Tweaks for building with macOS 26 by mitchute · Pull Request #11605 · NatLabRockies/EnergyPlus

mitchute · 2026-05-27T00:25:54Z

Pull request overview

A collection of tweaks needed to get EnergyPlus building on macOS 26 for Xcode- and Unix-generated makefiles.

Description of the purpose of this PR

Pull Request Author

Reviewer

…builds

jmarrec

I don't really get why we have XCode specific stuff but maybe it's really needed?

I pushed a branch with only 2 changes that allows me to build on Apple Clang 21 (macOS 26.5 fully up to date)

https://github.com/NatLabRockies/EnergyPlus/compare/build_apple-clang-21?expand=1

jmarrec · 2026-05-29T12:06:10Z

+    if(CMAKE_GENERATOR STREQUAL "Xcode")
+      target_compile_options(project_warnings INTERFACE -Wno-shorten-64-to-32)
+    endif()


I have tweaks locally to build on apple-clang 21, and I haven't needed to add this?

Also, is this really Xcode generator specific?

(I would have expected CMAKE_CXX_COMPILER_ID MATCHES "Clang")

jmarrec · 2026-05-29T12:06:23Z


 add_executable(generate_embeddable_epJSON_schema embedded/generate_embeddable_epJSON_schema.cpp)
-target_link_libraries(generate_embeddable_epJSON_schema PRIVATE project_options project_fp_options project_warnings fmt::fmt)
+target_link_libraries(generate_embeddable_epJSON_schema PRIVATE project_options project_fp_options project_warnings fmt::fmt nlohmann_json)


Haven't needed that either

jmarrec · 2026-05-29T12:15:47Z

+#ifndef CLI11_HAS_CODECVT
+#if defined(__APPLE__) && defined(__clang__)
+#define CLI11_HAS_CODECVT 0
+#elif defined(__GNUC__) && !defined(__llvm__) && !defined(__INTEL_COMPILER) && __GNUC__ < 5


I see what you're doing. This is better than my own

target_compile_options(project_warnings INTERFACE -Wno-deprecated-declarations) # wstring_convert deprecated in newer libc++

I wonder if it's better to have it inline in CLI11.hpp or if it should be in CMake via target_compile_definitions(xxx PRIVATE CLI11_HAS_CODECVT=0) (for targets energyplus and convertInputFormat). I guess inline is fine.

mitchute · 2026-06-01T15:01:32Z

@RKStrand can you give @jmarrec's branch a try? https://github.com/NatLabRockies/EnergyPlus/tree/build_apple-clang-21

RKStrand · 2026-06-01T15:38:19Z

@mitchute @jmarrec Ok, I just tried this. Pulled it down, ran configure and generate on CMake, cleaned the build folder in Xcode, and then tried to build. It was not successful. Here are the error messages:

mitchute · 2026-06-01T15:40:16Z

OK. I also see this with Xcode and Apple Clang 17. Can you try a Unix makefile build with CLion?

RKStrand · 2026-06-01T15:52:07Z

I was just able to successfully build using CLion--both energyplus and energyplus_tests. Was able to successfully run 1ZoneUncontrolled.idf using this energyplus build and the test suite completed 1260 test successfully. Not sure why it didn't do every unit test but at least it successfully completed what it did.

I'd love to get Xcode running again, but at least having CLion up and running is super helpful.

RKStrand · 2026-06-01T15:54:30Z

For the record, in case it helps, I'm running a Nov 2023 MacBook Pro 16-inch, with the Apple M3 Max chip, 128 GB of memory, macOS Tahoe26.5, and Xcode Version 26.5 (17F42).

mitchute · 2026-06-01T18:17:44Z

Was able to successfully run 1ZoneUncontrolled.idf using this energyplus build and the test suite completed 1260 test successfully. Not sure why it didn't do every unit test but at least it successfully completed what it did.

@RKStrand can you post anything regarding what caused the failure in completing all the tests?

RKStrand · 2026-06-01T21:31:10Z

I just tried this in debug and here is a place that it seems to have a problem:

mitchute · 2026-06-01T21:32:35Z

Thanks @RKStrand. I saw that as well.

…nerated builds

mitchute · 2026-06-02T14:40:50Z

@RKStrand can you try this branch for both? https://github.com/NatLabRockies/EnergyPlus/tree/build_apple-clang-21_msm

RKStrand · 2026-06-02T15:12:56Z

@mitchute Ok, here is what I got on my machine:

Big Picture for "msm" branch: Xcode builds and runs successfully, CLion encounters an error building energyplus_tests (details below)

Xcode:
energyplus--builds successfully (two warnings in epexpat related to xmltok_ but that didn't prevent the build) and ran 1ZoneUncontrolled.idf without issue
energyplus_tests--builds successfully (same warnings) and runs all tests (2229) successfully without any issues

CLion:
energyplus--buildings successfully and ran 1ZoneUncontrolled.idf without issue
energyplus_tests--build error as shown here...

.../third_party/gtest/googletest/include/gtest/gtest-printers.h:478:35: error: implicit conversion from 'char8_t' to 'char32_t' may change the meaning of the represented code unit [-Werror,-Wcharacter-conversion]
478 | PrintTo(ImplicitCast_<char32_t>(c), os);

It points to the (c) in that line as the issue.

mitchute · 2026-06-02T15:22:47Z

Right... I missed that change here. I'll push a patch soon.

mitchute · 2026-06-02T15:39:23Z

@RKStrand give it another shot now.

RKStrand · 2026-06-02T17:43:10Z

@mitchute Ok, after that, now I can say: Xcode--same as before (everything good).

CLion--energyplus still builds and runs. energyplus_tests builds now. What is a bit weird is that when I run it, it only does 1260 tests and then at some point, I get the following pop-up window (see left of image below).

If I click on ignore, it just stops right there. If I click on "Reopen", it launches a terminal window and tries to run energyplus_tests from the terminal. But that doesn't finish either because as you can see from the right side of the image, it seems to crash in a Solar Shading test.

This reverts commit b8ba550.

This reverts commit e6fbed0.

This reverts commit 43806ad.

…nerated builds" This reverts commit 090f8a1.

github-actions · 2026-06-02T23:18:03Z

⚠️ Regressions detected on macos-14 for commit `7601b2a`

Regression Summary

ERR: 1

jmarrec · 2026-06-03T20:52:34Z

I just tried this in debug and here is a place that it seems to have a problem:

Penumbra tests on arm64 debug have always been problematic for what it's worth.

mitchute

Assuming this passes the other builds, I'm done. @jmarrec the commit history on this is a disaster. Do we do a squash merge, or port it to a fresh branch and then just merge that?

mitchute · 2026-06-04T20:52:58Z

+    if(APPLE AND CMAKE_GENERATOR STREQUAL "Xcode")
+      set(_gtest_discovery_mode PRE_TEST)
+    else()
+      set(_gtest_discovery_mode POST_BUILD)
+    endif()


IIRC, this was due to something related to Xcode and code signing, which we are not doing during development builds.

mitchute · 2026-06-04T20:54:36Z


 add_executable(generate_embeddable_epJSON_schema embedded/generate_embeddable_epJSON_schema.cpp)
 target_link_libraries(generate_embeddable_epJSON_schema PRIVATE project_options project_fp_options project_warnings fmt::fmt)
+target_compile_options(generate_embeddable_epJSON_schema PRIVATE $<$<COMPILE_LANG_AND_ID:CXX,Clang,AppleClang>:-Wno-shorten-64-to-32>)


Turns this into a warning. I think this should just apply to AppleClang and I thought I changed it to that everywhere else, but now I see that I missed this one. I'll probably just leave it.

mitchute · 2026-06-04T20:56:14Z


 target_include_directories(airflownetworklib PUBLIC ${CMAKE_CURRENT_SOURCE_DIR}/include)
-target_link_libraries(airflownetworklib PUBLIC btwxt fmt::fmt)
+target_link_libraries(airflownetworklib PUBLIC btwxt fmt::fmt PRIVATE nlohmann_json)


My machine was complaining about the nlohman json.hpp #include <doj/alphanum.hpp>. This fixed it.

mitchute · 2026-06-04T20:57:13Z

 }
 #  else
-constexpr auto operator"" _a(const char* s, size_t) -> detail::udl_arg<char> {
+constexpr auto operator""_a(const char* s, size_t) -> detail::udl_arg<char> {


There were a couple random things like this that it didn't like.

mitchute · 2026-06-04T20:58:13Z

+          $<$<COMPILE_LANG_AND_ID:CXX,Clang>:-Wno-character-conversion>
+          $<$<COMPILE_LANG_AND_ID:CXX,Clang,AppleClang>:-Wno-deprecated-declarations>


Applies these based on what's available in each compiler.

mitchute · 2026-06-04T21:00:13Z

+if(APPLE AND CMAKE_GENERATOR STREQUAL "Xcode" AND OPENGL_FOUND)
+  set(OPENGL_FOUND FALSE)
+  message(STATUS "Disabling OpenGL for Xcode-generated builds to avoid Cocoa window creation during build/test execution")
+endif()


Penumbra want to create some windows behind the scenes when OpenGL is enabled, which it is by default. We could probably have worked through that, but since this isn't likely to be used by many of our developers, I figured we'd just turn it off when running from an Xcode generated build.

mitchute · 2026-06-05T16:19:59Z

Thanks @RKStrand for bearing with me to test this, and thanks @jmarrec for your review and contributions. Merging.

suppress implicit 64-to-32 bit conversion errors for xcode generated …

090f8a1

…builds

mitchute requested a review from RKStrand May 27, 2026 00:25

mitchute added the DoNotPublish Includes changes that shouldn't be reported in the changelog label May 27, 2026

mitchute added 3 commits May 27, 2026 13:55

possible changes for MacOS 26.5

43806ad

change from implicit to explicit cast

e6fbed0

guard against non-existant compiler flags

b8ba550

mitchute changed the title ~~Suppress Implicit 64-to-32 bit Conversion Errors for Xcode Generated Builds~~ Compiler Tweaks for macOS 26 May 27, 2026

mitchute changed the title ~~Compiler Tweaks for macOS 26~~ Tweaks for building with macOS 26 May 27, 2026

Build on Apple clang 21

a2a1fb5

jmarrec reviewed May 29, 2026

View reviewed changes

resolve some implicit conversion issues on xcode generated builds

f8bea8c

silence errors on Apple Clang, plus a few changes needed for Xcode ge…

b941b24

…nerated builds

patch gtest to avoid implicit conversion

aa93637

mitchute added 4 commits June 2, 2026 13:36

disable tests requiring OpenGL when on Apple

636a3f4

Merge branch 'develop' into build_apple-clang-21_msm

34071b3

Revert "guard against non-existant compiler flags"

c8f060a

This reverts commit b8ba550.

Revert "change from implicit to explicit cast"

c00475e

This reverts commit e6fbed0.

mitchute added 4 commits June 2, 2026 14:37

Revert "possible changes for MacOS 26.5"

54e2c79

This reverts commit 43806ad.

Revert "suppress implicit 64-to-32 bit conversion errors for xcode ge…

f7695de

…nerated builds" This reverts commit 090f8a1.

Merge branch 'build_apple-clang-21_msm' into xcode-generated-builds

dd0e3f1

compiler flags

c665b5b

fun fun fun

d226831

use a generator expression to simplify

1d76a07

mitchute commented Jun 4, 2026

View reviewed changes

mitchute merged commit df920a0 into develop Jun 5, 2026
11 checks passed

mitchute deleted the xcode-generated-builds branch June 5, 2026 16:20

		$<$<COMPILE_LANG_AND_ID:CXX,Clang>:-Wno-character-conversion>
		$<$<COMPILE_LANG_AND_ID:CXX,Clang,AppleClang>:-Wno-deprecated-declarations>

Conversation

mitchute commented May 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Pull request overview

Description of the purpose of this PR

Pull Request Author

Reviewer

Uh oh!

jmarrec left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mitchute commented Jun 1, 2026

Uh oh!

RKStrand commented Jun 1, 2026

Uh oh!

mitchute commented Jun 1, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

RKStrand commented Jun 1, 2026

Uh oh!

RKStrand commented Jun 1, 2026

Uh oh!

mitchute commented Jun 1, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

RKStrand commented Jun 1, 2026

Uh oh!

mitchute commented Jun 1, 2026

Uh oh!

mitchute commented Jun 2, 2026

Uh oh!

RKStrand commented Jun 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mitchute commented Jun 2, 2026

Uh oh!

mitchute commented Jun 2, 2026

Uh oh!

RKStrand commented Jun 2, 2026

Uh oh!

github-actions Bot commented Jun 2, 2026

⚠️ Regressions detected on macos-14 for commit 7601b2a

Uh oh!

jmarrec commented Jun 3, 2026

Uh oh!

mitchute left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mitchute commented Jun 5, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

mitchute commented May 27, 2026 •

edited

Loading

mitchute commented Jun 1, 2026 •

edited

Loading

mitchute commented Jun 1, 2026 •

edited

Loading

RKStrand commented Jun 2, 2026 •

edited

Loading

⚠️ Regressions detected on macos-14 for commit `7601b2a`