 |
What happens if you try this in 2.05? One of my requests for beta testing was to be sure whatever you're reporting doesn't also fail in 2.05. |
|
|
bump |
 |
I've dmapped from the console before in 2.05. It exits you from the mission but doesn't crash. |
 |
I have also regularly done the same with no issues. I have tested it just now on the latest package (testmapping the current map) and it is a direct crash. |
|
|
Goldwell has had the issue also - - http://forums.thedarkmod.com/topic/19162-beta-testing-206/page-17#entry416486 |
 |
Same here, the same behavior as Spooks. |
 |
Committed fix in svn rev 7353. During dmapping process, each console message triggers screen draw, so that user can see messages in realtime. This screen draw calls idRenderSystemLocal::EndFrame, which in turn runs game tics. Here is the full sequence of events: 1. Game shutdown is started (during dmapping process). 2. Entities are deleted one by one, the player is deleted first. 3. During one of the entities being deleted, console message is printed. 4. The idCommonLocal::VPrintf calls UpdateScreen because com_refreshOnPrint is enabled by dmap. 5. UpdateScreen calls EndFrame, ActivateFrontent, RunGameTics, game.RunFrame. 6. Something inside game.RunFrame tries to get player's location, so it crashes because player is NULL. This change fixes the problem on step 6: if player is null at the beginning of game.RunFrame, then function returns immediately. |
- Anonymous
