Stepping through the code

In my leisure. Hoping, mainly, that another open source version pops up somewhere that I can borrow good ideas. Eventually, in a day or so, I will have stepped through most of this. Taking time to let the industry come to some conclusions.

I fix bugs as I go, and noticed early that my lab cheats no longer work, everything has to be chained to syntax events between the dispatcher and callbacks. So the configuration has to be on the dispatch list and have a filter, then it has to chain over to the user script, another event. The TE script is now bound to the SQL - G interaction, the matching events. So I am going through that, but it all initializes correctly, and the dispatch routine actually runs, but badly at the moment.

What that means is that access to g internals, which normally occur below the syntax, will now have some slight interface to the syntax. So I will create an event called gspew, it has a snippet of code for sql events it might cause, g accidentally spewing a bit of the unwanted, perhaps we need a g logging snippet, creating logs of any changes made using g internals. I also have unconnected snippets of code for matching key values. Right now, all events are connected to the print triple micro action, stuff pops up on the screen.

The event list seems to expand. Overloaded link is an event, maybe calling one of the Dublin ore methods. URL leaps are events, table jumps. But still, I don;t see more than 8 major events, and probably just some variations on micro-methods on call back.
A table jump event, for example. The tables, in sql, are known apriori, so a table jump event is simply the standard statement on another table. All the priors of that table set up, whether it is square or nested or hybrid. Table jumping snippets are a snap.

Everything runs off the filter directed graph, the event grammar being the events masks and code of each ascending filter. This will work out, all the rest, jumping g, Dublin methods, higher level expansions and completion of nested stores. All that, just becomes that little spring loaded plunger on the pinball machine, a little snippet and your done or your off in dll methods for a while; entirely separate from the machine. The machine always trying to do just the necessary container management.

Table are done by chain reaction, when they are mentioned, we look a their info, if info is zero, then code fills info. Otherwise the table adapter has two gears, one for each of the no more than two binary tables. When the thing is square, the adapter spins out columns. After that, it is back to the key,link,pointer format. It is currently broken.

Hybrids, they have the [link.pointer][col1.col2.col3...,] form, great form, the machine can move this thing around.

GTPtraverse() //Graph Traversal Protocol

We need that web call!