fix: replace blanket -Wno-maybe-uninitialized with targeted pragmas (#979)#981
fix: replace blanket -Wno-maybe-uninitialized with targeted pragmas (#979)#981alandefreitas wants to merge 3 commits intoboostorg:developfrom
Conversation
GCC 7+ at -O3 emits false -Wmaybe-uninitialized from variant2 union storage when deeply inlining constexpr functions. Wrap the affected parse_sequence code in tuple_rule.hpp with a diagnostic pragma. fix boostorg#979
|
An automated preview of the documentation is available at https://981.url.prtest2.cppalliance.org/index.html If more commits are pushed to the pull request, the docs will rebuild at the same URL. 2026-02-24 21:55:26 UTC |
e16edf1 to
6d3338e
Compare
|
GCOVR code coverage report https://981.url.prtest2.cppalliance.org/gcovr/index.html Build time: 2026-02-24 17:19:12 UTC |
|
GCOVR code coverage report https://981.url.prtest2.cppalliance.org/gcovr/index.html Build time: 2026-02-24 22:02:59 UTC |
6d3338e to
d0f686e
Compare
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## develop #981 +/- ##
========================================
Coverage 99.15% 99.15%
========================================
Files 155 155
Lines 10045 10045
========================================
Hits 9960 9960
Misses 85 85
Continue to review full report in Codecov by Sentry.
🚀 New features to boost your workflow:
|
Remove the blanket
-Wno-maybe-uninitializedsuppression from all GCC builds and add a scoped pragma around the one site in our code that triggers the false positive (parse_sequenceintuple_rule.hpp). Also removes the unusedBOOST_URL_NO_INLINEmacro.The remaining false positives originate inside
variant2andoptionalheaders and need to be fixed upstream:The third commit is a temporary CI patch that copies fixed versions of those headers to prove the full fix works across GCC 7–15. It will be dropped once the upstream PRs are merged.