Announcement

Collapse
No announcement yet.

LLVM-Strip Introduced As An Alternative To GNU Strip

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

  • LLVM-Strip Introduced As An Alternative To GNU Strip

    Phoronix: LLVM-Strip Introduced As An Alternative To GNU Strip

    LLVM-Strip is the newest tool added to mainline LLVM...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    always surprising to read what was not in there, when one intuitively would have thought, of course such a basic thing would have been done with their assembler and linker proof of concept first, ...

    Comment


    • #3
      What's the reason to re-implement such a basic program? Is it issues with licensing?

      Comment


      • #4
        Originally posted by willmore View Post
        What's the reason to re-implement such a basic program? Is it issues with licensing?
        NIH syndrome

        Comment


        • #5
          Originally posted by willmore View Post
          What's the reason to re-implement such a basic program? Is it issues with licensing?
          Getting rid of the binutils dependency. The most hassles I have with clang is that a matching binutils needs to be available, given that clang is a native crosscompiler, it might thus need to pick from the correct version of any number of target binutils, potentially in different root-prefixes. I am stuck at clang 4.0 for most arm-none-eabi projects as the detection heuristics for that changed.
          (granted that most of the time, the linker + fitting scripts causes problems, and not the simple strip tool)

          Comment


          • #6
            Originally posted by willmore View Post
            What's the reason to re-implement such a basic program? Is it issues with licensing?
            Likely. the whole point of LLVM was "we don't like GNU compiler license"

            Comment


            • #7
              Originally posted by starshipeleven View Post
              Likely. the whole point of LLVM was "we don't like GNU compiler license"
              That was the *BSD and Apple perspective to embrace it. The original quest was a research project about dynamic compilation. Anyway, it is good to have competition; I'm pretty sure gcc would not have experienced such fast development pace on last years if it was not because llvm.

              Comment


              • #8
                Originally posted by acobar View Post
                That was the *BSD and Apple perspective to embrace it. The original quest was a research project about dynamic compilation. Anyway, it is good to have competition; I'm pretty sure gcc would not have experienced such fast development pace on last years if it was not because llvm.
                My point was only that the main reason for LLVM being what it is is that there are enough people/companies that don't like gcc's license. I'm not saying it's a bad thing.

                Not liking the license is still much better than just NIH syndrome.

                Comment


                • #9
                  Problem: we can't get the UNIX tools to work correctly
                  Solution: it makes us seem more important to have several gigabytes of software

                  Android did the same thing. There is no open source software for Android because to patch software you need several gigabytes of IDEs. But that's good, because Google gets ad money out of all the ad-supported trash that could have been open source.

                  Comment

                  Working...
                  X