Horgoš...
On 18th worked on indexing, i.e. a routine which would generate all the indexes from scratch, without trying to rebuild the existing indexes, because if anything got fucked up, there's no guarantee that the definitions of indexes remained untouched, so a regenerator wouldn't find the definitions from which to rebuild. So better have those defined in code in advance, before anything happens.
Removed yet another field from opisfmt.dbf because opisfmm4.prg isn't using it - in Sale's ancient version it was the ordinal number in the data dictionary, which he may have used for something, I don't need it.
I see I inserted a Ula somewhere, some window somewhere in the session entry in PolC is called after her. Well not really, it's not any monument, there's just one of the words in the context which began with these letters, and we were abbreviating things heavily. Hungarian language, with 14 vowels, can create many different syllables and the words can be quite short, but in practice they aren't because they get sometimes four or more suffixes in a row, or several words are stuck together to make new ones. For one, távirányító is a remote control - táv being far, irány a direction, ít a suffix to make a verb, ó to make a tool or performer of a verb. So we abbreviated a lot while trying to keep the meaning, which mostly worked, even now I can remember the meaing of most of the things. Most, I said.
On that 18th I started developing the core PolC, i.e. the session form. There were two, initially, SESHI and SESHN, but the former somehow vanished later, no trace and no idea what was it supposed to do (found it - it's histology!). The latter is the session entry form, recording who were the doctor and the patient, which room, when did it start, what was the diagnose (from the short list, always extendable and the big list always at hand), what was done (same for operation codes, with the extra facility to have operation groups, so anywhere between two or ten operations could have been picked at once, even less typing), comment, date to appear for a checkup, ... well, everything.
Solved a stupid bit in catal6 on 20th - the 23rd line (counting from zero) would mess with the previous content of the screen, so I saved the whole screen first, then restored it when exiting. Now I found that I could have a window on 23rd line, height one, no border (border takes a whole line), width 80, and on exit it just vanishes and the previous content is untouched.
In opisfmm4 in GenerAll I had to Set Exact On, so when looking for a record under hungarian collation (and serbian too, though it didn't exist, I used czech, almost good enough) it would distinguish between upper and lower case. Without it it just happened that it picks the function to refresh some field and attaches it to the first field, which was wrong. The tables having codepages was a novelty in fox 2.6, and while it solved several problems, it also created a few glitches. Of course, until retirement I never solved the problem of "almost good enough" collation for serbian, in neither latin nor cyrillic. Good enough was the best there was.
It included a new bug too, which passed unnoticed until it created larger problems - nothing stopped us from creating a collated index on a number or a date or anything binary, which meant that the values wouldn't appear sorted by their real values, but rather by their translated-for-collation values, which made some records inaccessible altogether, and the order of several others in three mother's cunts. Took me a while to understand what was going on.