-
Notifications
You must be signed in to change notification settings - Fork 62
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
cnl_example fails #887
Comments
The loss of bintray appears to prevent the unit tests as well. |
Hi @josmithiii. Which tests do you mean? |
Thanks for your reply! I'm trying to run the cnl tests:
|
Please give me a few days to look into this. I'm currently en route to CppOnSea and don't have a lot of bandwidth right now. |
Sorry for the long delay. As well as the conference running a week longer than expected(!), I'm facing some disruption at home due to house renovations. I'm not sure your issue is the same as #887. I'm not able to complete the repro steps above and when I work around them, I am not able to reproduce the issue. However, I'm very keen to help resolve the issue and adapt the project as necessary to avoid future problems. Therefore, could you file a separate bug using the bug template. Could you make sure that repro steps are from a system without a ~/.conan directory for maximum reproducibility. I recommend you avoid the Can you try and remove as many steps as possible that aren't necessary to reproduce the error? For reference, here's the repro steps I followed to try and match your own: # using llvm-14 as the default C and C++ compiler...
rm -rf ~/.conan/ ../build/*
CXX=/usr/bin/clang++ conan profile new --detect default
conan install --settings compiler.libcxx=libstdc++11 --build=missing --options test=unit ..
conan build .. |
Thanks very much! I'm getting much farther now. I've attached my slightly
modified version of your "repro steps", as well as its output.
The current problem seems to be this:
In file included from /Users/jos/w/cnl/test/unit/scaled_int/numbers.cpp:7:
/Users/jos/w/cnl/include/cnl/_impl/scaled_integer/../../_impl/scaled_integer/numbers.h:127:17:
error: constexpr variable 'pi_v<cnl::_impl::wrapper<unsigned long long,
cnl::power<-62, 2>>>' must be initialized by a constant expression
pi_v<cnl::scaled_integer<Rep, cnl::power<Exponent>>>{
^ ~
|
What version of Clang are you using? |
I see this on my Mac:
$ which clang
/usr/bin/clang
$ clang -v
Apple clang version 13.1.6 (clang-1316.0.21.2.5)
Target: arm64-apple-darwin21.5.0
Thread model: posix
InstalledDir:
/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin
|
- appears to take too many iterations on Apple Clang - reported under #887 but is unrelated
You don't appear to be using the clang profile and some tests are so complex that compiler limits on constant expressions are hit. I've tried hiding that particular test in the case there is no CNL testing profile in use but recommend using test/toolchain/clang.cmake or test/toolchain/clang-libc++.cmake for clang builds in order to enable all tests. Please can you try branch, osx-clang-fix and see if that gets you past the error? |
Near perfection!: index 2527f010..2d6a6747 100644
--- a/test/unit/scaled_int/elastic/elastic_scaled_int.cpp
+++ b/test/unit/scaled_int/elastic/elastic_scaled_int.cpp
@@ -187,7 +187,7 @@ namespace test_fraction_specific_long {
static_assert(specific > .333333333333333);
static_assert(specific < .333333333333334);
#else
- static_assert(identical(cnl::elastic_scaled_integer<63, cnl::power<-60>>{1.L / 3}, specific));
+ // static_assert(identical(cnl::elastic_scaled_integer<63, cnl::power<-60>>{1.L / 3}, specific));
#endif
} I am on a 64-bit ARM (Apple M1 Max). |
- appears to take too many iterations on Apple Clang - reported under #887 but is unrelated
Please, if possible could you create an issue with reproducible steps? Better still, do you have a branch in which you make an ARM build for OSX? I've tried setting |
This reproduces it on Mac OSX ARM just using the latest
It makes it to 79% and then prints
After that I see
Are you saying the above should go in a new issue? |
Partly due to loss of Bintray, but also because it's not tested, cnl_example doesn't build.
The text was updated successfully, but these errors were encountered: