On NetBSD 6.0.1, with the tests hacked to make these failures not fatal as they normally are: ./../bus/bus-test: Running SHA1 connection test ./../bus/bus-test: checking for memleaks 2 dbus_malloc blocks were not freed ... Shell echo service echoed the command line ./../bus/bus-test: checking for memleaks 2 dbus_malloc blocks were not freed ./../bus/bus-test: Running service files reloading test ./../bus/bus-test: checking for memleaks 2 dbus_malloc blocks were not freed ./../bus/bus-test: Running unix fd passing test ./../bus/bus-test: checking for memleaks 2 dbus_malloc blocks were not freed and I've also seen tests fail complaining that fd 5 was not close-on-exec (in the dbus-spawn tests). The "core" D-Bus developers are incredibly unlikely to fix this.
On FreeBSD 9.1 ./../bus/bus-test: checking for memleaks ./../bus/bus-test: Running signals test ./../bus/bus-test: checking for memleaks ./../bus/bus-test: Running SHA1 connection test ./../bus/bus-test: checking for memleaks 2 dbus_malloc blocks were not freed D-Bus not compiled with backtrace support so unable to print a backtrace Process 75491 sleeping for gdb attach
(In reply to comment #0) > > and I've also seen tests fail complaining that fd 5 was not close-on-exec > (in the dbus-spawn tests). Just opened a new ticket for that error, see #bug 72213 > > The "core" D-Bus developers are incredibly unlikely to fix this.
(In reply to comment #1) > On FreeBSD 9.1 > > ./../bus/bus-test: checking for memleaks > ./../bus/bus-test: Running signals test > ./../bus/bus-test: checking for memleaks > ./../bus/bus-test: Running SHA1 connection test > ./../bus/bus-test: checking for memleaks > 2 dbus_malloc blocks were not freed > D-Bus not compiled with backtrace support so unable to print a backtrace > Process 75491 sleeping for gdb attach I have patches for these two blocks memleak, still working on #bug 72213
Created attachment 90139 [details] [review] [PATCH 1/3] Fix memory leak for kqueue: replace tab with space
Created attachment 90140 [details] [review] [PATCH 2/3] Fix memory leak for kqueue: shutdown kqueue correctly
Created attachment 90141 [details] [review] [PATCH 3/3] Fix memory leak for kqueue: free memory before assign
Created attachment 90143 [details] [review] [PATCH v2] Fix memory leak for kqueue: shutdown kqueue correctly
Comment on attachment 90139 [details] [review] [PATCH 1/3] Fix memory leak for kqueue: replace tab with space Review of attachment 90139 [details] [review]: ----------------------------------------------------------------- I'll put it through "git show --ignore-space-change", but probably r+.
Comment on attachment 90143 [details] [review] [PATCH v2] Fix memory leak for kqueue: shutdown kqueue correctly Review of attachment 90143 [details] [review]: ----------------------------------------------------------------- > Also, it fixes other memory leaks I don't see any non-D-Bus memory leaks here? It fixes D-Bus memory leaks, and fd leaks. ::: bus/dir-watch-kqueue.c @@ +122,5 @@ > + _dbus_watch_unref (watch); > + watch = NULL; > + } > + > + if (kq >= 0) Redundant with the early-return at the top.
Created attachment 90153 [details] [review] [PATCH v3] Fix memory leak for kqueue: shutdown kqueue correctly Update commit message and drop redundant "if (kq >= 0)" clause.
Fixed in git for 1.7.10, thanks
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.