Announcement

Collapse
No announcement yet.

Linux 5.15's New "-Werror" Behavior Is Causing A Lot Of Pain

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

  • #11
    Originally posted by marek View Post
    Different gcc versions print different warnings. Other gcc versions have buggy warnings. Good luck building the kernel everywhere.
    I'm curious, is LLVM also as problematic on this issue?

    Comment


    • #12
      Ok now that I'm done laughing. It was just an obviously bad decision. I think Linus knew _exactly_ what he was doing wiht this move.

      Marek brought up a fantastic point about the different gcc/llvm versions. It's just a huge road block and nuisance that gets in the way of developers (expensive and busy and limited) time.

      They should *not* be fighting the compiler. They should spend their time writing awesome code. (which the compiler "warnings" are there to help assist). I think we're overthinking this.

      Comment


      • #13
        The reasonable middle ground is very reasonable. It will lead to people fixing their code moving forward without breaking current builds.

        Comment


        • #14
          Originally posted by perpetually high View Post
          Ok now that I'm done laughing. It was just an obviously bad decision. I think Linus knew _exactly_ what he was doing wiht this move.

          Marek brought up a fantastic point about the different gcc/llvm versions. It's just a huge road block and nuisance that gets in the way of developers (expensive and busy and limited) time.

          They should *not* be fighting the compiler. They should spend their time writing awesome code. (which the compiler "warnings" are there to help assist). I think we're overthinking this.
          You are still ignoring the glaring fact that compiler warnings are configurable and that the comittee will decide upon which warnings are relevant to the pursued level of code quality. After that there's a standard level which all code must meet.

          Initial setup phase is always painful and it only gets more painful after each year of ignorance. This is what Linus has obviously been dealing with always and it has now reached the point where it is too painful for a single person to withstand. It's time for those $200k-per-year overpaid engineers to get up from their asses and start to do the lifting too.

          Whining about varying errors among compilers and compiler versions is a non-argument. I would presume that the enabled warnings are selected in a manner which is compatible with the supported set of compilers. Obviously you're stepping into the unknown if you opt for a non-supported compiler anyway, then all the obstacles that you meet are your own fault.
          Last edited by curfew; 08 September 2021, 12:03 AM.

          Comment


          • #15
            There weren't that many replies on the thread, but maybe Linus got his point across. I also noticed a distinct lack of swearing in the discussion.

            Comment


            • #16
              curfew

              I'm not disagreeing with your post, I'm just saying it shouldn't be the default behavior. Unnecessary pain.

              Again, I understand the motive. Just don't think it's the correct course.

              The initial "blow-back" could be just temporary, or it could be a call to reverse. I'll let actual kernel developers discuss further.

              Comment


              • #17
                Originally posted by perpetually high View Post
                curfew

                I'm not disagreeing with your post, I'm just saying it shouldn't be the default behavior. Unnecessary pain.

                Again, I understand the motive. Just don't think it's the correct course.

                The initial "blow-back" could be just temporary, or it could be a call to reverse. I'll let actual kernel developers discuss further.
                Unnecessary pain or warning shot fired across the bow?

                Comment


                • #18
                  Originally posted by marek View Post
                  Different gcc versions print different warnings. Other gcc versions have buggy warnings. Good luck building the kernel everywhere.
                  With Linux it's even crazier than that sometimes. For example, I used to build custom Android roms, and simply building the exact same kernel source using a different gcc version would result in a kernel that would refuse to boot.

                  Comment


                  • #19
                    While I think that it should be enabled by default, I also think there should be a target toolchain infrastructure to proof against. If feels like a crazy move to enable this for everything without to know what downstream builds against. Like in 5.15 builds 5.15.1 builds 5.15.2 wont build because there was a llvm update that now creates a feature deprecation warning. Badumm.

                    Comment


                    • #20
                      Linus:

                      "And I'm most definitely not convinced when the "let's finally enable
                      -Werror after years of talking about it", people end up going "but but
                      but I have thousands of warnings".

                      That's the POINT of that commit. That "but but but I have thousands of
                      warnings" is not acceptable."

                      Heh heh. Love that man. 😊

                      Comment

                      Working...
                      X