From 45ae938a58418dde912406e782e21236ef52f700 Mon Sep 17 00:00:00 2001 From: Ryan Hughes Date: Sun, 27 Sep 2026 20:01:44 -0400 Subject: [PATCH] flea: put only copymanifest's O_TMPFILE test on tmpfs Moving the whole module broke the rest of it: Writer::create needs a runtime directory on a different filesystem from the copy, and with the test root on /dev/shm none qualifies. Only a_garbage_stream_falls_back_with_the_tree_untouched opens its manifest in its own test dir, which the builder's /tmp refuses (EOPNOTSUPP). --- pkgbuilds/flea/PKGBUILD | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/pkgbuilds/flea/PKGBUILD b/pkgbuilds/flea/PKGBUILD index 77365f55..23c3674f 100644 --- a/pkgbuilds/flea/PKGBUILD +++ b/pkgbuilds/flea/PKGBUILD @@ -115,8 +115,12 @@ check() { # back to the next create, so on the builder's /tmp the stranger is # identical and undo walks it back. tmpfs allocates inode numbers from a # counter and never reuses one, which is what the test assumes. + # copymanifest's garbage-stream test opens its O_TMPFILE manifest in its + # own test dir; the module's other tests must stay off tmpfs, because + # Writer::create wants a runtime dir on a different filesystem from the + # copy and finds none when both live on /dev/shm. local -a filesystem_tests=( - backend::copymanifest::tests:: + backend::copymanifest::tests::a_garbage_stream_falls_back_with_the_tree_untouched backend::menu_actions::tests:: backend::menudelete::tests:: backend::redo::tests::