simet-lmapd 0.15.2

simet-lmapd 0.15.2:

• minor Debian packaging fixes: enforce xz compression, some build environments do not support zstd yet.
• minor Debian packaging fixes: adjust maintainer name in debian/control

simet-lmapd 0.15.1:
( based on lmapd 0.4.1 commit 9c7a562 )

• lmapd: runner: run actions on their own process groups
  Run each task (lmap action) in their own process group, and and send SIGTERM to the whole process group when forcefully stopping it.

simet-lmapd 0.14.0 to 0.15.0:

• lmapctl: report: do not add incomplete/invalid report data.
  When generating a lmap report, ensure that only task output (output of an lmap schedule's action) that has been fully loaded without parsing and validation errors is accepted.
• lmapd: json-io: better error message for truncated input data.
• lmapd: workspace: restrict a schedule's active queue to regular files.
• lmapd: workspace: only move regular files to destination schedules
• lmapd: workspace: protect private namespace for actions too.
  Skip the private namespace (directories and files prefixed by "_") when moving action results and cleaning action workspaces. This gives actions a way to preserve state across separate runs of a schedule. Hidden files are also skipped when moving an action's result to destinations, but they are not preserved (they will be removed by lmapd_workspace_action_clean()).
  Note that actions cannot depend on the workspaces not being fully cleaned up when lmapd is restarted or reconfigured, as such workspaces depend on the lmapd task and schedule configuration.
• lmapd: runner: cleanup action workspace at schedule start.
  Cleanup any leftover data inside action workspaces of a schedule right before we start executing the schedule.
• lmapd: json: fix missing O_RDONLY on parse_file()
• lmapd: try a bit harder to detect stale pidfiles
  Ignore illegal PIDs in a pidfile (negative or zero), and check if a process with the PID read from the pidfile actually exists before considering it valid
• pidfile: remove stale comment, we don't use flock()
• lmapd: defend against internal errors that call event_free(NULL)
• lmapd: avoid crashes/assertion for too large random spreads on periodic events
• lmapd: refuse events of type periodic with a random spread equal to, or larger than its interval (period) as invalid configuration
• lmapd: defer the move of an action results until the whole schedule has run.
  This ensures the output of all actions will be in the same report, for example. We preserve the status-quo for now, and partial results of suppressed schedules with the stop-running flag are still going to be delivered to their destinations. That might change in the future.

simet-lmapd 0.15.1

simet-lmapd 0.15.1:
( based on lmapd 0.4.1 commit 9c7a562 )

• lmapd: runner: run actions on their own process groups
  Run each task (lmap action) in their own process group, and and send SIGTERM to the whole process group when forcefully stopping it.

simet-lmapd 0.14.0 to 0.15.0:

• lmapctl: report: do not add incomplete/invalid report data.
  When generating a lmap report, ensure that only task output (output of an lmap schedule's action) that has been fully loaded without parsing and validation errors is accepted.
• lmapd: json-io: better error message for truncated input data.
• lmapd: workspace: restrict a schedule's active queue to regular files.
• lmapd: workspace: only move regular files to destination schedules
• lmapd: workspace: protect private namespace for actions too.
  Skip the private namespace (directories and files prefixed by "_") when moving action results and cleaning action workspaces. This gives actions a way to preserve state across separate runs of a schedule. Hidden files are also skipped when moving an action's result to destinations, but they are not preserved (they will be removed by lmapd_workspace_action_clean()).
  Note that actions cannot depend on the workspaces not being fully cleaned up when lmapd is restarted or reconfigured, as such workspaces depend on the lmapd task and schedule configuration.
• lmapd: runner: cleanup action workspace at schedule start.
  Cleanup any leftover data inside action workspaces of a schedule right before we start executing the schedule.
• lmapd: json: fix missing O_RDONLY on parse_file()
• lmapd: try a bit harder to detect stale pidfiles
  Ignore illegal PIDs in a pidfile (negative or zero), and check if a process with the PID read from the pidfile actually exists before considering it valid
• pidfile: remove stale comment, we don't use flock()
• lmapd: defend against internal errors that call event_free(NULL)
• lmapd: avoid crashes/assertion for too large random spreads on periodic events
• lmapd: refuse events of type periodic with a random spread equal to, or larger than its interval (period) as invalid configuration
• lmapd: defer the move of an action results until the whole schedule has run.
  This ensures the output of all actions will be in the same report, for example. We preserve the status-quo for now, and partial results of suppressed schedules with the stop-running flag are still going to be delivered to their destinations. That might change in the future.

simet-lmapd 0.13.3

0.13.x: pre-release with support for tasks that output JSON multi-table results, and data-loss prevention on queue data-flow.
0.13.1: Minor fix to get correct header definitions on non-glibc systems.
0.13.2: Minor fix to ensure we never try to build the doxygen documentation since that build is failing right now.
0.13.3: Fix the report of ma-report-result-event-time (upstream bug). It was reporting ma-report-result-start-time, instead.