summaryrefslogtreecommitdiff
path: root/src/libs6
AgeCommit message (Collapse)Author
2020-12-18 More fixes to s6_supervise_linkLaurent Bercot
2020-12-11 Only realpath() if a given servicedir is not absoluteLaurent Bercot
2020-12-10 Add the possibility of down files to s6_supervise_linkLaurent Bercot
2020-12-10 bugfix: sarealpath() APILaurent Bercot
2020-12-10 Use mkptemp2Laurent Bercot
2020-12-09 Get rid of webipc.hLaurent Bercot
2020-12-07 Do not use 'scandir' as a variable name, thanks posixLaurent Bercot
2020-12-05 Fix build with nsss; fix s6_supervise_link() for logged servicesLaurent Bercot
2020-12-04 Add s6_supervise_link and s6_supervise_unlink for future s6-rc changeLaurent Bercot
2020-12-04 Big signal/command semantics change to svscan/supervise; add s6-svperms.Laurent Bercot
2020-12-02 Fix build against skalibs-2.10.0.0Laurent Bercot
2020-11-16 Fix supervise/svscan startup when the control dir already existsLaurent Bercot
2020-10-26 Align on new names of skalibs functionsLaurent Bercot
2020-10-26 Prepare for 2.9.2.1; add dangling symlink supportLaurent Bercot
2019-12-23 Make execline dependency optionalLaurent Bercot
This includes: - adding a configure option to disable execline support - duplicating el_semicolon() and el_getstrict() into the s6 library when execline is disabled at build time, making a compat shim and using it where needed (s6-svlisten, s6-ftrig-listen) - Creating alternatives for execlineb script spawning: * s6-log: '?' directive (done in the previous commit) * s6-notifyoncheck: require hardcoding the check in ./data/check * s6-ipcserver-access: no support for exec files, add a warning - Updating the relevant parts of the doc --enable-execline will remain the default, and whiners can choke on their tears.
2019-09-21 Remove tainnow.lib dependencyLaurent Bercot
2019-09-06 Adapt to new stopwatch APILaurent Bercot
2019-09-04 Separate wall clocks from stopwatchesLaurent Bercot
2019-07-03 Remove s6-fdholder-*c programsLaurent Bercot
The doc still needs to be updated.
2019-05-11 Accept empty regexes even on BSDsLaurent Bercot
2019-03-25 Reestablish directory checking in s6-svc/s6-svscanctl without commandsLaurent Bercot
2019-02-19 Adaptation to skalibs/posixishard.hLaurent Bercot
2019-02-04 Revert -I, but add uid/self and gid/self to uidgid accessrules checkingLaurent Bercot
2019-02-01Don't deprecate ftrig1_init: remove it. (major bump inc anyway)Laurent Bercot
2019-02-01 The ftrig1 must be read nonblocking by s6-ftrigrd, duh!Laurent Bercot
GOD WHY AM I SUCH AN IDIOT
2019-01-31 Make "s6-svc -wu foo" wait for foo to be up. Document supervise/ creation.Laurent Bercot
2019-01-09 Use mkptemp() in s6-ftrigrd. Prepare for 2.7.2.3.Laurent Bercot
2018-07-21 More build fixes for skalibs-2.7.0.0Laurent Bercot
2018-07-21 Adapt to skalibs-2.7.0.0Laurent Bercot
2018-03-20 Add death tally support to s6-supervise, throttling support to s6_svstatus_tLaurent Bercot
2017-12-22 Port ftrigr and s6lock to textclientLaurent Bercot
2017-11-27 Add s6_fdholder_start/end, prepare for 2.6.2.0Laurent Bercot
2017-11-22 Add SKALIBS_HASODIRECTORY supportLaurent Bercot
2017-11-10 More accurate exit code and clearer error messages for s6-svc and s6-svscanctlLaurent Bercot
2017-09-12 bugfix: s6_accessrules_backend_cdb could incorrectly return ERROR instead ↵Laurent Bercot
of DENY
2017-08-19 Fix fd leak in s6_svc_lock_releaseLaurent Bercot
2017-08-06 Bugfix: typo in s6_fdholder_getdump.cLaurent Bercot
2017-07-16 Fix ftrigw_fifodir_make on platforms where gid_t is unsigned (the BSDs, who ↵Laurent Bercot
would have thought!)
2017-07-16 Propagate fd_close() changes, fix deps for ftrigr_updateb()Laurent Bercot
2017-07-16 Add ftrigr_updateb()Laurent Bercot
2017-07-12 Add s6_svc_lock_take() and s6_svc_lock_release()Laurent Bercot
Allows fixing a race condition in s6-rc-init. s6-supervise also creates event/ before supervise/ so that locking the servicedir ensures event/ can be used.
2017-06-14 Add ftrigr_checksa(), rewrite s6_svlisten_loop() around itLaurent Bercot
- Fixes the race condition hit by permanent failure, i.e. two ftrig events close to each other - Requires storing the sequence of events client-side, so an additional stralloc, bleh - The visible struct ftrigr_s changes, so a major bump is needed -> prepare for 2.6.0.0 - ftrigr_check() is now a trivial wrapper around ftrigr_checksa()
2017-06-13 s6-ftrigrd: be silent when the client dies in the middle of an asyncout flushLaurent Bercot
Also add ./configure output to .gitignore
2017-06-13 bugfix: use memmove instead of memcpy in an overlapping stralloc in s6-ftrigrdLaurent Bercot
2017-03-13 Cleanup of superfluous includesLaurent Bercot
2017-03-12 Adapt to skalibs-2.5.0.0Laurent Bercot
2017-01-31 Actually fix the return value of ftrigr_update() (thanks jjk)Laurent Bercot
2017-01-30 bugfix: ftrigr_update() could return incorrect value for repeating events ↵Laurent Bercot
(thanks jjk)
2017-01-12 Types fix, first passLaurent Bercot
2016-10-16 Unconditionally switch s6-ftrigrd to surf (avoid boot stalling problems, ↵Laurent Bercot
and nobody cares about that RNG not being CS)