Merge ~athos-ribeiro/ubuntu/+source/pmdk:ftbfs-ppc64el into ubuntu/+source/pmdk:ubuntu/devel

[Athos Ribeiro]

Temp fix FTBFS issue on ppc64el - skip failing build time tests.

PPA: https://launchpad.net/~athos-ribeiro/+archive/ubuntu/pmdk-ftbfs/+packages

[Athos Ribeiro (athos-ribeiro)]

a ~ppa2 upload is pending (LP is under maintenance ATM).

We are still concerned with the fact that these can indeed be real bugs which may even affects other architectures as well.

Let's keep this MP on hold until we sort it all out.

[Athos Ribeiro (athos-ribeiro)]

Changed to skipping all tests on ppc64el. There are other tests failing for the python tests.

[Bryce Harrington (bryce)]

Per discussion at team meeting today + direction from Server team manager is that we should proceed with disabling at least the non-passing tests, with bug report(s) to follow up post-release.

Rationale is: 1) that a proper investigation + solution will likely be lower level, hardware-specific, and potentially more complex than the server team is able to perform in the time available pre-release, 2) we're lacking of certainty that upstream would prioritize issues on this architecture for expedient attention, and 3) removal of the ppc64el binaries for pmdk could have a cascading effect on rdepends (qemu, mariadb, ceph, ...).

[Bryce Harrington (bryce)]

There are several options:

1. Disable all tests for ppc64el (only). This is what's implemented in this MP.

2. Disable valgrind. This would skip at least some of the broken tests, however my gut says only a subset of the tests are experiencing valgrind issues (I counted 3 valgrind crashes), so this might be only a partial workaround.

3. Disable the tests currently failing. Athos and I counted 20; these could easily be disabled using just the same approach I used in *ubuntu1. But there might be additional tests beyond that (Athos mentions some python tests?) If some other new test case failed, we'd not be in any better situation for resolving it than we're in currently...

If we had another day I'd try #3 first, and if we had more days beyond that I'd instead do #2 and then #3. But given how late things are I agree with Athos that #1 is safest and most practical to get the build passing. We need to keep in mind that this is just to resolve the ftbfs, and the package still needs time to do dep8 testing and migration.

+1, please upload asap.

review: Approve

[Bryce Harrington (bryce)]

Thanks for not referencing the bug in the changelog; we'll leave that open and continue discussion on resolving it there.

[git-ubuntu bot (git-ubuntu-bot)]

Approvers: athos-ribeiro, bryce
Uploaders: athos-ribeiro, bryce
MP auto-approved

review: Approve

Unmerged commits

a23d3e6... by Athos Ribeiro on 2024-04-18

Update changelog

cfc97f8... by Athos Ribeiro on 2024-04-18

    - d/rules: skip ppc64el build time tests

04b94ba... by Athos Ribeiro on 2024-04-18

    - d/p/debian-changes: remove bogus file

diff --git a/debian/changelog b/debian/changelog
index b753303..f1a6159 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,3 +1,11 @@
+pmdk (1.13.1-1.1ubuntu2) noble; urgency=medium
+
+  * Fix FTBFS issues in ppc64el:
+    - d/rules: skip ppc64el build time tests
+    - d/p/debian-changes: remove bogus file
+
+ -- Athos Ribeiro <[e-mail]>  Thu, 18 Apr 2024 09:44:59 -0300
+
 pmdk (1.13.1-1.1ubuntu1) noble; urgency=medium
 
   * d/rules: Disable TEST5 on ppc64el to fix FTBFS
diff --git a/debian/patches/debian-changes b/debian/patches/debian-changes
deleted file mode 100644
index 3b637b7..0000000
--- a/debian/patches/debian-changes
+++ /dev/null
@@ -1,37 +0,0 @@
-Description: Autogenerated patch header for a single-debian-patch file.
- The delta against upstream is either kept as a single patch, or maintained
- in some VCS, and exported as a single patch instead of more manageable
- atomic patches.
-Forwarded: not-needed
-
----
---- /dev/null
-+++ pmdk-1.13.1/0001-d-rules-Disable-TEST5-on-ppc64el-to-fix-FTBFS.patch
-@@ -0,0 +1,27 @@
-+From c205347c197212ad46fc61c378470d0d25b49964 Mon Sep 17 00:00:00 2001
-+From: Bryce Harrington <[e-mail]>
-+Date: Tue, 16 Apr 2024 16:12:46 -0700
-+Subject: [PATCH]   * d/rules: Disable TEST5 on ppc64el to fix FTBFS     (LP:
-+ #2061913)
-+
-+---
-+ debian/rules | 4 ++++
-+ 1 file changed, 4 insertions(+)
-+
-+diff --git a/debian/rules b/debian/rules
-+index 134e00ee..b4217c0c 100755
-+--- a/debian/rules
-++++ b/debian/rules
-+@@ -48,5 +48,9 @@ ifeq (,$(filter nocheck,$(DEB_BUILD_OPTIONS)))
-+ 	rm -f src/test/obj_pool/TEST34
-+ 	# binutils regression #964457
-+ 	rm -rf src/test/scope/TESTS.py
-++ifeq ($(DEB_BUILD_ARCH),ppc64el)
-++	# https://github.com/pmem/pmdk/issues/6079
-++	rm -f src/test/obj_basic_integration/TEST5
-++endif
-+ 	+$(MAKE) check pycheck $(BUILD_ARGS)
-+ endif
-+-- 
-+2.34.1
-+
diff --git a/debian/patches/series b/debian/patches/series
index 01830c6..2be7c09 100644
--- a/debian/patches/series
+++ b/debian/patches/series
@@ -1,2 +1 @@
 check-manpages-noerror.patch
-debian-changes
diff --git a/debian/rules b/debian/rules
index deff278..4abc353 100755
--- a/debian/rules
+++ b/debian/rules
@@ -7,6 +7,11 @@ DEB_UPSTREAM_VERSION = $(shell dpkg-parsechangelog -S version | cut -d - -f 1 | 
 
 BUILD_ARGS = prefix=/usr libdir=/usr/lib/$(DEB_HOST_MULTIARCH) sysconfdir=/etc bashcompdir=/usr/share/bash-completion/completions NORPATH=1 BUILD_EXAMPLES=n BUILD_BENCHMARKS=n
 
+# LP: #2061913
+ifeq ($(DEB_BUILD_ARCH),ppc64el)
+export DEB_BUILD_OPTIONS+=nocheck
+endif
+
 %:
 	dh $@ --with bash-completion
 
@@ -44,8 +49,6 @@ ifeq (,$(filter nocheck,$(DEB_BUILD_OPTIONS)))
 ifeq ($(DEB_BUILD_ARCH),ppc64el)
 	# https://github.com/pmem/pmdk/issues/5458
 	rm -f src/test/obj_ctl_arenas/TEST3 src/test/pmem2_future/TESTS.py
-	# https://github.com/pmem/pmdk/issues/6079
-	rm -f src/test/obj_basic_integration/TEST5
 endif
 	+$(MAKE) check pycheck $(BUILD_ARGS)
 endif