Important: Branching update

  • Hi everybody,

    I'm not sure whether I'm the idiot or if GitHub's pull request tool has it twisted, but I ended up merging the codegen rewrite branch into the scriptnode branch (instead of the other way around).

    Anyways, I took this as an opportunity to get back to a branching model that is not completely pointless, so we're back on master / develop! That means:

    1. The branch formerly known as scriptnode is now the master branch. I just used the commit just before the messed up merge with codegen, so you can use this in production.
    2. The branch formerly known as scriptnode-codegen-rewrite is now the develop branch.

    I will add all features / bug fixes to the develop branch from now on, so unless you are working on a live project that has already been released, it's highly recommended to use the develop branch. In the foreseeable future, the develop branch will be merged into master, which is the exact time that HISE 3.0 is released.

    EDIT: Also wait an hour or two before switching, the merge process has messed up the compilation...

  • Is now the time to report all my GNU/Linux build errors or should I wait until the develop branch is a little further along?

  • Where should we put issues that we have with the develop branch? I am moving my almost finished instrument from scriptnode to the develop version and am finding some things.

  • @Christoph-Hart said in Important: Branching update:

    it's highly recommended to use the develop branch.

    In Mac, which XCode and Operating System version do you suggest as minimum requirments for this branch?

    Can we use Sierra & XCode8 for this?

  • @Christoph-Hart

    Do you want bugs we find in the develop branch in "The definitive feature request & bug fix roadmap"?

  • @Steve-Mohican said in Important: Branching update:

    Can we use Sierra & XCode8 for this?

    99% sure that's a no since you can't use that config for previous versions. I would guess Mojave and Xcode 10.3 10.2 are still the minimum requirements.

  • @Steve-Mohican I'm using Catalina and Xcode 10.2 and it builds successfully

  • @d-healey @ulrik Thank you for the info.

    What about to make the plugins Support M1 natively? Catalina is ok but I think XCode 10.2 wouldn't be enough for Native M1 Support?

  • @Steve-Mohican I don't know if XCode 10.2 is enough for building for M1 though, I have note tried it yet

  • Should we be using JUCE/Projucer 6 now (with develop)?

  • Yes. I think I've pushed new builds of the Projucer for Mac and Windows, but on Linux you can just compile the one in JUCE/extras/ in the HISE repo (it's the most recent version).

  • @Christoph-Hart Thanks, that built, shall I include it with my next pull request?

    I had to modify the makefile and jucer file slightly to get it to build successfully, I tested with the version from the JUCE git repo and had the same issue. Seems I just needed to delete my old Projucer config files folder.

  • Ooh, does this mean HISE will use the newer default sliders that JUCE has?
    I know, silly thing to be excited about...

  • I'm slowly getting close to a successful GNU/Linux build of HISE develop branch.

    I had to add -fpermissive flag to avoid a few errors and I had to change some of the math functions from their f variants because of that bug in the standard library on Linux.

    The remaining issues I'm not sure how to solve.

    I'm getting quite a few of these errors for different SNEX files.

    /HISE/hi_snex/snex_library/snex_jit_ExternalComplexTypeLibrary.cpp:658:38: error: call of overloaded ‘mov(asmjit::x86::Mem&, juce::int64)’ is ambiguous
    ../../../../../HISE/hi_backend/../hi_modules/../hi_scripting/scripting/scriptnode/api/DspNetwork.h:1222:63: error: ‘snex::cppgen::ValueTreeIterator’ has not been declared
     1222 |    cppgen::ValueTreeIterator::forEach(saveCopy, snex::cppgen::ValueTreeIterator::IterationType::Forward, stripValueTree);

    And then these two errors:

    ../../../../../HISE/hi_dsp_library/snex_basics/snex_DynamicType.cpp:21:50: error: conversion from ‘const juce::var’ to ‘int64_t’ {aka ‘long int’} is ambiguous
       21 | = reinterpret_cast<void*>((int64_t)value);
    ../../../../../HISE/hi_dsp_library/snex_basics/snex_TypeHelpers.cpp:152:69: error: call of overloaded ‘var(int64_t&)’ is ambiguous
      152 |   case Types::ID::Pointer: v = var(*reinterpret_cast<int64_t*>(data)); break;

    @Christoph-Hart If you can give me any tips here I can implement and test them.

  • @d-healey said in Important: Branching update:

    /HISE/hi_snex/snex_library/snex_jit_ExternalComplexTypeLibrary.cpp:658:38: error: call of overloaded ‘mov(asmjit::x86::Mem&, juce::int64)’ is ambiguous

    That error is ambiguous...

Log in to reply