newbie having an issue installing and running a test with cppyy

[SimonDev666]

hello, just started using cppyy again after a year or so. as per you use guide the following:

running on kali linux (latest version) ensured python3.12-dev was installed

$ python -m venv WORK $ source WORK/bin/activate (WORK) $ python -m pip install cppyy (WORK) $

created a simple python file in VS code in WORK 'Test_cppyy01.py'. just contains 'import cppyy'

decided to try from out VS code, so from terminal went into WORK and could see simple python file. ran file: 'python Test_cppyy01.py'

(Re-)building pre-compiled headers (options: -O2 -march=native); this may take a minute ... In file included from input_line_3:2: In file included from /usr/include/c++/14/string:42: In file included from /usr/include/c++/14/bits/char_traits.h:57: In file included from /usr/include/c++/14/bits/stl_construct.h:61: In file included from /usr/include/c++/14/bits/stl_iterator_base_types.h:71: /usr/include/c++/14/bits/iterator_concepts.h:1013:13: error: no matching function for call to 'begin' = decltype(ranges::access::__begin(std::declval<_Tp&>()));

100's of more error lines

any chance of telling me what i am doing wrong.

thanks Simon

[SimonDev666]

hello

seemed to have found a fix from your doc:

export CLING_STANDARD_PCH="none"

do or should i use pch headers?? if i should, how do i make or use them for all c++ generic??

[wlav]

Setting it to none means the headers will be read in. Ordinarily, this would only be used if you have an alternate (e.g. modules), it typically isn't a fix if there's an issue with the headers, as they're still read in dynamically. In this particular case, it may be that there's no #include <string> (explicit or implicit) in the test script, which removes the problem but only until you do. In other scenarios, setting the envar might be a fix if there's an issue with the installed PCH, but your script is rebuilding it already, so that can't be either.

As for the error itself, I think it's the limited support in Clang13 for concepts. Using EXTRA_CLING_ARGS='-O2 -march=native -std=c++17' would downgrade the C++ version and pre-empt it. (Work is on-going on Clang16 and Clang18 shortly after.)

[SimonDev666]

given the EXTRA_CLING_ARGS setting mentioned above, do I have download the source and rebuild the entire project?? what would command be, or is it as in your doc 'Install from Source' + the above EXTRA*

[SimonDev666]

i entered the following 'EXTRA_CLING_ARGS='-O2 -march=native -std=c++17'' in the VS Code python termainal under the activate venv and got no errors. so i may have answered my own above question

[wlav]

Yes, it's a run-time thing. PCH may need to be rebuild whenever switching args, but that's it. Aside, I pushed the first changes for LLVM16 yesterday; hopefully be done soon, after which problems like this should (famous last words) go away.

[SimonDev666]

i deleted my virtual env for python, did a complete apt-get update and upgrade and repeated all of the from the prev comments: " python -m venv WORK source ./WORK/bin/activate python -m pip install cppyy "

set these as from your doc and prev comments to me: " export CLING_STANDARD_PCH="none" set EXTRA_CLING_ARGS="-O2 -march=native -std=c++17" "

no run my code, it runs but now i get run-time errors

In file included from input_line_3:2: In file included from /usr/include/c++/14/string:42: In file included from /usr/include/c++/14/bits/char_traits.h:57: In file included from /usr/include/c++/14/bits/stl_construct.h:61: In file included from /usr/include/c++/14/bits/stl_iterator_base_types.h:71: /usr/include/c++/14/bits/iterator_concepts.h:1013:13: error: no matching function for call to 'begin' = decltype(ranges::access::begin(std::declval<_Tp&>())); ^~~~~~~~~ ..... /usr/include/c++/14/bits/iterator_concepts.h:989:19: note: because 'decay_copy(begin(t))' would be invalid: no matching function for call to 'begin' { decay_copy(begin(__t)) } -> input_or_output_iterator; ^ fatal error: too many errors emitted, stopping now [-ferror-limit=] Error: ./bin/rootcling: compilation failure (/tmp/allDictff783ed64c_dictUmbrella.h) /home/simon/Desktop/PythonVEnv/VirtEnv/lib/python3.12/site-packages/cppyy_backend/loader.py:139: UserWarning: No precompiled header available (failed to build); this may impact performance. warnings.warn('No precompiled header available (%s); this may impact performance.' % msg) ['/home/simon/Desktop/PythonVEnv/VirtEnv/lib/python3.12/site-packages/', '/home/simon/Desktop/Code/QT/Test_EmbedPython/Test_001_EmbedPython/', '/usr/lib/python312.zip', '/usr/lib/python3.12', '/usr/lib/python3.12/lib-dynload', '/usr/local/lib/python3.12/dist-packages', '/usr/lib/python3/dist-packages', '/usr/lib/python3.12/dist-packages'] (Re-)building pre-compiled headers (options: -O2 -march=native); this may take a minute ... here there done "

no matter how many times i run my code, i get the above errors, which i did not get before re-doing the current python virt env.

[wlav]

It doesn't look like these:

export CLING_STANDARD_PCH="none"
set EXTRA_CLING_ARGS="-O2 -march=native -std=c++17"

were actually set and/or propagated (not sure about the change in set/export either). Can you check on the Python prompt or within your script before importing cppyy what the values are of os.environ['CLING_STANDARD_PCH'] and os.environ['EXTRA_CLING_ARGS']?

As-is, if the standard pch is none, it shouldn't be rebuild and if the second is set (and rebuilding does happen), the values should appear in this line: options: -O2 -march=native as the values set instead of the defaults (as you have).

I'd add that if the only point is to change the used standard, I'd set STDCXX=17 on installation time, or force rebuild the PCH once (by setting the CLING_REBUILD_PCH once). Doing so will "bake in" the standard.