<html> <head> <base href="https://bugs.freedesktop.org/" /> </head> <body> <div> <a class="bz_bug_link bz_status_ASSIGNED " title="ASSIGNED - libinput_unref() frees memory even when clients hold references to devices or seats" href="https://bugs.freedesktop.org/show_bug.cgi?id=91872#c1">Comment # 1</a> on <a class="bz_bug_link bz_status_ASSIGNED " title="ASSIGNED - libinput_unref() frees memory even when clients hold references to devices or seats" href="https://bugs.freedesktop.org/show_bug.cgi?id=91872">bug 91872</a> from <a class="email" href="mailto:peter.hutterer@who-t.net" title="Peter Hutterer <peter.hutterer@who-t.net>"> Peter Hutterer</a> <pre>The options for fixing this issue introduce incompatible behavior, so I'm closing this as a WONTFIX, this bug is mostly for archival purposes only. If libinput_unref() changes to only do the work when all references are accounted for, callers have to release all existing references manually. This may cause a memory leak in existing callers. In addition, the call to libinput_suspend() would not happen until the resources are released, causing wakeups on top of the memory leaks. If libinput_unref() calls libinput_suspend() but leaves the memory until all references are released, the behaviour becomes unpredictable, and we get memory leaks in existing callers. Requiring an explicit call to libinput_suspend() breaks existing callers. So really, no good fix here and since the current behavior is also that of other libraries (e.g. libudev) we'll leave it as-is.</pre> </div> <hr> You are receiving this mail because: <ul> <li>You are the assignee for the bug.</li> </ul> </body> </html>