[RFC PATCH 0/4] [PIDNS] Trie and pidns-cache tests

[Ákos Uzonyi]

On Mon, 24 Aug 2020 at 20:07, Dmitry V. Levin <ldv at altlinux.org> wrote:
> On Mon, Aug 24, 2020 at 08:00:27PM +0200, Ákos Uzonyi wrote:
> > On Sun, 23 Aug 2020 at 23:55, Dmitry V. Levin <ldv at altlinux.org> wrote:
> > > On Sun, Aug 23, 2020 at 11:30:37PM +0200, Ákos Uzonyi wrote:
> > > > On Sun, 23 Aug 2020 at 00:36, Dmitry V. Levin <ldv at altlinux.org> wrote:
> > > > > OK, I've found a way to work this around and squeeze some coverage
> > > > > out of the test, feel free to squash it in:
> > > > >
> > > > > diff --git a/Makefile.am b/Makefile.am
> > > > > index f4bd0e0d9..f79d93437 100644
> > > > > --- a/Makefile.am
> > > > > +++ b/Makefile.am
> > > > > @@ -409,7 +409,7 @@ endif
> > > > >  CODE_COVERAGE_BRANCH_COVERAGE = 1
> > > > >  CODE_COVERAGE_GENHTML_OPTIONS = $(CODE_COVERAGE_GENHTML_OPTIONS_DEFAULT) \
> > > > >         --prefix $(shell cd $(abs_top_srcdir)/.. && pwd || echo .)
> > > > > -CODE_COVERAGE_IGNORE_PATTERN = '/usr/include/*'
> > > > > +CODE_COVERAGE_IGNORE_PATTERN = '/usr/include/*' '*/tests/*' '*/tests-m32/*' '*/tests-mx32/*'
> > > > >  strace_CPPFLAGS += $(CODE_COVERAGE_CPPFLAGS)
> > > > >  strace_CFLAGS += $(CODE_COVERAGE_CFLAGS)
> > > > >  strace_LDADD += $(CODE_COVERAGE_LIBS)
> > > > > diff --git a/tests/Makefile.am b/tests/Makefile.am
> > > > > index 39d948cf4..56e0e8eea 100644
> > > > > --- a/tests/Makefile.am
> > > > > +++ b/tests/Makefile.am
> > > > > @@ -66,7 +66,6 @@ libtests_a_SOURCES = \
> > > > >         test_ucopy.h \
> > > > >         tests.h \
> > > > >         tprintf.c \
> > > > > -       trie.c \
> > > > >         # end of libtests_a_SOURCES
> > > > >  libtests_a_CPPFLAGS = $(AM_CPPFLAGS) -D_FILE_OFFSET_BITS=64
> > > > >  check_LIBRARIES = libtests.a
> > > > > @@ -308,6 +307,11 @@ stack_fcall_mangled_SOURCES = stack-fcall-mangled.c \
> > > > >         stack-fcall-mangled-0.c stack-fcall-mangled-1.c \
> > > > >         stack-fcall-mangled-2.c stack-fcall-mangled-3.c
> > > > >
> > > > > +trie_test_SOURCES = trie_test.c trie_for_tests.c
> > > > > +trie_test_CPPFLAGS = $(AM_CPPFLAGS) $(CODE_COVERAGE_CPPFLAGS)
> > > > > +trie_test_CFLAGS = $(AM_CFLAGS) $(CODE_COVERAGE_CFLAGS)
> > > > > +trie_test_LDADD = $(LDADD) $(CODE_COVERAGE_LIBS)
> > > > > +
> > > > >  include gen_tests.am
> > > > >
> > > > >  if ENABLE_STACKTRACE
> > > > > diff --git a/tests/trie.c b/tests/trie.c
> > > > > deleted file mode 120000
> > > > > index 0dd9cab45..000000000
> > > > > --- a/tests/trie.c
> > > > > +++ /dev/null
> > > > > @@ -1 +0,0 @@
> > > > > -../trie.c
> > > > > \ No newline at end of file
> > > > > diff --git a/tests/trie_for_tests.c b/tests/trie_for_tests.c
> > > > > new file mode 100644
> > > > > index 000000000..13e564c4f
> > > > > --- /dev/null
> > > > > +++ b/tests/trie_for_tests.c
> > > > > @@ -0,0 +1 @@
> > > > > +#include "trie.c"
> > > >
> > > > I didn't know very well how this code coverage is configured, so thank
> > > > you for helping with this.
> > >
> > > The idea was to compile tests/trie_for_tests.c (which is trie.c) along
> > > with the test with coverage options.  For this reason I had to take
> > > tests/trie_for_tests.c out of libtests_a_SOURCES and convert it from a
> > > symlink into a real file so that the original trie.c is covered instead of
> > > a symlinked file.
> >
> > Yes, after reading the patch, I understood the main concept, but it
> > would have taken a lot of time to come up with this myself :).
> >
> > However the codecov result [1] of my last commit (which includes the
> > patch above) still shows that trie_free is not covered. I'm not sure
> > what's causing this, as trie_free is called in trie_test.
> >
> > [1] https://codecov.io/gh/AkosUzonyi/strace/src/6b0f5cf5aa6ac8473e64a1f9d7f9430a8356e410/trie.c
>
> Hmm, locally lcov says that 100% of trie.c is covered now.

OK, fine, then it's just some problem in codecov.