. /Noctis/Known Noctis IV Bugs, and.../ MESSAGE #63526
absent more than alex
written by Shadowlord on Aug 20, 2005 20:36
Okay, now here's R9b, which is also mostly a bugfix release:

Changes in R9b:
  • Fixed a bug in R9a in the sky() function, which could cause problems at or near star-density boundaries (and maybe elsewhere? Didn't check that, just fixed the bug). Among other things, this caused the "ACK! CAN'T FIND STAR" message to appear. When you start R9b, you'll get a scrolling message if this problem is affecting you. It'll tell you what to do (re-target and vimana back to the star you're at now, or if that doesn't work, target another star and vimana there). Any stars or planets you labelled with R9 or R9a will not be lost. More detail about the cause: In R9 all the different code which determined star coordinates were reconciled so that they all returned the same coordinates as each other. The sky() function is one of those. FSAT() is another one. FSAT() is used to find the sector coordinates of a star, and this is needed for labelling stars and planets. In R9a, noctis-0.cpp and noctis-1.cpp were split into separate DOS and windows versions, and all the ported code in the DOS version was reverted to R8.5, although I tried (and largely succeeded) to keep all the R9 features and bug fixes. However, I didn't notice or remeber that sky() had changed for R9 (other than being made windows-compatible), so the R8.5 version of sky() replaced the R9 version. Long story short, the coordinates returned by the R9a sky() didn't match the ones returned by FSAT(), which is what caused the problems. (Once I tracked down the problem, I just re-replaced sky() with the R9 version again to fix it, and added some detection and message code to explain how to recover from the error)
  • Fixed the slowdown around moons (particularly when the terrain-filter (the 'x' key) was on). The 'calculate-the-surface-of-these-planets-or-moons' variables (resident_map_1 and 2) weren't being cleared for moons after their surface was calculated. Thus their surface was recalculated every cycle. On a fast computer, this was only noticeable if the terrain filter was on.
  • Improved the speed of the surface-calculating code for the terrain filter a bit when re-calculating a previously-calculated surface. It probably won't have any effect unless you're flipping terrain mode on and off repeatedly, and even then it might not be noticeable.
  • And these are changes to the source code which were made to make the previously mentioned fixes possible:
    • Changed FSAT to return -2 if the target system is found but does not really exist due to rarity. (-1 already meant it didn't find a system, and previously -1 was returned when this happened.)
    • Changed getSectorCoords to have a return value which indicates a problem. 0 is no problem, -1 and -2 mean the same thing as when FSAT returns them (This calls FSAT, after all)


Edit: This also has a feature Mega recently added. I'll let him tell y'all about that.

Ataribaby said:
bug in R9a? when i start NICE first time and wanna set remote target then name of remote target stay stuck on first star what i hit and not change when i moving remote target pickup cursor browsing near stars via menu work ok
Let me know if this still happens in R9b?
└> last changed by Shadowlord on Aug 20, 2005 at 21:24
reading this thread
(not applicable to single message display)
. /Noctis/Known Noctis IV Bugs, and.../ MESSAGE #63526
14528, 7 queries, 0.044 s.this frame is part of the AnyNowhere network