Bug#1081515: gtk4: FTBFS with weston 14: many tests fail with --setup=wayland: Failed to open display
Control: tags -1 + patch
On Thu, 12 Sep 2024 at 22:41:10 +0200, Aurelien Jarno wrote:
> Weston 14 is crashing with SIGSEGV a following a few tests like flowbox
> or textiter, despite the test being successful. The following tests
> fails with no display.
The attached patch (proposed upstream by Jan Alexander Steffens) seems to
resolve this: I can build and test gtk4 successfully on amd64 and i386.
Also available as a merge request,
https://salsa.debian.org/xorg-team/wayland/weston/-/merge_requests/11
Thanks,
smcv
From: "Jan Alexander Steffens (heftig)" <heftig@archlinux.org>
Date: Sat, 14 Sep 2024 06:35:09 +0200
Subject: libweston/noop-renderer: Check shm_buffer for NULL
Copy the check from the pixman renderer.
Bug: https://gitlab.freedesktop.org/wayland/weston/-/issues/953
Signed-off-by: Jan Alexander Steffens (heftig) <heftig@archlinux.org>
Forwarded: https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1614
Bug-Debian: https://bugs.debian.org/1081515
---
libweston/noop-renderer.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/libweston/noop-renderer.c b/libweston/noop-renderer.c
index 06b4aeb..58d0b66 100644
--- a/libweston/noop-renderer.c
+++ b/libweston/noop-renderer.c
@@ -94,6 +94,12 @@ noop_renderer_attach(struct weston_paint_node *pnode)
}
shm_buffer = buffer->shm_buffer;
+ /* This can happen if a SHM wl_buffer gets destroyed before we attach,
+ * because wayland-server just nukes the wl_shm_buffer from underneath
+ * us. */
+ if (!shm_buffer)
+ return;
+
data = wl_shm_buffer_get_data(shm_buffer);
stride = buffer->stride;
height = buffer->height;
Reply to: