Ondřej Čertík

icon indicating a website link https://ondrejcertik.com/ icon indicating an email [email protected]

icon company Microsoft icon location Los Alamos, NM icon location The original author of the @sympy and @symengine libraries and the @lfortran and LPython (@lcompilers) compilers. Co-founder of the @fortran-lang organization

Results 1764 comments of Ondřej Čertík

Fix implicit interface type check for FORTRAN 77 sequence association

I tried to fix it in https://github.com/lfortran/lfortran/pull/8914 and it fixes this particular issue, but it breaks lapack_02.f90 test. So more work is needed. The basic idea is this: ASR must...

Sync to latest LF

This looks good, thanks for doing the merge!

Debugging LSP bug

• Hang Analysis - test_document_highlight issues textDocument/documentHighlight (request id 2) at test_document_highlight-parallel-client.log:69-82. No response ever arrives; instead the server keeps emitting telemetry each second. outgoingMessages’s seen counter advances (telemetry only)...

Debugging LSP bug

• Key log from the latest run is test_whole_document_formatting-parallel- client.log. It shows: - At 01:13:40.314Z the client sent textDocument/formatting (request id 2) for tests/function_call1.f90. - No response for that id...

Debugging LSP bug

• The latest hang is again in test_document_highlight. From test_document_highlight-parallel-client.log: - The client opened examples/expr2.f90, waited for diagnostics, then sent textDocument/documentHighlight (request id 2) at 06:34:10Z. - No response ever...

Segmentation fault in PRIMA on LLVM 14

This one might be related, LLVM 8: ``` 2025-11-18T09:41:27.2660596Z The corresponding X is: 2025-11-18T09:41:27.2661108Z 1.0000000000000000E+000 1.0000000000000000E+000 3.6160784864589629E-001 2025-11-18T09:41:27.2661742Z 1.0000000000000000E+000 -3.6160789815759925E-001 1.0000000000000000E+000 2025-11-18T09:41:27.2662452Z -1.0000000000000000E+000 1.0000000000000000E+000 -1.0000000000000000E+000 2025-11-18T09:41:27.2663227Z -8.7819855787179010E-009 -1.0000000000000000E+000 -1.0000000000000000E+000 2025-11-18T09:41:27.2663941Z...

Segmentation fault in PRIMA on LLVM 14

Another one: https://github.com/lfortran/lfortran/actions/runs/20323735436/job/58384605530 ``` +++ lfortran --separate-compilation checkexit.f90.o consts.F90.o datetime.o debug.F90.o evaluate.f90.o fprint.f90.o geometry.f90.o history.f90.o huge.F90.o inf.F90.o infnan.F90.o infos.f90.o initialize.f90.o linalg.f90.o memory.F90.o message.f90.o noise.o param.o pintrf.f90.o powalg.f90.o preproc.f90.o prob.o rand.o...

Segmentation fault in PRIMA on LLVM 14

This one is interesting: https://github.com/lfortran/lfortran/actions/runs/20323788219/job/58384750173 ``` +++ lfortran --std=f23 --cpp -DPRIMA_DEBUGGING=0 -DPRIMA_AGGRESSIVE_OPTIONS=0 -DPRIMA_INTEGER_KIND=4 -DPRIMA_REAL_PRECISION=32 -DPRIMA_QP_AVAILABLE=0 '-DPRIMA_TESTDIM='\''small'\''' ./tests/test.F90 -I../build/fortran -I./ The random seed is 2551 The test dimension is small chebyquad:...

CI: array_constructor_02_STD_F23 fails

And a similar error in: https://github.com/lfortran/lfortran/actions/runs/19913339457/job/57086799495 ``` 1160/1958 Test #1161: derived_types_79_STD_F23 ...................................Subprocess aborted***Exception: 0.02 sec ``` Also macOS latest LLVM 21.

Supporting SymEngine through xeus-cpp-lite

I think we should definitely include whatever is needed to run symengine interactively, and then have links in our README and a website to JupyterLite so that people can try...