process_status() returns hashref not arrayref - changed to keep processes grouped under ppid - previously arranged by pid ascending order; adapted kidreaper.pl to switch to JUST_TESTING mode if max_mb omitted; added gender to record display
kill process status radio box not checkbox - only kill 1 at a time; redirect after killing process status (inconsistent, sometimes fatal); adapted kidreaper.pl for cron with multiple service names
exempted print-run from Devel::Cycle analysis due to fastcgi timeout; brought rss back into process size display; switched process size data from array to hash for more control in tt rendering; amended local css location in print_view.tt
used kidreaper code in Resources::process_status() code to get more accurate ownership of child processes after server kills & restarts a child process (due to resource exhaustion?); using IPC::System::Simple for capture() & system() commands
* used Capture::Tiny in Local::DevelCycle - works OK with fastcgi process dumping fastcgi captured cycles to logfile, test harness & dev box fatal * user_agent string captured using HTTP::BrowserDetect
fix for brain-dead browser - repeating jQuery requires unqiue string (eg current timestamp) to prevent cacheing of result - was preventing page update with new_message result in IE11
* new messages tab * patched Local::DevelCycle to suppress "Unhandled type: REGEXP" warning so test scripts can use find_cycle() * bug fix to user_messages.tt - div_name clashed with css div name
adapted new_messages.cgi to use Template to render same tt as ajax/seek_new_messages; adjusted white-space above h2.header at top of (most) HILIS pages
reverted new_messages.pl print header() location - was causing error; enabled Devel::Gladiator arena_table for debugging HILIS4 memory allocation; new kidreaper.pl to kill processes by cron
