Changeset View
Standalone View
swh/perfecthash/Makefile
- This file was added.
| CFLAGS=-DHASH_DEBUG -Wall -I../.. -g -std=c++17 -fprofile-arcs -ftest-coverage | ||||||||||
douardda: why the `-I ../..` ? | ||||||||||
Done Inline ActionsSo that #include "swh/perfecthash/hash.h" works. dachary: So that #include "swh/perfecthash/hash.h" works. | ||||||||||
| LDFLAGS=-lcmph -lgtest -lpthread -lstdc++ -lstdc++fs -fprofile-arcs -ftest-coverage | ||||||||||
| hash: hash.o test_hash.o | ||||||||||
Done Inline Actionsthis hash tool is only built for testing purpose right? if so why not name it test_hash? douardda: this `hash` tool is only built for testing purpose right? if so why not name it `test_hash`? | ||||||||||
Done Inline ActionsI'll change that. dachary: I'll change that. | ||||||||||
| $(CXX) -o $@ $^ $(LDFLAGS) | ||||||||||
| hash.c: hash.h | ||||||||||
| test_hash.cpp: hash.h | ||||||||||
| format: | ||||||||||
| clang-format -i hash.c hash.h test_hash.cpp | ||||||||||
Done Inline Actionsno idea how hard that is and how relevant that'd be... (not a blocking point at all, just mentioning it) ardumont: no idea how hard that is and how relevant that'd be...
Still, might be this needs to be moved… | ||||||||||
Done Inline ActionsI don't see that it will be super complicated. But it's going to require the installation of non-python dependencies for everyone. I suppose it deserves a discussion? dachary: I don't see that it will be super complicated. But it's going to require the installation of… | ||||||||||
| check: hash | ||||||||||
| valgrind --leak-check=full --tool=memcheck ./hash | ||||||||||
| lcov -d . -c -o hash.lcov | ||||||||||
| rm -fr html ; genhtml -o html hash.lcov | ||||||||||
| firefox html/perfecthash/hash.c.gcov.html | ||||||||||
Done Inline Actions
maybe, eventually with a variable defined top-level in the makefile defaulting to firefox even? (I have such env variable in my env but it's been configured for so long, i don't recall whether that's standard or not) ardumont: maybe, eventually with a variable defined top-level in the makefile defaulting to firefox even? | ||||||||||
Done Inline Actionsmy main problem for this is its inability to play nicely when executed in ci/debbuild environment. douardda: my main problem for this is its inability to play nicely when executed in ci/debbuild… | ||||||||||
Done Inline Actionsoh yeah, when i wrote that, i assumed this target was for dev only, not ci. ardumont: oh yeah, when i wrote that, i assumed this target was for dev only, not ci. | ||||||||||
Done Inline ActionsFor the record the contentious line was removed: it is entirely unecessary whatever the context. dachary: For the record the contentious line was removed: it is entirely unecessary whatever the context. | ||||||||||
Done Inline ActionsThe html generation should be stripped for CI. I'm not sure how to make this particular test fit in the CI at the moment. dachary: The html generation should be stripped for CI. I'm not sure how to make this particular test… | ||||||||||
| clean: | ||||||||||
| rm -f *.o hash | ||||||||||
why the -I ../.. ?