Settings

Theme

Clice v1.0 Released: Welcome to Modern C++

ykiko.me

191 points by molenzwiebel · 111 comments

Reader

15 threads
assumed_throwaw

I discovered Clice a few months ago, but it wasn't stable enough for my needs at the time. It looked very promising, though. I'm glad to hear it's moving to v1.0.

I didn't expect this technical article to end on such a melancholy note. I'm sad that the author feels this way about his work of last years. But it was an interesting read, and I think many of us share these concerns and doubts. Hopefully, even in a world of AI agents, there will still be people that write and read code, and who will enjoy using Clice.

tancop

> in an era like this it means the effort I poured into clice over the past two and a half years is unlikely to have the lasting meaning I first imagined

Reading C++ without a good language server is hard. And if you don't read your agents code that makes you a stupid vibe coder. Everyone else will benefit from your project.

And it's good training data (also for humans) to learn modern C++ from a real codebase.

quailfarmer

Re the melancholy conclusion: I spent the last year of my school working on a big personal project, but when I started a full time job, this personal project was eclipsed, it no longer made sense in the context of the industry. But I recently returned for my 5yr reunion, and I found a bunch of current students who were still using my system! Upon reflection, it’s still one of the most fun projects I’ve ever done.

anonymous908213

> Over on X, you can see plenty of people asking Claude in a single sentence to rewrite the TypeScript compiler in another programming language and getting better performance out of it.

And yet nobody is using these rewritten Typescript compilers to compile real programs, because LLM-generated compilers fall apart the moment you move beyond whatever test programs the LLM was special-casing to work for impressive benchmarks. Drawing the conclusion that software dev has been 100x'd because people can post stupid attention-grabbing shit on HN and Twitter, when none of that software is actually used by anybody, is an extremely unfortunate conclusion to draw. And so, facing peer pressure and FOMO, the author abandoned their own project to the bots and allowed it to be slopped out to the point nobody will ever seriously use it. 100x, but at what cost? The damage being done to our community will take decades to recover from at this rate. We have turned the concept of software development into "who can make the most impressive landing page and headline" with no concern for making anything that is useful to anyone - not even to the author, for they are no longer learning.

rixed

Going on a tangent, but I'm always owndering if "modern C++" sounds rather appealing or off-putting to most people?

  • moltopoco

    Just one data point, because I have to get it off my heart. I loved Modern C++ when it meant C++11, but I've become more negative with each new language standard reaching my workplace.

    Using string_view over string by default is probably the #1 cause of memory errors that landed on my desk. I also keep hearing that "int x{0}" is preferable to "int x = 0" which allows for super dangerous implicit conversions, but then the same people rewrite push_back into emplace_back, which is much more lenient w.r.t. type conversions. It is also much easier to litter every function with [[nodiscard]] constexpr nothrow than it is to argue for readability.

    My best explanation is that "Modern C++" is making developers biased towards whatever language feature is newer, instead of making good decisions. I wish the community gave up on the moniker.

    • spacechild1

      `string_view` (and `std::span`) is great! Finally I don't have to use concrete types like `const std::string&` or `const std::vector<float>&` as function parameters when all I really want is just something that behaves like a string or array/vector type.

      In fact, `std::string_view` can lead to very elegant and efficient (zero-allocation) string parsing code.

      IME, taking a string_view as an argument is pretty unproblematic. Two areas where I have seen problems:

      1. returning a string_view from a function. Think twice before doing that.

      2. calling data() and passing the result to a function that expects a null-terminated C strings. (std::string_view is not guaranteed to be null-terminated.)

      ---

      Apart from that, there are so many great features and libraries in C++17 and C++20 that I use on a regular basis and that I would really miss:

      - structured bindings

      - if-statement initializers

      - fold expressions

      - inline variables (very useful for handling global state in header-only stuff)

      - if constexpr (replaces many uses of SFINAE)

      - concepts (replaces all of the remaining uses of SFINAE)

      - coroutines (really nice to use with ASIO)

      - std::optional

      - std::variant

      - std::filesystem

      - std::format (although I personally stick to fmt::format)

      - from_chars() and to_chars() (fast, local-independent string-number conversion routines)

      - std::bit

      - std::apply, std::invoke, std::bind_front

      - etc.

      • moltopoco

        I have no problem with any of these features, only with the culture surrounding them. The point should be to write maintainable, safe, and efficient code, but sometimes that is at odds with writing "modern" code because the boring old C++11 way works just fine. In my anecdotal experience (YMMV), "modern" then often trumps other concerns when it should never have been a goal in itself.

        • spacechild1

          Most of the features I have listed make the code more readable, which I would count as more maintainable. None of them have a runtime overhead.*) In fact, `<format>` and `<charconv>` are significantly faster than their counterparts.

          Personally, I don't use any modern C++ features just for the sake of it. That would be silly indeed! For example, I have never used ranges/views because I don't find them particularly compelling, at least for things I do.

          *) Creating a coroutine frame typically requires a heap-allocation, but so does the old ASIO callback style.

          • moltopoco

            We may be in agreement, so I'm sorry to split this hair: These features don't make the code more readable. They _can_ make it more readable.

            You can also replace a simple overloaded function with an implicit template that uses modern "if constexpr" to distinguish between input types. Or you can "modernize" a legacy function that takes a const string& to take string_view instead, even if the very first thing the function does is to turn it back into a string to pass its c_str() to a C API (leading to more net allocations). And so on.

            I always found the fashion trends of the C++ community counter-productive, and that is what I personally associate with the term "modern C++" nowadays (since around C++20). It was the same with iterators after C++98 - cool, now we can express algorithms with all the clarity and safety of C pointer arithmetics. With all due respect to the original STL, std::erase_if(cont,pred) is a godsend in its simplicity, but I suspect it would have been against the zeitgeist for a long time.

      • pwiecz

        > IME, taking a string_view as an argument is pretty unproblematic. Two areas where I have seen problems

        I've seen a few problema with automatic conversion of const std::string& arguments to std::string_view. They arised when the argument was later captured by value in a lambda and passed as an argument to e.g. std::thread. Instead of safely passing copy of the string now you pass refwrence to a memory owned by a different thread, which may be deallocated at any moment.

        • Conscat

          This can be prevented by `[[clang::lifetimebound]]` and I've been doing that lately.

      • Joker_vD

        > all I really want is just something that behaves like a string or array/vector type.

        What does behave like a std::string/std::vector but is not std::string/std::vector, respectively? In my experience, if you want a flat array behind a T* (or std::dequeue<T> at the worst) — so that the autovectorization can kick in — then that's precisely what you want, and no generalization to e.g. a linked list, or a heap, or a red-black tree is really possible.

        • spacechild1

          > What does behave like a std::string/std::vector but is not std::string/std::vector, respectively?

          `std::span` works with all vector/array-like types: `std::vector`, `std::array`, C-arrays and third-party vector/array types.

          Similarly, `std::string_view` works with `std::string`, C-strings and third-party string types.

          > so that the autovectorization can kick in

          It's not even about vectorization. Often you just have a function that takes a string-like or array-like thing, but you don't care about the exact container. Yes, you could make the argument a template (and constrain it with concepts), but then you have to put everything in headers.

    • William_BB

      To be fair, string_view has one very attractive feature over const ref string. String's char constructor is not explicit. With string view, you are guaranteed to not allocate by accident.

      • spacechild1

        Exactly! When you have a `const std::string&` argument, passing a C-string implicitly allocates a new temporary `std::string`.

        `std::string_view` accepts both `const std::string&` and `const char*` without hidden allocations.

        • moltopoco

          Understood, and I also use string_view where it makes sense. But while overusing std::string can cause unnecessary and hidden allocations, overusing string_view often leads to memory errors.

          I also love C++11 "auto", but overusing it then makes the string/string_view bugs harder to find.

          I am perfectly happy with with idiomatic C++11 plus features from recent standards, but I don't feel like that counts as "modern C++" anymore. A sibling comment put it better: idiomatic C++20 has become its own separate thing, and I personally don't love it, and it is certainly not more robust in my experience.

          • spacechild1

            > I also love C++11 "auto", but overusing it then makes the string/string_view bugs harder to find.

            Yes, I have soured a bit on the "almost always auto" mantra myself. There are cases where explicit types are both clearer and safer.

  • jandrewrogers

    In the case of C++, "modern" conventionally denotes "C++11 or later". Idiomatic code between "legacy" and "modern" dialects of C++ that they almost look like different languages. C++11 was a breaking change to the idioms of older C++.

    This is helpful context to the reader. Many C++ programmers are unfamiliar with either idiomatic legacy or modern dialects and would have difficulty following the code.

    There are arguably three different idiomatic dialects of C++ now: legacy, C++11, and C++20. We don't really have a common term for distinguishing between C++11 and C++20 even though the idiomatic code looks quite different.

    It has been an interesting evolutionary strategy. C++ standards meticulously avoid breaking changes but will happily do a major redesign of how idiomatic code is written.

  • htfy96

    Yes. Using C++23 in work with house-built coroutine abstraction is very fun and robust. Have to admit that it just requires certain level of skills of your colleague, and unfortunately general hires don't have.

  • omoikane

    Maybe they are referring to Effective Modern C++ book by Scott Meyers.

    I think "modern" is used to contrast newer C++ standards from C++98.

    • ibobev

      Or an older book.

      "Modern C++ Design: Generic Programming and Design Patterns Applied" by Andrei Alexandrescu

      Actually, he coined the term. Then Scott Meyers reused it, and now it is a common adjective used before C++ books, libraries, tooling, and so on.

    • jibalt

      They are quite explicitly referring to C++23

  • eluru

    Sounds tiring, CPP has been modernised, killed, solved, so many so so many times now, not even one was for real, this wont be the one either

rhodo

This is sick man, I think you should be really proud of what you built!

AshamedCaptain

The biggest issue with clangd (and a derivate like this) is that it is at least three orders of magnitude slower than non-clang based indexers like the old ones in Eclipse and Qt Creator. It used to be that I could index a huge codebase in a couple minutes; clangd takes days and therefore we have to trouble with stuff such as remote indices. Frankly, I'd trade anything in the world for a fast indexer that is NOT based on clang, even if it were not perfect (which is kind of the point).

  • vintagedave

    Are you sure? This is explicitly not a ‘derivative’ and has a fascinating coroutine with multiprocess architecture. I would not automatically expect it to be too slow.

    I agree re the old ones though. But they did far less. This design is heavy on completing templates. That’s tricky!

    • kamrann_

      I would assume they're referring to it being built on top of clang. While it does claim to have improved on some architectural decisions in clangd, and my testing of it today was indeed much better than clangd when it comes to handling modules in particular, at the end of the day the bottleneck for responsiveness remains by a distance the time it takes clang to recompiled a translation unit.

  • whizzter

    I think you should reconcile with Clang and hope that they can work on performance for LSP tasks, because while C++ 26 reflections does a lot of "conservative" things right, it also enabled consteval to create structures with novel shapes, so you need consteval in an indexer.

    ie.

    auto val = make_named<"x","y">(1, 2.3);

    // val is now a struct with an integer field named x and a double field named y

    • account42

      For the IDE use case its not clear support for that is worth giving up realtime parsing speeds. Of cource you can try to have your cake and eat it too by compiling a fast but limit parser with clangd but then you are doubling your bug surface.

  • qalmakka

    Clangd is slower precisely because it's very precise. The old indexers you mention, like the one in Eclipse CDT, never truly worked for me. They weren't that useful and were always lacking compared to IntelliSense. clangd was the first time the FOSS alternative truly matched and often trounced IntelliSense

    • kamrann_

      Yeah I think this is exacerbated by the ever increasing complexity of the language. Anything using workarounds to avoid implementing a full frontend is just going to fall apart on modern code depending on the standard library.

      • qalmakka

        Precisely. Clangd can successfully index the entirety of Unreal Engine for instance (after fixing how you generate compile_commands.json, and taking like 32 GB of RAM in the process), which is not a small feat. Let's see if this new kid can do the same, it has been indexing 43k files for 10 hours now so for one it is indeed slower than clangd, which took like only 5

  • mwkaufma

    Derivative?

        It is still built on LLVM/Clang, but unlike clangd and ccls, which are also Clang-based, clice was redesigned from the ground up to fix architectural problems that have gone unsolved in those servers for years.
qyron

Interesting project and great writeup! Looking forward to try this.

Neywiny

Interesting they show clangd crashing in vscode. That's been my biggest gripe with it. But they also show clice crashing in the same spot? Seems odd to not just fix it

  • quacker

    They explain exactly why in article:

    Because C++ is simply too complex, and on certain corner cases the compiler can easily crash or leak memory. That is especially true for incomplete code in the middle of editing: a huge downstream user base tests compilers on complete code, while very few people test them on half-written code, where the compiler mostly falls back on best-effort heuristics

    It's the compiler that crashes. Both clangd and clice are built on top of the same compiler (clang) which is why both experience a crash on the same code.

  • dwattttt

    I'm not a C++ expert, but I don't see anything in the shown snippet that would cause a language server crash, so I imagine the server is crashing on something off screen.

    I could understand them showing off crash recovery by using code deliberately written to crash the language server, and so not being something you can "fix".

    • Neywiny

      IME it's nothing about the code. It'll be weird setup issues that aren't traceable. In one IDE I end up having to just have a new workspace and copy the code over and somehow that fixes it. It's truly beyond me. At best I'll see an ENOENT error but that doesn't help me much.

  • qalmakka

    The problem is that the language is simply too complex with way too many corner cases.

    For instance, C++ the only programming language that comes to my mind where the grammar is type 0 in the Chomsky classification - it means it has to be parsed with a Turing machine and thus every C++ compiler suffers from the halting problem.

    This means that having a C++ compiler that will never crash or hang is literally mathematically impossible unless you stick to only parsing a subset of the grammar. Modern compilers bypass this by doing things like capping how many times templates can recurse, etc but this means that the frontend it will inevitably end up in unsavoury corners where it relies on its authors explicitly handling them - which sometimes they did not

    • account42

      > The problem is that the language is simply too complex with way too many corner cases.

      Bullshit. These are compiler bugs that can be fixed. It's just not a priority because it doesn't affect normal use with valid code.

      > This means that having a C++ compiler that will never crash or hang is literally mathematically impossible unless you stick to only parsing a subset of the grammar.

      This is only a theoretical concern since every compiler will accept arbitrarily large programs anyway due to being built on fixed-width integer architectures which can only address finite amounts of memory. Real compiler can set reasonable limits where they can avoid both hangs and crashes.

ggcr

Will be replacing clangd with this. Thanks OP!

BoingBoomTschak

The vibe smell is one of the strongest I've ever encountered. Who thought commit messages each dwarfing a Dumas novel was a good idea?

https://github.com/clice-io/clice/commits/main/

  • kuhsaft

    I'm not against agentic coding, but insanely verbose commit messages in place of actual decision making records are definitely an anti-pattern.

    Requirements, specs, and ADRs shouldn't be thrown out with agentic development. I've found that using rust-style `r[rule.id]` for requirements and specs with ADRs works a lot better than verbose commits.

    Requirements cover what, specs cover how, ADRs cover why, and git commits cover when.

  • nraynaud

    There is this saying that automation doesn’t reduce work, but it’s hard in the beginning to guess what the new work will be. I suspect that the hard human thing to do in the future is wading through the mountains of sloppy words.

    I’m baffled that AI companies still train their LLM to produce that many words, it’s wasting everybody’s time and destroy the environment.

nnevatie

> A completely hermetic, reproducible build powered by Bazel

Why not CMake, the de-facto (like it or not) build tool for C++?

signa11

reading this, i am reminded of verding's first rule of programming

    Any sufficiently complicated concurrent program in another language contains an ad hoc 
    informally-specified bug-ridden slow implementation of half of Erlang.
drnick1

Meh, I will stick with the kind of C++ that is useful and readable, basically C++98 or C with classes. There are now ten ways of doing the same thing, the language is a complete mess.

glebd

It's slop, unfortunately.

Keyboard Shortcuts

j
Next item
k
Previous item
o / Enter
Open selected item
?
Show this help
Esc
Close modal / clear selection