 Notes |
(0026198)
Kate (developer)
2014-05-18 15:21
|
Looking at it in debug mode, it looks like possibly what's taking so long to load the first time is running dactions on every floor in the game, not adjusting artefacts? |
|
(0026200)
Kate (developer)
2014-05-18 20:42
|
I bisected and narrowed down the huge delay on loading the game to ffe63d8: http://s-z.org/neil/git/?p=crawl.git;a=commitdiff;h=ffe63d874a0e [^]
_count_all_gold() looks like it is causing every level to be loaded and checked for gold. |
|
(0026205)
Kate (developer)
2014-05-19 02:08
|
Should be fixed, gold from all generated levels isn't counted on load now. Saves transferred from pre-Gozag will only have a 50 gold fee as a result, this seems probably fine though (and probably not a significant change anyway since previously only gold that had been left on the floor on old levels was counted, not gold that had already been picked up). |
|
(0026208)
Kate (developer)
2014-05-19 14:37
|
Hmm, apparently whatever was causing it to crash in Lair:2 is still happening. The save does load now since it's not visiting every generated level on load, but travelling to Lair:2 still crashes, so maybe it is an artefact problem after all. |
|
(0026245)
Kate (developer)
2014-05-21 00:40
|
Looking at the save locally, there is an artefact plate mail with the SustAb property in the Lair:2 stash, which shouldn't be possible at all (before 48ad150 it's a plate of {rC+ Dex+3}, afterwards it's {rC+ Dex+3 SustAb}). But I still can't get the save to actually crash locally with any combination of optimised/debug/whatever builds, so I'm not sure if that's actually the issue or not. |
|
(0026246)
Kate (developer)
2014-05-21 00:59
|
More notes, possibly entirely unrelated to the crash: statues all across the dungeon appear to be remembered as stone arches when looking at them on the X map. Travelling to look at them reveals that they are in fact still statues, though. |
|
(0029684)
Kate (developer)
2015-10-27 23:52
|
This Atomjack save file has a similar crash and contains the Staff of Olgreb {rPois SustAt}, so it seems likely that it must be somehow related to artefacts incorrectly getting the SustAt property on save upgrade??? |
|
(0029808)
st (reporter)
2015-11-16 21:42
|
http://crawl.develz.org/saves/dumps/mudo-2d311b1-151116-2041.tar.bz2 [^]
Here's another broken save |
|
Atomjack-CDO-20150922.cs [^] (1,850,443 bytes) 2015-10-27 23:49
MarvinPA-6f76ed4-140517-1631.tar.bz2 [^] (3,650,751 bytes) 2016-06-25 20:00 
