Changelog for s6. In 2.7.2.0 ---------- - s6-ipcserver-socketbinder now takes a -B option to make the socket blocking. - s6-svscan now takes a "-d notif" option to notify readiness. This is *shallow readiness*, i.e. readiness will be reported when s6-svscan is ready to receive s6-svscanctl commands; it does *not* mean all the services started at s6-svscan launch will be ready. - New optional file in a service directory: down-signal. It allows you to customize the signal sent by s6-svc -d (and now s6-svc -r too). - New s6-svc option: -r. Sends the signal defined in the down-signal file (still SIGTERM by default). In 2.7.1.1 ---------- - Bugfixes. In 2.7.1.0 ---------- - Bugfixes. - s6-svwait et al. now exit 102 instead of hanging if s6-supervise dies. - New commands: s6-svdt, s6-svdt-clear, s6-permafailon - s6-tai64nlocal can now print GMT times with the -g option. In 2.7.0.0 ---------- - Bugfixes. - s6-ftrigrd and s6lockd internally use the textclient API/ABI instead of skaclient. This is cleaner, faster code. - New "-a perms" option to s6-ipcserver and s6-ipcserver-socketbinder. In 2.6.2.0 ---------- - Bugfixes. - New functions: s6_fdholder_start() and s6_fdholder_end(). In 2.6.1.1 ---------- - Bugfixes. In 2.6.1.0 ---------- - Bugfixes. - Addition of the s6_svc_lock_take() and s6_svc_lock_release() functions, to fix a race condition in s6-rc. - New program: s6-notifyoncheck, to implement polling for readiness and reroute the result into the s6 readiness notification mechanism. In 2.6.0.0 ---------- - Bugfixes. One of the bugfixes required a visible ABI change, which explains the major version bump. - Documentation fixes. In 2.5.1.0 ---------- - Bugfixes. - s6-ftrig-listen1 now prints to stdout the last event (the last byte) it received and made it match the regexp it was given. - s6-supervise now recognizes the "timeout-kill" file in a service directory. If this file exists and contains a nonzero integer t, the behaviour of a "s6-svc -d" command is changed from "send a SIGTERM and a SIGCONT" to "send a SIGTERM and a SIGCONT, then if the service is still not dead after t milliseconds, send a SIGKILL". In 2.5.0.0 ---------- - Added this NEWS file. :) - Major types overhaul to make them more POSIXly correct: compatibility with skalibs-2.5.0.0. - Services can now report permanent failure by having a ./finish script exit 125. - s6-svwait, s6-svlisten and s6-svlisten1 now exit the number of services that reported permanent failure. They exit 99 on timeout. - When a service has a ./nosetsid file that contains "setpgrp", it is run in a new process group, but not in a new session. If the ./nosetsid file contains anything else, the service is run in the same session and process group as the supervisor, and if there is no such file, it's run in a new session (same behaviour as before). - s6-log now writes a warning to stderr when it closes the stdout stream. - s6-log's individual file size limit bumped to 256 MB. - Various bugfixes. (thanks jjk!)