Develop branch build status

[d.healey]

@Christoph-Hart Thanks, I'll give it a go this afternoon

[d.healey]

HISE builds but I'm getting some errors with snex workbench. I'll just paste a small part of the output here but there is a lot more of the same.

../../../../../HISE/hi_snex/snex_public/snex_jit_FunctionData.cpp:791:39: note: called from here
  791 |    case ID::Integer: return c4_ttvv<R>(f, a1, a2, (int)a3, a4);
      |                             ~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~
../../../../../HISE/hi_snex/snex_public/snex_jit_FunctionData.cpp:785:54: error: inlining failed in call to ‘always_inline’ ‘static R snex::jit::VariadicCallHelpers::ReturnFunctions::c4_ttvv(const snex::jit::FunctionData&, T1, const snex::VariableStorage&, const snex::VariableStorage&, const snex::VariableStorage&) [with R = void*; T1 = void*]’: recursive inlining
  785 |   template <typename R, typename T1> variadic_call R c4_ttvv(const FunctionData& f, T1 a1, const VariableStorage& a2, const VariableStorage& a3, const VariableStorage& a4)
      |                                                      ^~~~~~~
../../../../../HISE/hi_snex/snex_public/snex_jit_FunctionData.cpp:792:38: note: called from here
  792 |    case ID::Double: return c4_ttvv<R>(f, a1, a2, (double)a3, a4);
      |                            ~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~
../../../../../HISE/hi_snex/snex_public/snex_jit_FunctionData.cpp:785:54: error: inlining failed in call to ‘always_inline’ ‘static R snex::jit::VariadicCallHelpers::ReturnFunctions::c4_ttvv(const snex::jit::FunctionData&, T1, const snex::VariableStorage&, const snex::VariableStorage&, const snex::VariableStorage&) [with R = void*; T1 = void*]’: recursive inlining
  785 |   template <typename R, typename T1> variadic_call R c4_ttvv(const FunctionData& f, T1 a1, const VariableStorage& a2, const VariableStorage& a3, const VariableStorage& a4)
      |                                                      ^~~~~~~
../../../../../HISE/hi_snex/snex_public/snex_jit_FunctionData.cpp:793:38: note: called from here
  793 |    case ID::Float:  return c4_ttvv<R>(f, a1, a2, (float)a3, a4);
      |                            ~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~
../../../../../HISE/hi_snex/snex_public/snex_jit_FunctionData.cpp:785:54: error: inlining failed in call to ‘always_inline’ ‘static R snex::jit::VariadicCallHelpers::ReturnFunctions::c4_ttvv(const snex::jit::FunctionData&, T1, const snex::VariableStorage&, const snex::VariableStorage&, const snex::VariableStorage&) [with R = void*; T1 = void*]’: recursive inlining
  785 |   template <typename R, typename T1> variadic_call R c4_ttvv(const FunctionData& f, T1 a1, const VariableStorage& a2, const VariableStorage& a3, const VariableStorage& a4)
      |                                                      ^~~~~~~
../../../../../HISE/hi_snex/snex_public/snex_jit_FunctionData.cpp:794:39: note: called from here
  794 |    case ID::Pointer: return c4_ttvv<R>(f, a1, a2, (void*)a3, a4);
      |                             ~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~
../../../../../HISE/hi_snex/snex_public/snex_jit_FunctionData.cpp:785:54: error: inlining failed in call to ‘always_inline’ ‘static R snex::jit::VariadicCallHelpers::ReturnFunctions::c4_ttvv(const snex::jit::FunctionData&, T1, const snex::VariableStorage&, const snex::VariableStorage&, const snex::VariableStorage&) [with R = void*; T1 = void*]’: recursive inlining
  785 |   template <typename R, typename T1> variadic_call R c4_ttvv(const FunctionData& f, T1 a1, const VariableStorage& a2, const VariableStorage& a3, const VariableStorage& a4)

[Christoph Hart]

Can you compare the build flags - this isn‘t something that has changed since yesterday.

[d.healey]

@Christoph-Hart I didn't notice anything in Projucer however it occured to me that yesterday we only built the debug version, so I just tested and the debug version builds, the problem is with the release version only.

[Christoph Hart]

Ah ok that makes sense. But HISE builds as release build?

[d.healey]

Yes release version of HISE works.

[arminh]

Anyone tried to build VST3 x86? Im getting issue like this.

c:\users\admin\desktop\hise-53483599ac7ff8ac358f13d102ff56a825cd0056\hi_dsp_library\snex_basics\snex_types.h(409):
 error C2338: misaligned poly handler (compiling source file ..\..\JuceLibraryCode\include_stk_wrapper.cpp) [C:\Users\admin\Desktop\New Plugin\Binaries\Builds\VisualStudio2017\New Plugin.SharedCode.vcxproj]

[Christoph Hart]

32bit is deprecated. Just don't :)

[arminh]

@Christoph-Hart it's bad :D Now I can't run plugins on Windows ARM :P

[d.healey]

@Christoph-Hart said in Develop branch build status:

I think I have fixed all issues except for the one where the export caused the workbench to crash (I tried to fix that one, but it's a shot in the dark).

I set AllowCompilation to true. I then go to Tools >> Compile DSP networks and instant crash. Did I miss a step?

This is from gdb

Thread 1 "snex_workbench" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7ffff7329c80 (LWP 26609)]
0x0000000000e7eca0 in snex::jit::FunctionData::operator=(snex::jit::FunctionData const&) ()

[crd]

I am having trouble building the latest develop version on Mojave using Xcode 10.3. I am getting this error:

[d.healey]

@crd Christoph just added that yesterday, it's not an important feature so for now try building from a commit before that addition.

[d.healey]

@Christoph-Hart My old friend has returned with the latest commit :)

../../../../../HISE/hi_dsp_library/snex_basics/snex_TypeHelpers.cpp:160:68: error: call of overloaded ‘var(int64_t&)’ is ambiguous
  160 |  case Types::ID::Pointer: v = var(*reinterpret_cast<int64_t*>(data)); break;

I think it's possibly fixed in this pull request - https://github.com/christophhart/HISE/pull/192

And another error I don't recognise

../../../../../HISE/hi_scripting/scripting/scriptnode/ui/NodeComponent.cpp:90:3: error: cannot convert ‘scriptnode::NodeComponent::Header::Header(scriptnode::NodeComponent&)::<lambda(juce::ValueTree&, bool)>’ to ‘const ChildChangeCallback&’ {aka ‘const std::function<void(juce::ValueTree, bool)>&’}
   90 |  });

[Christoph Hart]

Yes I'm chewing through macOS compilation at the moment and the second one also appears here. I'll commit a fix to those in a few minutes.

[d.healey]

@Christoph-Hart Thanks, just tried the latest and all compiled, but as soon as I tried to go to the script editor HISE crashed, going to try a debug build now and see if I can narrow it down. Unfortunately I'm going to be out all afternoon so won't get to play around with it much until this evening.

Script editor works in debug build - I'll experiment with different optimization levels.

I notcied the home/end keys are screwy in the new editor...

Update: It crashes with optimisation lvl 2 so I should be able to debug this when I get back.

Update 2: Crash is triggered by this function, I recall we had an issue here before
void mcl::HighlightComponent::paintHighlight(Graphics& g)

Line 68 for (auto sr : document.getSearchResults())

[Christoph Hart]

I've changed this line to something else, but I have no idea why it would crash or if the new way fixes it, so please have a go :)

[d.healey]

@Christoph-Hart Thanks, I'll try it this evening :)

[d.healey]

@Christoph-Hart said in Develop branch build status:

I've changed this line to something else, but I have no idea why it would crash or if the new way fixes it, so please have a go :)

Yep that seems to have fixed it (in debug at least, building release now).

End/Home keys are still messed up. Where should I hunt in the codebase to try and fix this myself?

[Christoph Hart]

HISE/hi_tools/mcl_editor/code_editor/TextEditor.cpp at 3c0afbf057a7323458a721ad1967fa4722289d1c · christophhart/HISE

The open source framework for sample based instruments - HISE/hi_tools/mcl_editor/code_editor/TextEditor.cpp at 3c0afbf057a7323458a721ad1967fa4722289d1c · christophhart/HISE

GitHub (github.com)

Try

#if JUCE_WINDOWS || JUCE_LINUX

instead.

[d.healey]

@Christoph-Hart Yep that did it, thanks!