To work, via Horgoš.
Next day gen_cat6 (generator for catal6) got some parameter checking. Also messed with nmenu.prg, tmenu.prg, something in MEC for that doctor in Székesvár, and (I guess Brlja did) something for Ledinje.
Another neat feature was the attempt to have a docstrimg - if a piece of code in GenerAll began with a comment, which would begin and end with a *+ line and end with *-, anything between these lines would go into the .doc field in the menu table, in records from which such a piece of code was called (directly or as a parameter for a form or a catal6 generated routine or a few other cases). While it wouldn't fill much, it would at least help fill the help.
In PolC, local indexing routine - for the local tables that each workstation had, then for recoding patients (those entered with temp code, or where a mistake was made), so it would just change this code. These would happen when a patient is a baby (no state issued ID number yet), unconscious (so ID unknown) or a foreigner (so they can't decide what to use). Then there was a special add-on for the cases when they send a patient to a CT out of town, seems to be something insurance related or their own CT couldn't handle. Ah, yes, those guys had MR too, that's why.
There were two troubles with the lady in Sekešvar. The first, worse and unsolvable (unsoluble!) was her coffee. Thin, too sweet and just bad, and also unavoidable, she knows that coffee is programmers' fuel.
The other was skimping on hardware, so her disk was never big enough, which she compensated with compression. The compression gimmick was the fashion of the times (well, at least two years ago), and the Stacker company was famous with it for a while. The Stacker app would keep the filesystem zipped, which wasn't bad for many things, specially if your work mostly consisted of saving an odd word document six times an hour. Those documents contain oodles of unused or sparsely populated space, so they compress really well, super, great... But when you have a database and churn records in it from two more workstations, updating it twice a minute, and you don't use a genuine Stacker but stoop low for whatever piece of shit m$ put in its latest Dos (six, or was it still the five?), you risk a total fuckup of your disk content. m$ was just then in one of its phases when it sits and watches which of the add-on makers will strike it rich, and then buys or recreates the app and inserts it into Dos. It touts that as an epic technological breakthrough, disregarding the fact that everybody knows that's neither new, nor theirs, nor a breakthrough... but the „everybody“ is the IT crowd only. The rest of folks will believe whatever m$'s advertising says. Eh, the power of it.
I spent a nice few hours pulling her machine out of the shit, and she was happy for that one day. Not so much later, don't know why - whether there was some discrepancy between promises, lack of visits and possibly a snag with the payment, I never found out, and, frankly, was glad that the issue remained silent. This wasn't a happy app.
In unosc.prg a bug: „in _u_pregled it calls ref_hdr, where _msgcol and _statcol are mentioned before they are filled. I put _msgcol=_u_mscp and _statcol=_u_stcp“. This somehow worked before, must be I added something recently and broke it.