Please make the following clarifications in the test plan and SRU "where problems could occur":
a) Use the scripts as planned in the original test plan, but extend them to all architectures, because we don't want frame pointers in s390x and ppc64el, only arm64 and amd64 if I'm not mistaken (not on 32bits). b) We need some form of smoke test to make sure the package still runs. Here it's a library, so it will have to be done with its reverse dependencies. Perhaps enough reverse dep8 tests are triggered? c) If (a) cannot be done for some reason, or perhaps as an improvement on (a), check the build logs for each architecture and look for the frame pointer flags being present and absent where required. d) Please verify the diff in the compiler flags. I'm not sure if my comment #3 was exhaustive, please verify. If it's correct, then that diff from comment #3 suffices. e) In particular, we are looking for "problematic" flags now being enabled. For example, and I didn't spot it in this build (but I could be wrong, as I just did a quick check), the "-Wl,-Bsymbolic-functions" linker flag is known to have caused problems with libraries and plugins in the past (like in autofs[1][2]). That's the kind of regression we really have to be careful to avoid, and detect early. 1. https://bugs.launchpad.net/ubuntu/+source/autofs/+bug/1470687 2. https://bugs.launchpad.net/ubuntu/+source/autofs/+bug/1472115 -- You received this bug notification because you are a member of Ubuntu Touch seeded packages, which is subscribed to nspr in Ubuntu. https://bugs.launchpad.net/bugs/2084460 Title: [SRU] Please enable frame-pointer on Noble Status in nspr package in Ubuntu: Incomplete Bug description: [ Impact ] In order to help developers with profiling, Ubuntu enabled frame- pointers for packages in the main archive. https://ubuntu.com/blog/ubuntu-performance-engineering-with-frame- pointers-by-default The `Affects` packages were missed due to the packaging issues. These have been fixed on Oracular and can now be ported to Noble. [ Test Plan ] Look at the DW_AT_producer in the associated debug ELF file for flags. -fno-omit-frame-pointer should be present after the fix. The following scripts can be used as a reference https://gist.github.com/julian-klode/85e55553f85c410a1b856a93dce77208 https://gist.github.com/julian-klode/95818246eaef0ac6a54588f7f368e25c The nspr package comes with a build time test-suite that should ensure some level of functionality from the re-built library. override_dh_auto_test: ifeq ($(filter nocheck,$(DEB_BUILD_OPTIONS)),) $(MAKE) -C nspr/pr/tests $(MAKE) -C nspr/lib/tests # Skip gethost because it needs DNS, and thus networking. # Skip fdcach, peek and vercheck because they fail. # Skip socket because it freezes. # Skip getproto because it fails on some buildds. # Skip nblayer because it freezes on armel. cd nspr/pr/tests && grep -v '^\(fdcach\|gethost\|getproto\|nblayer\|peek\|socket\|vercheck\)$$' ./runtests.sh | sh - $(CURDIR)/nspr/dist cd nspr/lib/tests && LD_LIBRARY_PATH=$(CURDIR)/nspr/dist/bin$(addprefix :,$(LD_LIBRARY_PATH)) ./base64t cd nspr/lib/tests && LD_LIBRARY_PATH=$(CURDIR)/nspr/dist/bin$(addprefix :,$(LD_LIBRARY_PATH)) ./string endif [ Where problems could occur ] No source changes are made. The packaging fixes have enabled other security hardening flags. This could have unintended effects. [ Other Info ] changes in dpkg have been made to make sure frame-pointers are not enabled on s390x and ppc64el. Bug reference with similar SRU - https://bugs.launchpad.net/ubuntu/+source/authbind/+bug/2081707 To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/nspr/+bug/2084460/+subscriptions -- Mailing list: https://launchpad.net/~touch-packages Post to : touch-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~touch-packages More help : https://help.launchpad.net/ListHelp
