux you will have no issues.

But there may be a conceptual problem with what you are doing. To my understanding you aim to verify existence of specific configuration files required for your program to work right, if those files are missing than the program should fail. But that kind of failure due to missing configuration files, is an expected and valid result of your code. Yet, you unit-test this as if missing files should fail the test, as if missing files are an indication that something wrong with your code, this is wrong.

Missing files are not indication of your code not working correct and Unit-test should not be used as a validator to make sure the files exist prior executing the program, you will likely agree that unit-test is not part of the actual process and it should only aim to test your code and not preconditions, the test should compare an expected result (mock result of your code) vs. actual result and certainly not meant to become part of the code. That unit test looks like a validator that should be in the code.

So unless those files are produced by your specific code (and not the deployment) there is no sense testing that. In such case you need to create a configuration validator code - and your unit test could test that instead. So it will test that the validator expected result with a mock input you provide. But the thing here is that you would know that you only testing the validation logic and not the actual existence of the files.

When exactly does segmentation fault happen (=when is SIGSEGV sent)?

When you attempt to access memory you don’t have access to, such as accessing an array out of bounds or dereferencing an invalid pointer. The signal SIGSEGV is standardized but different OS might implement it differently. "Segmentation fault" is mainly a term used in *nix systems, Windows calls it "access violation".

Why is the process in undefined behavior state after that point?

Because one or several of the variables in the program didn’t behave as expected. Let’s say you have some array that is supposed to store a number of values, but you didn’t allocate enough room for all them. So only those you allocated room for get written correctly, and the rest written out of bounds of the array can hold any values. How exactly is the OS to know how critical those out of bounds values are for your application to function? It knows nothing of their purpose.

Furthermore, writing outside allowed memory can often corrupt other unrelated variables, which is obviously dangerous and can cause any random behavior. Such bugs are often hard to track down. Stack overflows for example are such segmentation faults prone to overwrite adjacent variables, unless the error was caught by protection mechanisms.

If we look at the behavior of "bare metal" microcontroller systems without any OS and no virtual memory features, just raw physical memory - they will just silently do exactly as told - for example, overwriting unrelated variables and keep on going. Which in turn could cause disastrous behavior in case the application is mission-critical.

Why is it not recoverable?

Because the OS doesn’t know what your program is supposed to be doing.

Though in the "bare metal" scenario above, the system might be smart enough to place itself in a safe mode and keep going. Critical applications such as automotive and med-tech aren’t allowed to just stop or reset, as that in itself might be dangerous. They will rather try to "limp home" with limited functionality.

Why does this solution avoid that unrecoverable state? Does it even?

That solution is just ignoring the error and keeps on going. It doesn’t fix the problem that caused it. It’s a very dirty patch and setjmp/longjmp in general are very dangerous functions that should be avoided for any purpose.

We have to realize that a segmentation fault is a symptom of a bug, not the cause.

It's not about the reference problem. And constexpr variable and constexpr function are different things. Quoted from the answer https://stackoverflow.com/a/31720324:

The reference does not have a preceding initialization from the point of view of i, though: It's a parameter. It's initialized once ByReference is called.

This is fine, since f does have preceding initialization. The initializer of f is a constant expression as well, since the implicitly declared default constructor is constexpr in this case (§12.1/5). Hence i is initialized by a constant expression and the invocation is a constant expression itself.

And about "preceding initialization", quoted from this:

It does mean "be initialized", but it's more importantly about the visibility of a preceding initialization within the context of the expression being evaluated. In your example, in the context of evaluating func(0) the compiler has the context to see the initialization of rf with 0. However, in the context of evaluating just the expression rf within func, it doesn't see an initialization of rf. The analysis is local, as in it doesn't analyze every call-site. This leads to expression rf itself within func not being a constant expression while func(0) is a constant expression.

Corresponding to your case, the line:

Is there a way to avoid this string coercion and use objects of type Buf or Blob ?

Unfortunately not at present. However, one can use the Latin-1 encoding, which gives a meaning to every byte, so any byte sequence will decode to it, and could then be matched using a grammar.

Also, I don't see any fundamental reason why regex should be reserved only to string, a NFA of automata theory isn't intriscally made for characters instead of bytes.

There isn't one; it's widely expected that the regex/grammar engine will be rebuilt at some point in the future (primarily to deal with performance limitations), and that would be a good point to also consider handling bytes and also codepoint level strings (Uni).

Recent SG9 discussion on LWG3564 concluded that the intended design is that x and as_const(x) should be required to be substitutable with equal results in equality-preserving expressions for which both are valid. In other words, they should be "equal" in the [concepts.equality] "same platonic value" sense. Thus, for instance, it is not valid for x and as_const(x) to have entirely different elements.

The exact wording and scope of the rule will have to await a paper, and we'll have to take care to avoid banning reasonable code. But certainly things like "x is a range of pairs but as_const(x) is a range of ints" are not within any reasonable definition of "reasonable".