MantisBT - Zandronum
View Issue Details
IDProjectCategoryView StatusDate SubmittedLast Update
0002059Zandronum[All Projects] Bugpublic2015-01-09 02:282018-09-30 23:57
ReporterDusk 
Assigned ToTorr Samaho 
PriorityhighSeveritymajorReproducibilityalways
StatusclosedResolutionfixed 
PlatformOSOS Version
Product Version2.0-beta 
Target Version1.4Fixed in Version1.4 
Summary0002059: Dying by hitscan results in undead state online
DescriptionIt appears that when a bot kills the player by hitscan online, the player may wind up in an undead state, where their health is 0 but can still move around and (in some cases anyway) interact, even pick up health and get a positive health again.

It doesn't appear to happen with projectiles or offline, but this is erratic enough that I really cannot say for sure.
Steps To Reproduce* zandronum -host -file undead_state_test.pk3 -deathmatch -warp 1
* zandronum -file undead_state_test.pk3 -connect localhost

Add a bot and let it kill you with the pistol, the undead state should trigger eventually.
Additional Information
TagsNo tags attached.
Relationships
Attached Files? undead_state_test.pk3 (574) 2015-01-09 02:28
https://zandronum.com/tracker/file_download.php?file_id=1370&type=bug
png Screenshot_Doom_20150109_042749.png (106,548) 2015-01-09 02:28
https://zandronum.com/tracker/file_download.php?file_id=1371&type=bug
png
Issue History
Date ModifiedUsernameFieldChange
2015-01-09 02:28DuskNew Issue
2015-01-09 02:28DuskFile Added: undead_state_test.pk3
2015-01-09 02:28DuskFile Added: Screenshot_Doom_20150109_042749.png
2015-01-10 11:46Torr SamahoNote Added: 0011349
2015-01-10 11:46Torr SamahoStatusnew => feedback
2015-01-10 11:47Torr SamahoNote Edited: 0011349bug_revision_view_page.php?bugnote_id=11349#r6368
2015-01-10 13:35Torr SamahoNote Edited: 0011349bug_revision_view_page.php?bugnote_id=11349#r6369
2015-01-10 13:38Torr SamahoNote Edited: 0011349bug_revision_view_page.php?bugnote_id=11349#r6370
2015-01-10 13:38Torr SamahoNote Edited: 0011349bug_revision_view_page.php?bugnote_id=11349#r6371
2015-01-10 14:03Torr SamahoNote Added: 0011351
2015-01-10 14:03Torr SamahoAssigned To => Torr Samaho
2015-01-10 14:03Torr SamahoStatusfeedback => assigned
2015-01-10 14:34cobaltStatusassigned => needs testing
2015-01-10 14:34cobaltTarget Version2.0 => 1.4
2015-01-10 14:34cobaltDescription Updatedbug_revision_view_page.php?rev_id=6373#r6373
2015-01-10 14:34cobaltSteps to Reproduce Updatedbug_revision_view_page.php?rev_id=6375#r6375
2015-01-10 14:34cobaltNote Added: 0011353
2015-01-10 14:36Torr SamahoNote Added: 0011354
2015-01-11 10:54Torr SamahoNote Edited: 0011354bug_revision_view_page.php?bugnote_id=11354#r6382
2015-01-14 12:17DuskNote Added: 0011372
2015-01-14 12:17DuskStatusneeds testing => resolved
2015-01-14 12:17DuskResolutionopen => fixed
2015-01-14 12:17DuskFixed in Version => 1.4
2015-01-14 12:17DuskSummaryDying by bot hitscan results in undead state online => Dying by hitscan results in undead state online
2015-01-14 12:17DuskDescription Updatedbug_revision_view_page.php?rev_id=6384#r6384
2015-01-14 12:17DuskSteps to Reproduce Updatedbug_revision_view_page.php?rev_id=6385#r6385
2015-01-14 12:18DuskNote Edited: 0011372bug_revision_view_page.php?bugnote_id=11372#r6387
2018-09-30 23:57Blzut3Statusresolved => closed

Notes
(0011349)
Torr Samaho   
2015-01-10 11:46   
(edited on: 2015-01-10 13:38)
I could reproduce this locally. Unfortunately, this only seems to happen rarely (for me it was roughly once in 5 to 10 deaths). Can somebody check whether this only happens with bots also or also with human players? And can somebody track down in which revision this broke (after confirming that this didn't happen in 1.3, but I'm pretty sure we would have heard about this, if it did).

EDIT: After going into the undead state, the bot immediately shoot me again, which caused a "proper death".

EDIT2: This does happen in 1.4, too. EDIT3: But at least it did not happen during a short test in 1.3
(0011351)
Torr Samaho   
2015-01-10 14:03   
I think this is caused by Water's puff prediction patch. I'll look at this in detail.
(0011353)
cobalt   
2015-01-10 14:34   
Issue addressed by commit 839cd8f167c8: Fixed: In case "cl_hitscandecalhack 1", the client erroneously handled damage of hitscan attacks on its own (fixes 2059).
Committed by Benjamin Berkels [Torr Samaho] on Saturday 10 January 2015 15:25:22

Changes in files:
 docs/zandronum-history.txt | 2 +-
 src/p_map.cpp | 6 ++++++
 2 files changed, 7 insertions(+), 1 deletions(-)
(0011354)
Torr Samaho   
2015-01-10 14:36   
(edited on: 2015-01-11 10:54)
Should be fixed now. It was caused by the client side prediction of bloodsplatter effects, which didn't exit P_LineAttack early enough.

EDIT: In particular, this was not limited to bots as attackers.
(0011372)
Dusk   
2015-01-14 12:17   
(edited on: 2015-01-14 12:18)
Nice catch. It appears fixed to me.