Sony Proposes Changing LLVM Clang Default To C++20 Mode

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • kpedersen
    replied
    Originally posted by lowflyer View Post
    But I have the impression that you are blinded by your idea of "micro$oft is a decade behind"
    A decade? Hah, in terms of C, they are much further behind than a mere decade.

    Originally posted by lowflyer View Post
    C is no longer relevant [...] especially to them
    They are so far behind that they decided to no longer market their compiler as a C compiler? Either you are a mug if you believe that is what they wanted, or you probably want to ask them to remove the "modern C" reference from their product page:

    Originally posted by lowflyer View Post
    Just for the records: C99 in GCC was completed with GCC 4.5 which came out in April 14, 2010.
    Not quite. The C99 support was in earlier than 4.0 but was deemed fully compliant by 4.5, including spec bugs and corner cases. Some vendors, including Microsoft *never* reach full compliance. And never will.

    ... and when did MSVC get C99 again? A decade after that? In terms of full compliance? 2025 and counting...

    Originally posted by lowflyer View Post
    C is no longer relevant generally
    Oh, right. You aren't very experienced then. Thats fine, though due to that, there is no need to continue this discussion.
    Last edited by kpedersen; 15 January 2025, 10:52 AM.

    Leave a comment:


  • KernelCrasher
    replied
    Originally posted by brad0 View Post
    Michael being pathetic as usual.
    Be jerk elsewhere.

    Leave a comment:


  • coder
    replied
    Originally posted by lowflyer View Post
    Ask Alexa.
    I see what you did there (wikipedia link). It's worth noting that some of those are non-x86 and some are actually based on LLVM. Once you eliminate the ones that don't even support C++17, the list shrinks quite rapidly!

    Sorry, I assumed you had more detailed knowledge of which contenders seemed worth watching. My bad. I'll try not to make any presumptions of your expertise, in the future.
    Last edited by coder; 15 January 2025, 05:51 AM.

    Leave a comment:


  • lowflyer
    replied
    Originally posted by kpedersen View Post

    I feel you didn't quite get the point. It was never about C99 being the end goal, it was more about it took them a decade longer to reach it than the others. Likewise with C11, likewise with C17, they are well behind the competition due to a lack of manpower and expertise in that specialist field.

    As an example, this announcement is a fairly comical example being many years later than all the other popular compiler vendors: C11 and C17 support arriving in 2020!
    Bien au contraire! I think I did get your point very well.

    But I have the impression that you are blinded by your idea of "micro$oft is a decade behind" that you forgot to read the rest of my post. You were too busy to blur out your conclusion: "lack of manpower" that you didn't see mine: "C is no longer relevant generally and especially to them". Which has the same effect in the end. So we practically agree.

    Just for the records: C99 in GCC was completed with GCC 4.5 which came out in April 14, 2010.

    Leave a comment:


  • lowflyer
    replied
    Originally posted by coder View Post
    Such as? I'm genuinely curious.
    Ask Alexa. - Or perhaps these days it's possible to find some AI out there that can help you with existential questions like these. I think it was called something like "alter vista" ... No. I think the name was later changed to something that sounds more like "gugg el".

    Leave a comment:


  • kpedersen
    replied
    Originally posted by lowflyer View Post

    I don't think that it was just a "manpower" issue. They reached C11 and C17. I deem the quarter-of-a-century-old C99 irrelevant.
    I feel you didn't quite get the point. It was never about C99 being the end goal, it was more about it took them a decade longer to reach it than the others. Likewise with C11, likewise with C17, they are well behind the competition due to a lack of manpower and expertise in that specialist field.

    As an example, this announcement is a fairly comical example being many years later than all the other popular compiler vendors: C11 and C17 support arriving in 2020!

    Originally posted by carewolf View Post

    Msvc isn't a c compiler. They only support the c required by the standards
    Indeed. That's the "marketing angle". As I mentioned above, with their limited manpower, "naturally they made the (correct) business decision to compromise and cut features to focus on *just* C++."

    In other words, it only isn't a C compiler because they failed to keep up. The rest is just marketing "fibs". I am sure if they ever managed to catch up (rather than simply migrate to clang), suddenly MSVC will "be a C compiler" again

    Now imagine if they *also* had to support architectures other than Wintel and waarch64 again, they would have likely ditched MSVC almost a decade ago (similar to how Intel and Embarcadero did).
    Last edited by kpedersen; 14 January 2025, 02:18 PM.

    Leave a comment:


  • carewolf
    replied
    Originally posted by kpedersen View Post

    MSVC is a good example. They didn't have enough manpower to reach C99 conformance even a decade after the competitors. So naturally they made the (correct) business decision to compromise and cut features to focus on *just* C++.

    Another example with MSVC is the very limited ASan-family support. I suspect they will migrate to clang within this decade though.
    Msvc isn't a c compiler. They only support the c required by the standards, and while they used to suck greatly, they have the last few years been ahead of clang and gcc, especially in the standard library implementation of c++ features which has always lacked behind for gcc and been ignored by clang

    Leave a comment:


  • carewolf
    replied
    Isn't the biggest problem for clang C++20 libc++?

    Leave a comment:


  • coder
    replied
    Originally posted by lowflyer View Post
    coder , A downgrade to a 2-horse race will make room for others. There are a number of contenders.
    Such as? I'm genuinely curious.

    Leave a comment:


  • lowflyer
    replied
    Originally posted by kpedersen View Post

    MSVC is a good example. They didn't have enough manpower to reach C99 conformance even a decade after the competitors. So naturally they made the (correct) business decision to compromise and cut features to focus on *just* C++.

    Another example with MSVC is the very limited ASan-family support. I suspect they will migrate to clang within this decade though.
    I don't think that it was just a "manpower" issue. They reached C11 and C17. I deem the quarter-of-a-century-old C99 irrelevant. On windows and the micro$oft side, the C language is pretty much a dead horse by now. Their focus is not *just* C++, they have their CLR variant of C++ (and C# and a few other languages). The have their own family of sanitizers. Asking micro$oft for ASan support is a bit odd. I don't know about you but when I want a steak, I don't go to the hairdressers.

    coder , A downgrade to a 2-horse race will make room for others. There are a number of contenders.

    Leave a comment:

Working...
X