Update bind9 CI test to use meson #2562
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #2562 +/- ##
==========================================
+ Coverage 78.73% 78.74% +0.01%
==========================================
Files 645 645
Lines 110658 110658
Branches 15648 15648
==========================================
+ Hits 87124 87137 +13
+ Misses 22834 22820 -14
- Partials 700 701 +1 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
| @@ -53,6 +59,7 @@ ls | |||
|
|
|||
| aws_lc_build ${SRC_ROOT} ${AWS_LC_BUILD_FOLDER} ${AWS_LC_INSTALL_FOLDER} -DBUILD_TESTING=OFF -DBUILD_TOOL=OFF -DBUILD_SHARED_LIBS=1 | |||
| export LD_LIBRARY_PATH="${AWS_LC_INSTALL_FOLDER}/lib" | |||
| export LD_PRELOAD=libjemalloc.so.2 | |||
There was a problem hiding this comment.
LD_PRELOAD represents a path. is there a jemalloc lib in BIND9_SRC_FOLDER? if so, why do we need it?
There was a problem hiding this comment.
There's no jemalloc lib in BIND9_SRC_FOLDER. I added it because I ran into a segmentation fault, which came from a memory allocator conflict between bind9 and aws-lc. Adding LD_PRELOAD solves the issue by forcing both to use the same memory allocator at runtime. But to be completely honest I don't like this approach too much.
We didn't have this problem building with jemalloc when bind9 was still using automake so I suspect it's probably due to how meson is set up in bind9.
Another option is to disable jemalloc altogether. Does that seem like a better way to go?
There was a problem hiding this comment.
I see. This work-around effectively disables jemalloc by pointing LD_PRELOAD to a non-existent file. Doing that explicitly in the build configuration seems like a better way to go.
Issues:
Resolves P257046914
Description of changes:
Bind9 recently switched from automake to meson for its build system, leading to broken CI test on our end. This PR fixes that.
Callouts:
I disabled jemalloc for bind9 build because I ran into memory allocator conflict with jemalloc. We didn't have this problem when bind9 was still using automake so I suspect it's probably due to how meson is set up in bind9. Disabling jemalloc shouldn't affect the test outcome.
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license and the ISC license.