genomics sample linkage form; xsd validation for genomics_xml.pl; bugfix for audit/turnaround_percentile.tt - subtraction only performed if values exist

patched genomic_xml script - error in gel_id generation; moved receiving_system & receiving_facility hl7 items to object constructor; reverted QueryLog to return if fastcgi - logging process size was killing print run - change introduced on live system 23/3/16

bugfix - previous change to QueryLog stopped all sql logging by setting gglobal SKIP_LOG_QUERY inside code block

logging fastcgi process size

changed user message handling: * site.status.tt calls ajax/seek_new_messsages as jQuery on page load * site/status.tt calls new_messages.cgi every minute until 1 min before session timeout * set idle-timeout value in session user_profile, created at new login if no existing user profile * user session user_profile idle-timeout value in LIMS::_authen_config $authen_timeout sub * new $LIMS::Local::QueryLog::SKIP_LOG_QUERY to avoid logging specific queries (eg ajax/seek_new_messages) * updated new_messages.cgi with logging and formatted return msg to same as ajax/seek_new_messages

patient merge includes demographics data; moved mech.htm to scripts dir

added process size logging facility to fastcgi process; authorised/incomplete records worklist displays outstanding lab tests; added cd SCRIPT_DIR to fcgi script so git version info loaded when restarted from outside HILIS4 dir

minor adjustment to QueryLog formatting to make process-size easier to read

cosmetic changes to _warn_handler and minor performance enhancement by accessing elements of coderef directly, not via temporary vars (, , etc)

enclosed QueryLog in sub so it's only called when required and doesn't have to determine pid process size each time, has side-effect of dramatic speed up for dev server restarts when RDBO dumps 'Making method' output

process size timeout issues - added pid 3169 and pid size data to querylog to track growing process size problem; added centre and timestamp to fastcgi stdout; added auto-refresh to process status & logged-in users pages

RDBO debug method changed again - now have to specifically ask for $RBDO::Debug to and $RDBO::Manager::Debug using DDBO_DEBUG & RDBO_DEBUG_ON env flags - difficult to get desired outcome from fastcgi processes, crons, devel server and test scripts wrt log files and stderr

corrected mistake made in v1464 which switched off query logging in production mode; minor edits to new_referral_source.pl

changed LIMS::RDBO Rose debugging flags to require RDBO_DEBUG env param to be set - eg lims_server.pl, locally in test scripts, or export console ENV param

QueryLog BEGIN block looks for ENV var ROSEDB_DEVINIT = devinit_prod to load Dispatch::File::Rolling - previous method looked for devinit_devel & loaded Dispatch::File, otherwise Dispatch::File::Rolling. This failed in test suite if rolling sql.log already created for current date, with 644 permissions for Apache user. Now test suite and dev server both use same log files, and rolling log files only used by Apache user