please just check CONFIG_INOTIFY_USER .. from linux 2.6.32 makefile: CONFIG_INOTIFY: Say Y here to enable legacy in kernel inotify support. Inotify is a file change notification system. It is a replacement for dnotify. This option only provides the legacy inotify in kernel API. There are no in tree kernel users of this interface since it is deprecated. You only need this if you are loading an out of tree kernel module that uses inotify. Reproducible: Always
Please describe why ? Afaik, tracker-0.7 does not need this anymore as it is handled by glib/gio.
maybe it does not need it, although it's used indirectly through gio and my guess is tracker is not usable without inotify. anyway, my point was that checking for CONFIG_INOTIFY (as the ebuild does now) is wrong.
Without a more detailed explanation, I'll stick to what the source says: $ egrep "include.*inotify.h" * -r src/libtracker-miner/tracker-monitor.c:#include <sys/inotify.h> [and more...]
did you even read my report? it has nothing to do with tracker using inotify or not. it's about CONFIG_INOTIFY vs. CONFIG_INOTIFY_USER, where the first one is only used by kernel modules (or not anymore, it's deprecated).
erm sorry I'm a bit sleepy, what is the request about ? Should we change the check, remove it ?
this is what i propose: diff --git a/app-misc/tracker/tracker-0.7.10.ebuild b/app-misc/tracker/tracker-0.7.10.ebuild index 0f40e6a..55eddac 100644 --- a/app-misc/tracker/tracker-0.7.10.ebuild +++ b/app-misc/tracker/tracker-0.7.10.ebuild @@ -85,7 +85,7 @@ function notify_inotify() { } function inotify_enabled() { - linux_chkconfig_present INOTIFY && linux_chkconfig_present INOTIFY_USER + linux_chkconfig_present INOTIFY_USER } pkg_setup() {
so if I understand this right, we now ignore invalid setups for older kernels, right ?
CONFIG_INOTIFY_USER is available since 2.6.18 (http://cateee.net/lkddb/web-lkddb/INOTIFY_USER.html) and earlier kernels depended on CONFIG_INOTIFY. so you could not have CONFIG_INOTIFY_USER enabled an not CONFIG_INOTIFY. i see no problems here.
FYI, just found bug #283763, udev had the same problem.
Fixed in overlay. Thanks for reporting.