2 * check systemd-tmpfiles for selinux context hookup for mknod(), symlink() and similar
4 * swap units that are activated by one name but shown in the kernel under another are semi-broken
6 * Dangling symlinks of .automount unit files in .wants/ directories, set up
7 automount points even when the original .automount file did not exist
8 anymore. Only the .mount unit was still around.
10 * make polkit checks async
12 * properly handle .mount unit state tracking when two mount points are stacked one on top of another on the exact same mount point.
14 * add 'set -e' to scripts in test/
15 * make test in test/ work with separate output dir
17 * suppress log output on shutdown when "quiet" is used
21 * make anaconda write timeout=0 for encrypted devices
23 * make sure pkexec works fine with pam_systemd works fine with audit=0
25 * create /var/log/journal/
27 * external: maybe it is time to patch procps so that "ps" links to
28 libsystemd-logind to print a pretty service name, seat name, session
29 name in its output. Currently it only shows cgroup membership, but
30 that's sometimes kinda hard to parse for a human.
33 - don't unconditionally filter out duplicate settings
34 - support high-level cgroup setting syntax in systemctl. Example: "systemctl set-cgroup-attr MemoryLimit 5K"
35 - support writte string mapping even for non-high-level settings
36 - add man page for systemctl commands
37 - make sure we work fine with multi-line strings
39 * kernel cmdline switch to turn off predictable network interface names
41 * journal is not closed properly at shutdown when run in a container?
43 * introduce new "journal" group in place of adm?
46 - localectl: add listing support for X11 keymaps, by parsing /usr/share/X11/xkb/rules/xorg.lst
47 - localectl: support new converted x11→console keymaps
49 * logind: Class property should probably know "background" or so as value for cron jobs, and the inhibition checks should filter those out too.
51 * timer logic is confused by units which are skipped due to failing condition
52 http://lists.freedesktop.org/archives/systemd-devel/2013-February/008816.html
56 * convert /etc/bash_completion.d/ to /usr/share/bash-completion/
57 (systemd (and everything else) should install zero static data in /etc)
59 * tmpfiles: when traversing the tree, check for bind mount points with nametohandle()
61 * fedora: connect the timer units of a service to the service via Also= in [Install], and maybe introduce timers.target
63 * fedora: F20: go timer units all the way, leave cron.daily for cron
65 * add a tool that lists active timer units plus their next elapstion and the time the units ran last
67 * man: document the very specific env the shutdown drop-in tools live in
69 * shutdown logging: store to EFI var, and store to USB stick?
71 * man: extend runlevel(8) to mention that runlevels suck, and are dead. Maybe add runlevel(7) with a note about that too
73 * systemctl: maybe add "systemctl add-wants" or so...
75 * man: add a link to socket activation blog from systemd.socket(5)
77 * systemctl status: show drop-in snippets for service files in addition to service file path themesevles.
79 * man: add more examples to man pages
81 * man: maybe sort directives in man pages, and take sections from --help and apply them to man too
83 * man: systemctl: clarify that -p applies to "systemctl show" only
85 * add "# export SYSTEMD_PAGER=" to bash login
87 * /usr/bin/service should actually show the new command line
89 * fedora: suggest auto-restart on failure, but not on sucess and not on coredump. also, ask people to think about changing the start limit logic. Also point people to RestartPreventExitStatus=, SuccessExitStatus=
91 * write UI tool that pops up emergency messages from the journal as notification
93 * think about window-manager-run-as-user-service problem: exit 0 → activate shutdown.target; exit != 0 → restart service
95 * msgcatalog determine default language of entries from the catalog
96 file name. i.e. foobar.de_DE.catalog would set the default entry
97 name for the entries to de_DE if they aren't explicitly suffixed
100 * figure out what we do about hostnames/fqdn in hostnamectl/hostnamed
102 * use "log level" rather than "log priority" everywhere
104 * unit_name_mangle() is currently not used on "systemctl enable" and friends, only on "systemctl start"
106 * ensure sd_journal_seek_monotonic actually works properly.
108 * timedate: have global on/off switches for auto-time (NTP), and auto-timezone that connman can subscribe to.
110 * support --root= in msgcatalog compiler
112 * Honour "-" prefix for InaccessibleDirectories= and ReadOnlyDirectories= to
113 suppress errors of the specified path doesn't exist
115 * dev-setup.c: when running in a container, create a tiny stub udev
116 database with the systemd tag set for all network interfaces found,
117 so that libudev reports them as present, and systemd's .device units
120 * merge service_kill() logic into unit_kill_context() and unify the code paths
122 * introduce ExecCondition= in services
125 - fsck hookup for the ESP mount is missing
126 - write man page for efi boot generator
127 - honor language efi variables for default language selection (if there are any?)
128 - honor timezone efi variables for default timezone selection (if there are any?)
129 - introduce bootctl (backed by systemd-bootd) to control temporary and persistent default boot goal plus efi variables
131 * maybe do not install getty@tty1.service symlink in /etc but in /usr?
133 * re-enable "make check" for gtk-doc (broken for unknown reason)
135 * fstab: add new mount option x-systemd-after=/foobar/waldo to allow manual dependencies to other mount points
136 https://bugzilla.redhat.com/show_bug.cgi?id=812826
138 * print a nicer explanation if people use variable/specifier expansion in ExecStart= for the first word
140 * mount: turn dependency information from /proc/self/mountinfo into dependency information between systemd units.
143 - logind: optionally, ignore idle-hint logic for autosuspend, block suspend as long as a session is around
144 - When we update the kernel all kind of hibernation should be prohibited until shutdown/reboot
145 - logind: wakelock/opportunistic suspend support
146 - Add pretty name for seats in logind
147 - logind: allow showing logout dialog from system?
148 - logind: spawn user@..service on login
149 - logind: non-local X11 server handling
150 - logind: add equivalent to sd_pid_get_owner_uid() to the D-Bus API
151 - pam_systemd: try to get old session id from cgroup, if audit sessionid cannot be determined
152 - pam: when leaving a session explicitly exclude the ReleaseSession() caller process from the killing spree
153 - logind: GetSessionByPID() should accept 0 as PID value
155 * exec: when deinitializating a tty device fix the perms and group, too, not only when initializing. Set access mode/gid to 0620/tty.
157 * DeviceAllow/DeviceDeny: disallow everything by default, but whitelist /dev/zero, /dev/null and friends
159 * service: watchdog logic: for testing purposes allow ping, but do not require pong
162 - import and delete pstore filesystem content at startup
163 - journald: also get thread ID from client, plus thread name
164 - journal: when waiting for journal additions in the client always sleep at least 1s or so, in order to minimize wakeups
165 - add API to close/reopen/get fd for journal client fd in libsystemd-journal.
166 - fallback to /dev/log based logging in libsystemd-journal, if we can't log natively?
167 - declare the local journal protocol stable in the wiki interface chart
168 - journal: reuse XZ context
169 - sd-journal: speed up sd_journal_get_data() with transparent hash table in bg
170 - journald: when dropping msgs due to ratelimit make sure to write
171 "dropped %u messages" not only when we are about to print the next
172 message that works, but alraedy after a short tiemout
173 - check if we can make journalctl by default use --follow mode inside of less if called without args?
174 - maybe add API to send pairs of iovecs via sd_journal_send
175 - journal: when writing journal auto-rotate if time jumps backwards
176 - gatewayd: should run under its own UID
177 - journal: add a setgid "adm" utility to invoke from libsystemd-journal, which passes fds via STDOUT and does PK access
178 - journactl: support negative filtering, i.e. FOOBAR!="waldo",
179 and !FOOBAR for events without FOOBAR.
180 - journal: when rotating, copy over old acls/access mode
181 - journal: document why we do not give ownership to journal files to the user that created them but use FS ACLs for that
182 - journal: send out marker messages every now and then, and immediately sync with fdatasync() afterwards, in order to have hourly guaranteed syncs.
183 - journal: when we haven't written anything in a while, sync to disk and mark file as offline, in order to be more often than not in a clean state
184 - journal-send.c, log.c: when the log socket is clogged, and we drop, count this and write a message about this when it gets unclogged again.
185 - journal: find a way to allow dropping history early, based on priority, other rules
186 - journal: When used on NFS, check payload hashes
187 - journal: When used on NFS make sure wake up sd_journal_wait() every 2s, to handle missing inotify
188 - document that people can use file system ACLs to manage access to journal files, with example
189 - Introduce journalctl -b <nr> to show journal messages of a previous boot
190 - journald: check whether it is OK if the client can still modify delivered journal entries
191 - journal live copy, based on libneon (client) and libmicrohttpd
192 - journald: add kernel cmdline option to disable ratelimiting for debug purposes
193 - refuse taking lower-case variable names in sd_journal_send() and friends.
194 - journald: we currently rotate only after MaxUse+MaxFilesize has been reached.
195 - journal: deal nicely with byte-by-byte copied files, especially regards header
196 - journalctl: show multiline log messages sanely, expand tabs, and show all valid utf8 messages
197 - journal: store euid in journal if it differs from uid
198 - journal: sanely deal with entries which are larger than the individual file size, but where the components would fit
201 - document unit_name_mangle()
202 - document that deps in [Unit] sections ignore Alias= fileds in
203 [Install] units of other units, unless those units are disabled
204 - man: clarify that time-sync.target is not only sysv compat but also useful otherwise. Same for similar targets
205 - Document word splitting syntax for ExecStart= and friends
206 - document that units from /etc override those from /usr and /run
207 - document the exit codes when services fail before they are exec()ed
208 - document that service reload may be implemented as service reexec
209 - document in wiki how to map ical recurrence events to systemd timer unit calendar specifications
212 - systemctl list-jobs - show dependencies
213 - add systemctl switch to dump transaction without executing it
214 - Add a verbose mode to "systemctl start" and friends that explains what is being done or not done
215 - "systemctl disable" on a static unit prints no message and does
216 nothing. "systemctl enable" does nothing, and gives a bad message
217 about it. Should fix both to print nice actionable messages.
218 - print nice message from systemctl --failed if there are no entries shown, and hook that into ExecStartPre of rescue.service/emergency.service
219 - add new command to systemctl: "systemctl system-reexec" which reexecs as many daemons as virtually possible
220 - systemctl enable: improve the success messages (i.e. more human readable, less shell-like)
221 - systemctl enable: fail if target to alias into doesn't exist? maybe show how many units are enabled afterwards?
222 - systemctl: "Journal has been rotated since unit was started." message is misleading
223 - support "systemctl stop foobar@.service" to stop all units matching a certain template
224 - Something is wrong with symlink handling of "autovt@.service" in "systemctl list-unit-files"
226 * When shutdown.target is queued begin with an asynchronous sync()?
228 * introduce ntp.service (or suchlike) as symlink that is used to arbitrate between various NTP implementations
230 * deal with sendmail/postfix exclusivity
233 - configurable jitter for timer events
234 - timer events with system resume
235 - timer units should get the ability to trigger when:
236 o CLOCK_REALTIME makes jumps (TFD_TIMER_CANCEL_ON_SET)
238 - Support 2012-02~4 as syntax for specifying the fourth to last day of the month.
240 * update the kernel's TZ (sys_tz) when DST changes
242 * sync down the system time to the RTC when:
243 - CLOCK_REALTIME makes jumps (the user explicitely requested a time set)
244 - DST/timezone changes && ntp is active && RTC-in-localtime (never do it without ntp)
245 This takes care of syncing ntpdate updates to the RTC, and DST updates for localtime
246 mode, it will never touch the RTC if the no reliable time source is active or the
247 user did not request anything like it.
250 - implement conditional properties (dmi matches)
251 - hwdb --filter=ID_DRIVE_*
253 * if booted in "quiet" mode, and an error happens, turn on status
254 output again, so that the emergency mode isn't totally
255 surprising. Also, terminate plymouth.
257 * libunwind support for coredump pattern hook, and includes this in
258 the message for coredumps. After all, libunwind is now capable to
259 unwind coredumps since a few weeks ago. This probably requires that
260 we have nice support for multi-line messages on display in
261 logs-show.c. Alternatively: use libelfutil, which seems to be the
262 better supported alternative.
264 * figure out relation of --all and --full in the various tools
266 * add libsystemd-password or so to query passwords during boot using the password agent logic
268 * If we show an error about a unit (such as not showing up) and it has no Description string, then show a description string generated form the reverse of unit_name_mangle().
270 * fedup: add --unit to systemctl switch-root somehow
271 * fedup: don't delete initrd on switch-root
275 - timedated: export boolean that clarifies whether NTP is even available
276 - timedated: refuse time changes when NTP is on
278 * clean up date formatting and parsing so that all absolute/relative timestamps we format can also be parsed
280 * introduce generic AUGMENT_PID=, AUGMENT_DEVICE= fields
282 * on shutdown: move utmp, wall, audit logic all into PID 1 (or logind?), get rid of systemd-update-utmp-runlevel
284 * add "provisioning" instructions to setup an empty /etc + /var
285 - used to setup a new container from a shared /usr
286 - superset of tmpfiles model
287 - instructions shipped by packages and stored in /usr/lib/
288 - compose /etc/passwd and /etc/group, copy files
289 - able to create uid + gid used by packages, for file ownership
291 * make repeated alt-ctrl-del presses printing a dump, or even force a reboot without
292 waiting for the timeout
294 * high level net_prio setting in execution context
296 * hostnamed: before returning information from /etc/machine-info.conf check the modification data and reread. Similar for localed, ...
298 * currently x-systemd.timeout is lost in the initrd, since crypttab is copied into dracut, but fstab isn't
300 * WorkingDirectory: support env var replacements like in ExecStart= so that people can use $HOME
302 * refuse boot if /etc/machine-id is not useful (or set taint?)
305 - nspawn: consider changing users for -u with su, so that NSS resolving works correctly
306 - nspawn: implement personality changes a la linux32(8)
307 - nspawn: --read-only is not applied recursively to submounts
308 - nspawn: make use of device cgroup controller by default
309 - bind mount read-only the cgroup tree higher than nspawn
310 - nspawn: investigate whether we can support the same as LXC's
311 lxc.network.type=phys mode, and pass through entire network
312 interfaces to the container
313 - nspawn: maybe add a way to drop additional caps, in addition to add additional caps
314 - nspawn: maybe explicitly reset loginuid?
315 - nspawn: move all containers to name=systemd:/containers/, and introduce externally visible names for containers
316 - nspawn: make it work for dwalsh and shared /usr containers -- tmpfs mounts as command line parameters, selinux exec context
319 - cryptsetup-generator: warn if the password files are world-readable
320 - cryptsetup-generator: add RequiresMountsFor= to cryptseup service files referencing a file, similar for devices
321 - cryptsetup-generator: allow specification of passwords in crypttab itself
322 - move cryptsetup key caching into kernel keyctl?
323 https://bugs.freedesktop.org/show_bug.cgi?id=54982
324 - when key file cannot be found, read it from kbd in cryptsetup
326 * move debug shell to tty6 and make sure this doesn't break the gettys on tty6
328 * hw watchdog: optionally try to use the preset watchdog timeout instead of always overriding it
329 https://bugs.freedesktop.org/show_bug.cgi?id=54712
331 * after deserializing sockets in socket.c we should reapply sockopts and things
333 * make timer units go away after they elapsed
335 * http://lists.freedesktop.org/archives/systemd-devel/2012-September/006502.html
336 (network and remote-fs on shutdown)
338 * come up with a nice way to write queue/read_ahead_kb for a block device without interfering with readahead
340 * move PID 1 segfaults to /var/lib/systemd/coredump?
342 * create /sbin/init symlinks from the build system
344 * Query Paul Moore about relabelling socket fds while they are open
346 * system.conf should have controls for cgroups
348 * allow writing multiple conditions in unit files on one line
350 * explore multiple service instances per listening socket idea
352 * testing tool for socket activation: some binary that listens on a socket and passes it on using the usual socket activation protocol to some server.
354 * shutdown: don't read-only mount anything when running in container
356 * MountFlags=shared acts as MountFlags=slave right now.
358 * ReadOnlyDirectories= is not applied recursively to submounts
360 * drop PID 1 reloading, only do reexecing (difficult: Reload()
361 currently is properly synchronous, Reexec() is weird, because we
362 can't delay the response properly until we are back, so instead of
363 being properly synchronous we just keep open the fd and close it
364 when done. That means clients don't get a successful method reply,
365 but much rather a disconnect on success.
367 * remember which condition failed for services, not just the fact that something failed
369 * use opterr = 0 for all getopt tools
371 * properly handle loop back mounts via fstab, especially regards to fsck/passno
373 * allow services with no ExecStart= but with an ExecStop=
375 * dracut-shutdown needs to be ordered before unmounting /boot
377 * initialize the hostname from the fs label of /, if /etc/hostname does not exist?
379 * rename "userspace" to "core-os"
382 - syscall filter: add knowledge about compat syscalls
383 - syscall filter: don't enforce no new privs?
384 - syscall filter: option to return EPERM rather than SIGSYS?
385 - syscall filter: port to libseccomp
386 - system-wide seccomp filter
388 * systemd-analyze: data collection tools should be lightweight (few dependencies); data analysis tools can be heavyweight
390 * .device aliases need to be implemented with the "following" logic, probably.
392 * load-fragment: when loading a unit file via a chain of symlinks
393 verify that it isn't masked via any of the names traversed.
395 * introduce Type=pid-file
397 * maybe allow services with ExecStop= set, but no ExecStart=?
399 * change Requires=basic.target to RequisiteOverride=basic.target
401 * support rd.luks.allow-discards= kernel cmdline params in cryptsetup generator
403 * drop accountsservice's StandardOutput=syslog and Type=dbus fields
405 * when breaking cycles drop sysv services first, then services from /run, then from /etc, then from /usr
407 * move passno parsing to fstab generator
409 * automount: implement expire:
410 - set superblock timeout AUTOFS_DEV_IOCTL_TIMEOUT_CMD
411 - periodically run AUTOFS_DEV_IOCTL_EXPIRE_CMD
412 - every timeout/4 (original autofs logic)
413 - blocking, needs a thread
415 - receive expire packet on pipe if kernel tells the timeout is over
417 - answer expire packet on pipe with AUTOFS_DEV_IOCTL_{READY,FAIL}_CMD
418 - AUTOFS_DEV_IOCTL_EXPIRE_CMD returns
420 * services which create their own subcgroups break cgroup-empty notification (needs to be fixed in the kernel)
422 * ExecOnFailure=/usr/bin/foo
424 * ConditionSecurity= should learn about IMA and SMACK
427 - remove all (misguided from day 1) userspace firmware_class handling
429 - unify utf8 validator code with shared/
431 - add trigger --subsystem-match=usb/usb_device device
433 * cleanup syslog 'priority' vs. 'level' wording
435 * dbus upstream still refers to dbus.target and shouldn't
437 * when a service has the same env var set twice we actually store it twice and return that in systemctl show -p... We should only show the last setting
439 * support container_ttys=
441 * introduce mix of BindTo and Requisite
443 * add DeleteSocketsOnStop=yes|no option to socket units
445 * There's currently no way to cancel fsck (used to be possible via C-c or c on the console)
447 * add option to sockets to avoid activation. Instead just drop packets/connections, see http://cyberelk.net/tim/2012/02/15/portreserve-systemd-solution/
449 * default unix qlen is too small (10). bump sysctl? add sockopt?
451 * figure out whether we should leave dbus around during shutdown
453 * dbus: in fedora, make /var/lib/dbus/machine-id a symlink to /etc/machine-id
455 * dbus: move dbus to early boot
457 * save coredump in Windows/Mozilla minidump format
459 * support crash reporting operation modes (https://live.gnome.org/GnomeOS/Design/Whiteboards/ProblemReporting)
461 * clean up session cgroups that remain after logout (think sshd), but eventually run empty
463 * when an instanced service exits, remove its parent cgroup too if possible.
465 * default to actual 32bit PIDs, via /proc/sys/kernel/pid_max
467 * be able to specify a forced restart of service A where service B depends on, in case B
468 needs to be auto-respawned?
470 * when a bus name of a service disappears from the bus make sure to queue further activation requests
472 * tmpfiles: apply "x" on "D" too (see patch from William Douglas)
474 * for services: don't set $HOME in services unless requested
476 * hide PAM/TCPWrap options in fragment parser when compile time disabled
478 * when we automatically restart a service, ensure we restart its rdeps, too.
480 * allow Type=simple with PIDFile=
481 https://bugzilla.redhat.com/show_bug.cgi?id=723942
483 * move PAM code into its own binary
485 * implement Register= switch in .socket units to enable registration
486 in Avahi, RPC and other socket registration services.
488 * make sure systemd-ask-password-wall does not shutdown systemd-ask-password-console too early
490 * support sd_notify() style notification when reload begins (RELOADING=1), reload is finished (READY=1), and add ReloadSignal= then to use in combination
492 * support sd_notify() style notification when shutting down, to make auto-exit bus services work (STOPPING=1)
494 * verify that the AF_UNIX sockets of a service in the fs still exist
495 when we start a service in order to avoid confusion when a user
496 assumes starting a service is enough to make it accessible
498 * Make it possible to set the keymap independently from the font on
499 the kernel cmdline. Right now setting one resets also the other.
501 * and a dbus call to generate target from current state
504 - drop /.readahead on bigger upgrades with yum
505 - move readahead files into /var (look for them with .path units?)
506 - readahead: use BTRFS_IOC_DEFRAG_RANGE instead of BTRFS_IOC_DEFRAG ioctl, with START_IO
507 - readahead: when bumping /sys readahead variable save mtime and compare later to detect changes
508 - readahead: make use of EXT4_IOC_MOVE_EXT, as used by http://e4rat.sourceforge.net/
510 * add support for /bin/mount -s
512 * GC unreferenced jobs (such as .device jobs)
514 * write blog stories about:
515 - hwdb: what belongs into it, lsusb
516 - enabling dbus services
518 - how to make changes to sysctl and sysfs attributes
520 - how to pass throw-away units to systemd, or dynamically change properties of existing units
521 - how to integrate cgconfig and suchlike with systemd
522 - testing with Harald's awesome test kit
524 - how to develop against journal browsing APIs
525 - the journal HTTP iface
526 - non-cgroup resource management
527 - dynamic resource management with cgroups
528 - refreshed, longer missions statement
529 - calendar time events
530 - init=/bin/sh vs. "emergency" mode, vs. "rescue" mode, vs. "multi-user" mode, vs. "graphical" mode, and the debug shell
531 - how to create your own target
532 - instantiated apache, dovecot and so on
533 - hooking a script into various stages of shutdown/rearly booot
535 * allow port=0 in .socket units
537 * support systemd.mask= on the kernel command line.
539 * reuse mkdtemp namespace dirs in /tmp?
541 * recreate systemd's D-Bus private socket file on SIGUSR2
543 * Support --test based on current system state
545 * investigate whether the gnome pty helper should be moved into systemd, to provide cgroup support.
547 * maybe introduce ExecRestartPre=
549 * dot output for --test showing the 'initial transaction'
551 * port over to LISTEN_FDS/LISTEN_PID:
552 - rpcbind (/var/run/rpcbind.sock!) HAVEPATCH
556 - libvirtd (/var/run/libvirt/libvirt-sock-ro)
557 - bluetoothd (/var/run/sdp! @/org/bluez/audio!)
560 * fingerprint.target, wireless.target, gps.target, netdevice.target
562 * io priority during initialization
564 * drop cap bounding set in readahead and other services
570 - natively watch for dbus-*.service symlinks (PENDING)
571 - allow specification of socket mode/umask when allocating DBusServer
572 - allow disabling of fd passing when connecting a AF_UNIX connection
573 - allow disabling of UID passing for AUTH EXTERNAL
574 - always pass cred data along each message
575 - teach dbus to activate all services it finds in /etc/systemd/services/org-*.service
577 * fix alsa mixer restore to not print error when no config is stored
579 * make cryptsetup lower --iter-time
581 * patch kernel for xattr support in /dev, /proc/, /sys?
583 * NTP: the kernel's 11-minutes-mode syncs the system time to the RTC, but only
584 in an ~30 minutes window. It does not adjust larger differences. Find a way
585 to tell the kernel, to always do a full time sync when the RTC is in UTC and
586 we are in 11-minutes-mode. When we trust the system time to NTP we also want
589 * kernel: add device_type = "fb", "fbcon" to class "graphics"
592 - plot per-process IO utilization
593 - group processes based on service association (cgroups)
594 - document initcall_debug
598 * look for close() vs. close_nointr() vs. close_nointr_nofail()
600 * check for strerror(r) instead of strerror(-r)
602 * Use PR_SET_PROCTITLE_AREA if it becomes available in the kernel
604 * %m in printf() instead of strerror(errno);
608 * set_put(), hashmap_put() return values check. i.e. == 0 doesn't free()!
610 * use secure_getenv() instead of getenv() where appropriate
612 * link up selected blog stories from man pages and unit files Documentation= fields
614 Scheduled for removal or fixing:
616 * xxxOverridable dependencies (probably: fix)