-
-
Couldn't load subscription status.
- Fork 5.9k
Update to trusty (2nd try) #1897
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
|
isn't that a leak in the X libraries? |
|
If so, only we can do is disabling this test? |
|
valgrind always reports leaks in X libraries. It's probably related to their own way of allocating memory. Nothing we can do about it, ignore.... |
|
This link explains how to suppress reporting of some leaks with lsan: |
758ce51 to
c7ddf98
Compare
|
It seems that the leaks in X libraries were successfully suppressed, however, another problem occurred. |
|
that is seems to be a problem on the Mac builds. Perhaps because it is linking against the wrong ruby headers? I don't have a mac, so not sure. Also unfortunately, we don't see the leak logfile because of this: There is one remaining problem on the linux system: the po check script aborts somehow: Unfortunately, we don't see the error here... |
|
Sorry, I mistakenly updated the build matrix. ASAN should have been skipped on macOS. Latest log: |
|
https://travis-ci.org/k-takata/vim/jobs/260685607#L9852 Now the leaks in X libraries were successfully suppressed, but another leak detected. |
|
Hm, diff --git a/.travis.yml b/.travis.yml
index 17f041807..9a0b44dfa 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -46,7 +46,7 @@ matrix:
FEATURES=huge SRCDIR=./src CHECK_AUTOCONF=no ASAN_OPTIONS="print_stacktrace=1 log_path=asan"
"CONFOPT='--enable-perlinterp --enable-pythoninterp --enable-rubyinterp --enable-luainterp'"
- os: linux
- compiler: gcc
+ compiler: clang
env: BUILD=yes TEST=test SANITIZER_CFLAGS="-g -O1 -DABORT_ON_INTERNAL_ERROR -DEXITFREE -fsanitize=address -fno-omit-frame-pointer"
FEATURES=huge SRCDIR=./src CHECK_AUTOCONF=no ASAN_OPTIONS="print_stacktrace=1 log_path=asan"
"CONFOPT='--enable-perlinterp --enable-pythoninterp --enable-rubyinterp --enable-luainterp'" |
Codecov Report
@@ Coverage Diff @@
## master #1897 +/- ##
==========================================
- Coverage 75.08% 74.78% -0.31%
==========================================
Files 76 77 +1
Lines 125301 125674 +373
Branches 0 28120 +28120
==========================================
- Hits 94088 93988 -100
- Misses 31213 31686 +473
Continue to review full report at Codecov.
|
|
did ASAN with gcc work? |
|
Yes, it worked. However, the coverage decreased about 4% on coveralls and about 1% on codecov. |
|
Hm, I believe this is a problem of coverage and clang: I googled this message: I don't know why we use cpp-coveralls=0.3.2, can you try with the latest version cpp-coveralls=0.4.0? |
Also show if_xx version only if +eval. Update FileList.
|
The gcov version mismatch was solved by using llvm-cov instead of gcov when compiler is clang. if_python3 didn't worked because python 3.5 which was installed by pyenv was used. I added scripts for showing all if_xx versions: The coverage still decreases a bit, but I suppose it is because of compiler version or something, and I don't think there are anything we can do for this. I think this PR is ready to be merged. |
It seems that #1884 failed because
libgtk2.0-devwas missing.If
libgtk2.0-devis installed, almost all jobs passes: https://travis-ci.org/k-takata/vim/builds/259091225But still one job fails: https://travis-ci.org/k-takata/vim/jobs/259091233
Any ideas?