Reset SIGCHLD to default

[catagits/FCGI-ProcManager.git] / ChangeLog

diff --git a/ChangeLog b/ChangeLog
index c66ddbd..cfd5438 100644 (file)
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,43 @@
+0.23
+    * Fix pm_wait routine to exit without a warning
+    * Fix pm_wait to correctly detect child processes which have died.
+    * Fix to return SIGCHLD handler to it's default state in the
+      pm_manage loop so that loading code which sets SIGCHLD to ignore
+      won't interfear with normal operation.
+
+0.22 2011-08-04 Tomas Doran  <bobtfish@bobtfish.net>
+    * Ship a version of FCGI::ProcManager::Constrainted which actually
+      works as advertised.
+
+0.21 2011-08-02 Tomas Doran  <bobtfish@bobtfish.net>
+    * Ship a working dist.
+
+0.20 2011-08-02 Gareth Kirwan  <gbjk@cpan.org>
+    * ProcManager::Constrained: Add a new FCGI::ProcManager::Constrainted
+      subclass which will restart workers safely after a certain number of
+      requests, or at certain memory thresholds.
+    * ProcManager: Fix start_delay causing a permanent hang.
+
+0.19 2009-07-22 Gareth Kirwan  <gbjk@cpan.org>
+
+    * ProcManager.pm: Add a pm_title option to configure the process title
+      assigned to $0
+
+0.18 2007-12-18 James Jurach  <muaddib@erf.net>
+
+     * ProcManager.pm: Remove constraint that parent process cannot die.
+
+0.17
+
+       * ProcManager.pm: Remove constraint that process can't change pid.
+         This resolves the daemon exiting after first child dies
+         if the process id of the daemon has changed since starting.
+         This commonly happens when a script waits to check for the
+         socket being available and working before forking and exitting.
+         It can also happen in a race condition of fork && exit when
+         the exit is delayed until after the forked child has already
+         spawned the fastcgi handlers.
+
 2001-04-23 18:12  James Jurach  <muaddib@erf.net>
 
        * ProcManager.pm: SIGHUP handler now issues SIGTERM to children but