Did a bunch of code changes around today's coffee roasting session. All little things where most of the functionality was already present and I just needed to hook things up to the UI. Also set my Mom's phone up to receive email since she'll want to have that capability on her next vacation for stuff she can't schedule before she's on the road. Sending doesn't work because I couldn't find the settings for using TLS for that on her device.
There are a couple minor tweaks that I still want to make, but I think it's all stuff that I had in the last public release that should be trivial to port into the new code base when I have the time to look at that.
Combined with interpolation in the reference columns, it's now easier to see how live data compares with the plan for people who want that expressed in numbers instead of pictures. Last night's change adds a bit of visual distinction (for now just a yellow background, but it's trivial to change that to several other styling options) to the row containing the latest measurements so it's a bit easier to spot that at a glance.
I think I'm probably just about done with enhancements to the table view for the next version of Typica. The two big changes from previous versions are that roast profile translation can now be reflected in the table instead of being a graph only feature and I've put in the option to show whatever the most recent measurement is whether that's a row that would otherwise be displayed or not.
So whatever was causing the problem in the first place is may still be hiding somewhere unfixed just no longer causing directly visible problems. I'll use it the next time I roast coffee and see how that goes, but this is something I'll want to come back to later to clean things up properly.
Just did my least favorite kind of bug fix. Gave up on trying to figure out why the erroneous behavior was happening (a variable was being updated with double the expected offset) and rewrote the code to just recalculate the value every time it's needed, at which point everything seemed to start working properly. It's a cheap calculation infrequently performed so it's not like anybody will notice, but I kind of hate that a change that shouldn't have altered the behavior at all ended up fixing it
I think I need to write some more tests. I've got a set of 3 features that individually work perfectly, but when combined, not so much. It's better than it was and close enough to right that it's still usable, but I'm clearly not thinking about this quite right and it takes too long to set up the faulty conditions in live use.
Removed the green coffee from the coffee roasting area. That's not a sign that I'm going out of business. New lights are getting installed tomorrow morning and I just wanted the stuff out of the way and not have the risk of stuff falling into the bags. The old lights were salvaged from another shop before getting installed here 26 years ago so they're quite elderly now and probably not worth paying to repair instead of getting something slightly brighter, more efficient, and easier to clean.
Author of Typica software for coffee roasters.