Stefan Glienke

Your json is wrong - an empty array is represented as [], not as null

Nullable<T> you mean?

I wonder when that will happen given the oldest mention of that version I remember was in 2010

you must be using a version before 10.3 - this has been fixed: https://quality.embarcadero.com/browse/RSP-11302

I am also glad we could find a solution - fyi the actual fix was done a bit different as suggested in the comments of that issue. Also thanks to @jbg who gave some input on the subject and @Bruneau who we worked with to get this solved.

Some of your functions have a defect as in returning an empty array when no delimiter is found - they must return a 1 element array with the input string if they should follow RTL behavior. Also you can remove some unnecessary branching and make the code simpler: function CustomSplitWithPrecountByIndex2(const aString: string; const aDelimiter: Char): TArray<string>; var i, resultLen, resultIdx, tokenPos, inputLen, lastDelimiterPos: Integer; begin inputLen := aString.Length; lastDelimiterPos := 0; resultLen := 1; for i := 1 to inputLen do if aString[i] = aDelimiter then begin Inc(resultLen); lastDelimiterPos := i; end; SetLength(Result, resultLen); resultIdx := 0; tokenPos := 1; for i := 1 to lastDelimiterPos do if aString[i] = aDelimiter then begin SetString(Result[resultIdx], PChar(@aString[tokenPos]), i - tokenPos); tokenPos := i + 1; Inc(resultIdx); end; SetString(Result[resultIdx], PChar(@aString[tokenPos]), inputLen - lastDelimiterPos); end;

Should be an (*1)array of (*2)pointer to libvlc_media_track_t passed by (*3)reference, no?

Spring4D has ISet<T> and 2.0 will introduce IMultiSet<T>

Isn't it usually Sheldon who comes up with ridiculous stuff just to win an argument?

@Lajos Juhász Already reported as RSP-24383 @A.M. Hoornweg Might be RSP-19835 that got you.

I think the reason why people keep coming back is that there is uncertainty about when it can be assumed that the data is aligned properly. A local or global variable for example is not guaranteed to be naturally aligned.

100% this - any other program built with a programming language that directly is built ontop of winapi works as well - you only need to change applications in order to use and embrace new features or recommendations (such as not storing configuration or user specific data next to the binaries in programfiles)

Just tested with some integers (where sorting stability does not matter anyway): - for random data non generic code is slightly faster simply because they can use the comparison operators instead of doing rather costly comparer calls (even if the comparer itself is the default one) - approx 2 times - already sorted almost identical speed - reversed sorted data timsort is like 8 times faster

If you actually clicked on the link I provided and followed the other links provided in that answer by Peter Cordes (who is like the Jon Skeet on asm related stuff on SO) you will ultimately find a quote from the Intel documentation (and the AMD one agreeing on that) which states: Which is why I wrote that it works when target is aligned naturally which means by 64 bit - that in itself ensures that it does not cross a cache line. Now arguing that the presented solution does not work when this is not given is kinda moot. Thomas stated that he wants to "set a double variable in a thread safe manner" - which I assume is meant literally and a double variable will be 64-bit aligned.

movq xmm0,[value] movq [target],xmm0 should do the trick when target is naturally aligned - for reference.

The lack of pdb support in Delphi makes it tedious to use because you only get addresses reported which you then have to manually look up.

Just pointing this out so you don't get yourself into trouble: IANAL but the fact that the IPP is under a commercial license or a free license if you qualify (time limited if I read it correctly - but I just quickly skimmed through it) might make it arguable to actually distribute any parts of it. If you know more about it I would be glad to be wrong. And as said before - even though your intentions are surely to make it easy for users - putting the source for those projects with an explanation how to build them would get you on the safe side.

Why isn't the code of those dlls open source as well? If they are simply directly taken from those Intel libraries, provide a link how to get them directly from the original source.

I would like to add: "because good compilers can turn such code into the most efficient code."

Notice that I wrote "efficient code" - it's not only about applications running fast (enough) but also energy efficient which most people tend to forget.

Then you are almost at an Introsort (Spring4D uses that instead of plain Quicksort)

Always understand the reason behind guidelines, suggestions, principles: why is it a good thing to do x - not just because famous person y said so. Because when you understand why something is a good or bad thing to do you are able to evaluate the situations where you want or dont want to follow them and what implications arise.

The question stands - there is no "yes/no" about that - their reference count is threadsafe - an assignment is not because its not atomic (which is the very same issue that Dalija wrote about in her blog post). About this discussion I think it is nonsense - the issue being pointed out by Marco arises when you access something with a broader scope than where you are currently - i.e. a global variable or a variable outside of the current nested routine - that also includes passing the same variable by reference twice or more and modifying it. f(i,i) would also be defect if it was implemented like this: while a > 0 do begin Inc(b); Dec(a); end; "Better" for strings would raise the question: better for what? Memory consumption? UTF8, string interning. Multithreading? Immutability. Memory locality? Short string optimization (i.e. avoiding the memory indirection) - there are probably more and most of them would be a complete breaking change even more than Delphi 2009.

What is thread safety anyway?

I uninstalled SourceTree (which I had been using for years before with decreasing joy) the day I started using Fork and never looked back - back then Fork was completely free and changed to a paid license later but those 50 bucks were very well spent.