Rendered at 19:19:52 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
alankarmisra 1 days ago [-]
Ooo I love these things. I wrote a C++ -> Clang -> JS transpiler so I could show value change states on top of the C++ source code in html and allow people to step through the whole state change. They can also vary the input and see the values change cause the JS layer could recompute the values - which is why I bothered with the transpiling - otherwise I could have just output the state changes and parsed them as CSV for the steps. Fun project!
I wish the readme example showed a lambda capture. Now that's where compiler magic REALLY shows up.
sltkr 23 hours ago [-]
C++ lambdas are just structs that overload operator(); once you realize that they become a lot less mysterious. But it's a great example of how this tool is useful to learn about those things (of course you could also read the C++ standard but who has time for that?)
(That also shows how C++'s range-based for-loops work, which is similarly some syntactic suger for a fairly mundane for-loop, but details are interesting: unlike a typical for-loop, the loop variable is declared inside the loop body.)
StilesCrisis 21 hours ago [-]
I said shows a lambda _capture_. This involves the compiler automatically determining which variables cross the scope boundary, and injecting code for each one.
frabert 4 hours ago [-]
I'm curious to understand which part you think is particularly interesting. The compiler clearly knows which variables are declared in what scope, so what's mysterious about figuring out which values need to be captured?
Someone 13 hours ago [-]
> That also shows how C++'s range-based for-loops work, which is similarly some syntactic suger for a fairly mundane for-loop, but details are interesting: unlike a typical for-loop, the loop variable is declared inside the loop body
That’s not true in C++, is it? There, this does not compile
int foo(int i) {
int add(int x) { return i + x; } // nested function
return add(3);
}
int bar(int i {
alias lambda = (x) => x + i; // lambda
return lambda(3);
}
}
I must be the only one who (while I like the tool) feel a little strange, somewhere between sad, angry, annoyed and disappointed (can’t really express it), that a language needs (or at least is of good use) such a tool.
Like: I wish what is the first interpretation of my code to be obvious. Is not about advanced optimization, or nitty gritty compiler tricks, is about the language. This tool is useful in any C++ compiler. How did we end with such a complex language?
I mean (sorry, very personal take) the programming language is for me a tool to help me think about a problem and its solution. In this era of LLM more than before, is not a tool to get the machine to do something, is a language for formalizing ideas, problems and solutions, like math. Should not be so obscure, that you need to be a lawyer to understand it, right? Or I am very much old and totally in the weeds with my ideas?
jcelerier 7 hours ago [-]
Which language do you think doesn't need it? For instance for me I wish I had such a tool whenever I used any interpreted or JIT language, to understand clearly to which semantics whatever is happening maps and how their respective front-end optimizes. Things as simple as "a" + "b" in JavaScript: are you sure you know how every js runtime maps it? Do they fold the strings as an optimization during parsing ? Do they preallocate a larger native string type and push? Do they just do something that ends up an equivalent of std::u16string("a") + std:: u16string("b")?
kccqzy 1 days ago [-]
No a language doesn’t need it. It’s an educational tool for those new to the language or new to a particular language feature. With practice language features just become automatic so you don’t think about them when thinking about a problem.
In the README example, the C++ practitioner just doesn’t think about those automatically synthesized copy constructors and such for POD types, nor do they think about the fact that assignment is desugared to operator= or that an assignment can involve a static_cast.
f1shy 1 days ago [-]
>> needs (or at least is of good use) such a tool.
Of course there are people who do not need that tool. The thing is, other languages doesn’t need such an explanation tool at all, is very much enough with reading a small book. I personally would like C++ to not really need it, but clearly as I stated: maybe “the only one”
kccqzy 23 hours ago [-]
> other languages doesn’t need such an explanation tool at all
Plain old C has -E to expand macros. Rust has cargo expand. Haskell has -ddump-simpl. Common Lisp has macroexpand. Any language that has a complicated syntax or allows the user to customize the syntax needs such a tool. Whether such a tool actually exists reflects more on popularity and community desire.
drysine 1 days ago [-]
>other languages doesn’t need such an explanation tool at all
Are you sure about that?
esikich 1 days ago [-]
In what language is compiling simple? Do you not like compiled languages? I'm not sure what you're complaining about here.
f1shy 1 days ago [-]
Lisp is simple to compile. Even C (at the level this tool works anyway), also basic, and maybe other 100 languages (again, not the whole compiler, but what this tool shows).
Compilation per se is not difficult. Doing it with good error recovery, good feedback in case of error, and the optimization is difficult.
This tool does not tell me those parts (for that is gotbolt) this is about how the first layer of interpretation works, and not for making compilers, but for understanding what the code actually means, because is not obvious at first sight.
dasyatidprime 1 days ago [-]
I agree with an overall thrust of “C++ is overcomplicated and has many moving parts and a lot of implicit stuff happening in the background that makes it difficult to understand”. But I do want to briefly note that Lisp environments have macro expansion steppers for a reason too.
pjmlp 1 days ago [-]
Classical Lisp 1.5 yeah, modern Lisps not so much, hence why there are famous publications like Lisp In Small Pieces.
Only if you mean K&R C or the early C89/90 compilers, before the extension explosion, newer language standards, and modern ISAs.
rramadass 1 days ago [-]
You have misunderstood the language and the tool.
Compilers for all languages have the concept of "lowering" the level of source-language abstractions from high-level to low-level viz. Written Source -> Simpler Source -> AST -> IR -> Machine Code. One type is "desugaring" i.e. transforming complex syntax to simpler ones so that the following stages of compilation have to operate only on a smaller and hence more manageable language subset.
This tool shows both desugaring and other lowering (eg. implicit transformations, type deductions etc.) within the C++ source language by regenerating the source from the AST thus giving you full visibility until that stage in the compilation process.
I don't think it's only C++ though, Python decorators, Rust macros, almost every language's compiler write code you never see. C++ just does more of it and makes it harder to guess. Being able to look at the expanded version is useful in any language that does this.
veexx103 18 hours ago [-]
I’ve worked on something similar, but the code reconstructed from the Clang AST fails to compile due to various issues.
Can these issues be resolved?
kvemkon 22 hours ago [-]
> See your source code ...
What I'd like to see, is how much time is spent to compile each block of C++ code (sliced as fine as possible).
kizxc4395 13 hours ago [-]
This is so interesting
IshKebab 1 days ago [-]
This is what let me understand lambdas. Great tool.
drysine 1 days ago [-]
Good for C++'s coroutines too
pjmlp 1 days ago [-]
It also helps to know the .NET model, as C++20 co-routines were inspired by how they work and related magic methods on awaitable types.
tweedler290 1 days ago [-]
thats crazy. i made this https://github.com/oooscoos/Benzi which does the same thing but is language agnostic (supports 9 other languages). urs probs goes in deeper into c++ than mine tho
abbefaria27 1 days ago [-]
This is very different and doesn’t relate to what the linked repo does.
https://youtu.be/ZPGysUX62OY
https://cppinsights.io/
Here's an example: https://cppinsights.io/s/f04b6896
(That also shows how C++'s range-based for-loops work, which is similarly some syntactic suger for a fairly mundane for-loop, but details are interesting: unlike a typical for-loop, the loop variable is declared inside the loop body.)
That’s not true in C++, is it? There, this does not compile
(https://cppinsights.io/s/26405687: error: use of undeclared identifier 'i')https://en.cppreference.com/cpp/language/for:
“A for statement equivalent to:
Except that […]”and
“While in C names declared in the scope of init-statement and condition can be shadowed in the scope of statement, it is forbidden in C++:
“Like: I wish what is the first interpretation of my code to be obvious. Is not about advanced optimization, or nitty gritty compiler tricks, is about the language. This tool is useful in any C++ compiler. How did we end with such a complex language?
I mean (sorry, very personal take) the programming language is for me a tool to help me think about a problem and its solution. In this era of LLM more than before, is not a tool to get the machine to do something, is a language for formalizing ideas, problems and solutions, like math. Should not be so obscure, that you need to be a lawyer to understand it, right? Or I am very much old and totally in the weeds with my ideas?
In the README example, the C++ practitioner just doesn’t think about those automatically synthesized copy constructors and such for POD types, nor do they think about the fact that assignment is desugared to operator= or that an assignment can involve a static_cast.
Of course there are people who do not need that tool. The thing is, other languages doesn’t need such an explanation tool at all, is very much enough with reading a small book. I personally would like C++ to not really need it, but clearly as I stated: maybe “the only one”
Plain old C has -E to expand macros. Rust has cargo expand. Haskell has -ddump-simpl. Common Lisp has macroexpand. Any language that has a complicated syntax or allows the user to customize the syntax needs such a tool. Whether such a tool actually exists reflects more on popularity and community desire.
Are you sure about that?
Compilation per se is not difficult. Doing it with good error recovery, good feedback in case of error, and the optimization is difficult.
This tool does not tell me those parts (for that is gotbolt) this is about how the first layer of interpretation works, and not for making compilers, but for understanding what the code actually means, because is not obvious at first sight.
Only if you mean K&R C or the early C89/90 compilers, before the extension explosion, newer language standards, and modern ISAs.
Compilers for all languages have the concept of "lowering" the level of source-language abstractions from high-level to low-level viz. Written Source -> Simpler Source -> AST -> IR -> Machine Code. One type is "desugaring" i.e. transforming complex syntax to simpler ones so that the following stages of compilation have to operate only on a smaller and hence more manageable language subset.
This tool shows both desugaring and other lowering (eg. implicit transformations, type deductions etc.) within the C++ source language by regenerating the source from the AST thus giving you full visibility until that stage in the compilation process.
PS: Explanatory video presentation by the author - https://www.youtube.com/watch?v=VJ6ZvDRYzNE&t=558s
Can these issues be resolved?
What I'd like to see, is how much time is spent to compile each block of C++ code (sliced as fine as possible).