1. Flat earth... Have you heard of the game
called "Whack-A-Mole"? Look it up...
2. CM93 chart outlines: This needs some discussion.
Gleaning chart (cell) outlines requires reading and decoding the cell completely, since we don't know a priori how many M_COVR objects are present. These are the objects which describe the cell coverage.
So, there is a performance issue. The question is, "How far down into the cell stack should we proceed in order to find useful smaller scale cell outlines"
Example: At your favorite location Minerva Reef, the cell outline we want to see is a D-cell. If we are looking at an A-cell, we must traverse the stack all the way to D in order to see the required outline. OK. no problem, since there is only one covering B cell, and no C cell at all at this location. Fast enough.
However, if we apply this logic worldwide, then the performance becomes unacceptable, since we will be reading B, C, and D cells whenever we look at A cells. I tried an experimental version. Zooming out at A scale shows lots of outlines, but is unusably slow.
We need to limit the stack drill-down in some sensible, useful way. Perhaps as some function of current
display scale? For instance, drill down far enough to display outlines at a scale of 25% of current
effective display scale?
I don't know the answer, and I solicit comments.
3. segfault at Rockall.
Seems to be a bug in GDK system code. I saw it once, but was unable to reproduce. Looking into it.