Hello Nilesh, thanks for the upload right! > Had a quick look, it looks nice. However would you consider fixing > blhc which seems to fail salsa CI? > IMO it'd be straightforward to do a patch here > > https://salsa.debian.org/science-team/freefempp/-/jobs/1809852 I think CI blhc detection is wrong because freefem++ use a script called "ff-c++" to build the source instead of calling directly g++, which triggers false positive. The ff-c++ script actually calls g++ with the correct flags. For example, blhc detects this line as wrong: eval ./ff-c++ NewSolver.cpp -lumfpack -lamd -lcholmod -lklu - I/usr/include/suitesparse -lumfpack -lamd -lcholmod -lklu - I/usr/include/suitesparse -llapack -lblas but in the build-log we can check that ff-c++ calls g++ with these options, which are fine: g++ -c -fPIC -g -ffile-prefix- map=/home/francois/dev/freefem++/freefempp=. -fstack-protector-strong - Wformat -Werror=format-security -DNDEBUG -O3 -mmmx -mavx -std=c++14 - DBAMG_LONG_LONG -DNCHECKPTR -fPIC '-I./include' '- I/usr/include/suitesparse' '-I/usr/include/suitesparse' -g -ffile- prefix-map=/home/francois/dev/freefem++/freefempp=. -fstack-protector- strong -Wformat -Werror=format-security -DNDEBUG -O3 -mmmx -mavx - std=c++14 -DBAMG_LONG_LONG -DNCHECKPTR -fPIC -Wdate-time - D_FORTIFY_SOURCE=2 -I/usr/include/suitesparse - I/usr/include/hdf5/serial -I/usr/include 'NewSolver.cpp' -Wl,-z,relro - Wl,-z,now -rdynamic Do you have a patch in mind to discard these false positive cases? Thanks, François
Attachment:
signature.asc
Description: This is a digitally signed message part