Openembedded Core Discussions
 help / color / mirror / Atom feed
From: sunilkumar.dora@windriver.com
To: openembedded-core@lists.openembedded.org
Cc: raj.khem@gmail.com, randy.macleod@windriver.com,
	sundeep.kokkonda@windriver.com, deepesh.vatharaj@windriver.com
Subject: [PATCH 07/11] pango: Don't turn clang 23's -Wunused-but-set-global into an error
Date: Sun, 30 Aug 2026 18:28:47 +0530	[thread overview]
Message-ID: <20260830125853.1139807-8-sunilkumar.dora@windriver.com> (raw)
In-Reply-To: <20260830125853.1139807-1-sunilkumar.dora@windriver.com>

From: Sunil Dora <sunilkumar.dora@windriver.com>

clang 23 added -Wunused-but-set-global to the -Wunused-but-set-variable
group, which pango promotes to an error:

  pango/pango-fontset.c:36:1: error: variable 'pango_fontset_parent_class'
  set but not used [-Werror,-Wunused-but-set-global]

It fires on the parent_class variables emitted by GLib's G_DEFINE_TYPE
macros. Keep the warning visible but not fatal. gcc builds are
unaffected as meson only passes flags the compiler supports.

Reported upstream with patches attached: the meson change as pango
issue #900 and the underlying GLib macro fix as glib issue #4038.

Signed-off-by: Sunil Dora <sunilkumar.dora@windriver.com>
---
 ...ror-on-clang-s-unused-but-set-global.patch | 28 +++++++++++++++++++
 meta/recipes-graphics/pango/pango_1.58.0.bb   |  1 +
 2 files changed, 29 insertions(+)
 create mode 100644 meta/recipes-graphics/pango/pango/0001-meson-don-t-error-on-clang-s-unused-but-set-global.patch

diff --git a/meta/recipes-graphics/pango/pango/0001-meson-don-t-error-on-clang-s-unused-but-set-global.patch b/meta/recipes-graphics/pango/pango/0001-meson-don-t-error-on-clang-s-unused-but-set-global.patch
new file mode 100644
index 0000000000..2c3788abdf
--- /dev/null
+++ b/meta/recipes-graphics/pango/pango/0001-meson-don-t-error-on-clang-s-unused-but-set-global.patch
@@ -0,0 +1,28 @@
+From: Sunil Dora <sunilkumar.dora@windriver.com>
+Date: Wed, 20 Aug 2026 10:45:00 +0530
+Subject: [PATCH] meson: don't error on clang's -Wunused-but-set-global
+
+clang 23 added -Wunused-but-set-global to the -Wunused-but-set-variable
+group, so the -Werror=unused-but-set-variable in test_cflags now turns
+it into an error. It fires on the <type>_parent_class variables that
+GLib's G_DEFINE_TYPE family of macros emit, which pango cannot fix, and
+on a few deliberately write-only globals in the test suite.
+
+Keep the warning visible but not fatal. gcc does not know this warning
+group, so cc.get_supported_arguments() drops the flag there.
+
+Upstream-Status: Submitted [https://gitlab.gnome.org/GNOME/pango/-/issues/900]
+Signed-off-by: Sunil Dora <sunilkumar.dora@windriver.com>
+---
+--- a/meson.build
++++ b/meson.build
+@@ -148,6 +148,9 @@
+     '-Werror=empty-body',
+     '-Werror=write-strings',
+     '-Werror=unused-but-set-variable',
++    # clang 23 made this part of the group above; it fires on GLib's
++    # G_DEFINE_TYPE boilerplate, which pango cannot fix.
++    '-Wno-error=unused-but-set-global',
+     '-Wundef', # FIXME: https://bugzilla.gnome.org/show_bug.cgi?id=792481
+   ]
+ 
diff --git a/meta/recipes-graphics/pango/pango_1.58.0.bb b/meta/recipes-graphics/pango/pango_1.58.0.bb
index 6e1aebad1d..0c9e26f39c 100644
--- a/meta/recipes-graphics/pango/pango_1.58.0.bb
+++ b/meta/recipes-graphics/pango/pango_1.58.0.bb
@@ -18,6 +18,7 @@ UPSTREAM_CHECK_REGEX = "pango-(?P<pver>\d+\.(?!9\d+)\d+\.\d+)"
 GIR_MESON_ENABLE_FLAG = "enabled"
 GIR_MESON_DISABLE_FLAG = "disabled"
 
+SRC_URI += "file://0001-meson-don-t-error-on-clang-s-unused-but-set-global.patch"
 SRC_URI[archive.sha256sum] = "bc5bad6213ad4886a47d1e80292fd850b64159b50db67917a43d9ea80ee2298a"
 
 DEPENDS = "glib-2.0 glib-2.0-native fontconfig freetype virtual/libiconv cairo harfbuzz fribidi"
-- 
2.43.0



  parent reply	other threads:[~2026-08-30 12:59 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-08-30 12:58 [PATCH 00/11] clang/llvm: Upgrade to 23.1.0 sunilkumar.dora
2026-08-30 12:58 ` [PATCH 01/11] clang/llvm: Upgrade to 23.1.0 release sunilkumar.dora
2026-08-30 19:53   ` Khem Raj
2026-09-01  3:37     ` [OE-core] " Dora, Sunil Kumar
2026-09-01 12:49       ` Richard Purdie
2026-08-30 12:58 ` [PATCH 02/11] spirv-llvm-translator: Upgrade to 23.1.0 sunilkumar.dora
2026-08-30 19:43   ` Khem Raj
2026-09-01  4:07     ` [OE-core] " Dora, Sunil Kumar
2026-08-30 12:58 ` [PATCH 03/11] openmp: Build via the LLVM runtimes entry point sunilkumar.dora
2026-08-30 19:45   ` Khem Raj
2026-09-01  4:33     ` [OE-core] " Dora, Sunil Kumar
2026-09-03  1:11       ` Khem Raj
2026-08-30 12:58 ` [PATCH 04/11] rust: Fix build and oe-selftest failures with LLVM 23 sunilkumar.dora
2026-08-30 19:47   ` Khem Raj
2026-09-01  5:17     ` [OE-core] " Dora, Sunil Kumar
2026-09-03  1:08       ` Khem Raj
2026-08-30 12:58 ` [PATCH 05/11] rust: Do not pass the removed x86 amx-tf32 feature to " sunilkumar.dora
2026-08-30 12:58 ` [PATCH 06/11] clang: Disable clangd's decision-forest completion model on powerpc sunilkumar.dora
2026-08-30 12:58 ` sunilkumar.dora [this message]
2026-08-30 12:58 ` [PATCH 08/11] ovmf: Fix memcpy alias prototype for clang 23 sunilkumar.dora
2026-08-30 12:58 ` [PATCH 09/11] vulkan-samples: Don't turn clang 23's -Wdangling-gsl into an error sunilkumar.dora
2026-08-30 12:58 ` [PATCH 10/11] gettext: Export a fully-flagged Objective-C compiler sunilkumar.dora
2026-08-30 12:58 ` [PATCH 11/11] clang/llvm: Ignore tests fail with clang/llvm 23 upgrade sunilkumar.dora
2026-08-30 19:53 ` [PATCH 00/11] clang/llvm: Upgrade to 23.1.0 Khem Raj
2026-09-01  4:59   ` [OE-core] " Dora, Sunil Kumar
2026-09-03  1:10     ` Khem Raj
2026-09-03 23:46       ` Khem Raj
2026-09-04  4:49         ` Dora, Sunil Kumar
2026-09-05 13:26           ` Dora, Sunil Kumar
2026-09-05 18:01             ` [OE-core] " Khem Raj

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260830125853.1139807-8-sunilkumar.dora@windriver.com \
    --to=sunilkumar.dora@windriver.com \
    --cc=deepesh.vatharaj@windriver.com \
    --cc=openembedded-core@lists.openembedded.org \
    --cc=raj.khem@gmail.com \
    --cc=randy.macleod@windriver.com \
    --cc=sundeep.kokkonda@windriver.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox