Adjustments related to the inactivity timer. The TimeClock Summary update triggered by the inactivity timer established a database connection that wasn't subsequently released. This messed with all the work that just went into making the database connection graphs nice and pretty ;-)
Changed how punch-ins via fingerprint work, to more closely match how they work without fingerprints. Seems still some issues around what happens when scanning in after scanning out, that kind of thing. Might not be there yet, will see...
Logging more information about punches generally, fingerprint scan punches specifically.
Related to the above, more work on first punch protocol.
Adjusted DB timeouts to be a little longer so that they don't timeout during a server restart.
Adjusted TimerCheckIn calculation - was not calculating the 10 min (or whatever) checkin delay properly.
Updated EurekaLog to 7.7.8.58. Not specifically needed for anything, just the latest version after a major TimeClock update.
Adjusted the rules for \"First Punch Protocol\" to have a 901 punched after a 902 in the case of a returning worker. Seems this was the way it was, and got mixed up with the PSF 24hr stuff.
Updated to accommodate longer than 8-character usernames. Only impacts Supervisor Functions, and even then only in very specific ways for some reason. Probably should look into that more.
Updated Access History logging to be more... interesting. Can now see individual TimeClock Attendance selections, more performance data.
Changes in how the Fingerprint Bypass work.
Probably other things I've forgotten. It's been, like, 2 days.
Revamped how lists are drawn, IE activities, locations, checklists, etc. to be better/faster.
Revamped how Supervisor Functions are drawn to be better/faster and at regular resolution. Drawing text after scaling background image rather than before.
Lots of work around how logging is done, more logging entries, more timing added to the log, more, More, MORE.
Changed around how network failures are managed. Hopefully more resilient than it was before.
Played around a bit with all kinds of timeout issues, hopefully going to dispense with the nightly bump on the graphs as the server is rebooted.
Lots of work around improving the performance of the Attendance screens.
Added a \"TimeClock Summary\" section at the top-right of the main TimeClock screen. Shows who is working nearby.
Changed how TimeClock GH and PH defaults are calculated. Should work better for raw punches, which will help with the TimeClock Summary.
Changed around how the background images and the rest of the image processing is handled. Much more performance and more accuracy in everything. Replaced some of the underlying background images with more options for making the 3D look more palatable.
Work on the ThankYou screen to better position photos without changing their aspect ratio (tricky given that the aspect ratio is different after it is painted....) and added more info to the ThankYou message to indicate data input and the processing of the punch.
Work on improving overall startup time. Now around 5sec over the VPN. Should be much faster locally.
Probably more that I can't think of off of the top of my head.
Minor update for checklists. When recording a checklist, the number used is recorded in the SKU field that is subsequently visible in the Greenhouse Viewer and other places. This value had been entered previously as \"PICK-203\" or whatever number was selected. This update changes it to use the first four letters of the type of checklist, currently either FORK, ELEC or PICK. Should help distinguish what checklist was being done when looking at the labour data, should anyone be interested in that.
Minor change to the queries that are first run after a long network disconnect - the BUZZER and M09 queries. The BUZZER query is first run after a CheckIn has started. M09 is first run when someone punches in, which may come after a long disconnect.
These queries were both updated with a different 'Error Message' so that if an error occurs, it is reported visually as a network error rather than a query error. These aren't complex queries, so it is unlikely that the query has failed but far more likely that the network has failed.
Normally this failure will recover eventually when the network access is restored, but the error is still presented on the screen for some reason as an 'Excessive Error' type of thing. Maybe this is an issue to take up in the next EurekaLog update.
The intent here is to give an error more indicative of the problem. Very few TimeClocks use the Buzzer system for example (currently ONE as of this writing I think) and ERR-M09 sounds more cryptic than it needs to be.
Fixed problem with Checklists not being recorded at all. Also added checklist output to the \"main\" TimeClock log so at least there's a record if the query fails in this fashion again.
Wrapped some exception handling around the code that plays sounds to see if it helps with the random exceptions that are generated. An odd place to have an exception it seems.
Added some extra code to the logging that checks for updates. Seems sometimes there is a StrToFloat error, suggesting that the database query isn't always returning data. Which doesn't make much sense, but the logging should indicate if that's actually the case or not.
Pretty major update to many things. First up, the code that draws the list of activities was updated. Seems that the labels were all overlapping one another, making the drawing of the menu excruciatingly slow even on modern computers. Nice and fast now, just by trimming the size of the labels.
Hid everything to do with 'messages' appearing at the top-right of the TimeClock - they weren't being used, and just interfered with the screen updates.
Updated Checklist system. Bit of an overhaul really. Now properly supports multiple checklists, and handles things better when multiple checklists are available, no checklists, checklists without questions, and so on. Pretty major overhaul of how checklists are started here.
Changed the RESTART mechanism for upgrades, hopefully will dispense with the endless looping upgrades after a new version is installed that needs to be deployed more quickly.
Fixed some miscellaneous bugs in code that comes up from time to time - moving around RunQuery invocations so that they are closer to the query processing.
Changed the Mantis interface to add the Database/TimeClock number to the BugAccount field to make it easier to see which TimeClock is affected.
One more time with this bleepin query! The thing that was fixing the \"worker last punched\" query seemed to be ok, but apparently some inline comments are not compatible with some installed versions of the DB2 client. Go figure. Anyway, this update is merely to remove those comments from the query. Hasn't really been a habit, adding comments to queries, but I guess it isn't likely to become one anytime soon.
Another change to the mechanism used to check for whether the worker is present or not, complicated by the 24-hr processing system. The last iteration made it possible to punch a worker who had punched out within the past 12 hours, so this update makes it so that the worker is considered punched out if their last activity was, well, punch out. No clue why this hasn't been the case for the past 20 years, but there you have it.