Doomsday Engine

2008-11-06

22:29 Bug #565: Post level Infine seg fault

Please can you attach the Infine scripts you are using to this tracker item. We will also need a step-by-step of how ... danij

21:11 Bug #565 (Closed): Post level Infine seg fault

If a map exits while running an infine def (i.e a fade out) and the map also has an infine def set to run after it. T... vermil

21:24 Feature #1475 (New): Spawnmobj "deaf" things

Spawnmobj can set the facing angle/height (position) of a thing and the actual def it is attached to can be set to on... vermil

21:14 Bug #566 (Progressed): [Resampler] Low quality sound samples play distorted

I know this is technically not a bug because sounds of such quality are not officially supported by Dday...
Dday c... vermil

16:28 Bug #541: Jittery view bobbing

Re-opened as there does still seem to be a problem.
To replicate:
Enable auto-run.
Use the default view bobbing sett... danij

2008-11-05

08:44 Bug #563: Use of uninitialised value of size 8

After an extended debugging session with Yagisan, we have reached the conclusion that this isn't a Doomsday issue. I'... danij

02:59 Bug #563 (Closed): Use of uninitialised value of size 8

SVN 6033 is causing valgrind to halt when starting jhexen. It halts at this point
Failed to locate high resolution... yagisan

05:05 Revision 112ba1a7 (github): Fixed: Bug #2223434, the player viewdata was not being initialized at all. R_NewSharpWorld is always called every tic regardless whether there is a map loaded or not and because there are no player mobjs at this time, this would result in memcpy'ing uninitialized garbage around.

danij

03:03 Bug #564 (Closed): Conditional jump or move depends on uninitialised value(s)

SVN 6033 is causing valgrind to halt when starting jhexen. It halts at this point.
GL_NewTexture: Uploading (469:1... yagisan

02:34 Bug #562 (Closed): Conditional jump or move depends on uninitialised value(s)

SVN 6033 is causing valgrind to halt when starting jhexen. It halts at this point
==2562== Conditional jump or mov... yagisan

02:27 Bug #561 (Closed): Conditional jump or move depends on uninitialised value(s)

SVN 6033 is causing valgrind to halt when starting jhexen. It halts at this point.
==2562== Conditional jump or mo... yagisan

02:18 Bug #559: jDRP: Chaingun invisible on autoswitch while firing

Its a duplicate item #746738: https://sourceforge.net/tracker2/?func=detail&aid=746738&group_id=74815&atid=542099
Th... danij

00:49 Bug #559: jDRP: Chaingun invisible on autoswitch while firing

I'm sorry, duplicate of what? I searched for this bug before filing it, but didn't manage to find anything. haarp

01:02 Revision 5746ae53 (github): Fixed potential segfault in XG which could attempt to dereference a NULL xline_t ptr in an XGTrav_* routine if passed a NULL linedef_t ptr OR if the ptr was to a dummy line (P_ToXLine would fail, returning NULL).

Cleaned up. danij

2008-11-04

22:51 Bug #554: Line_active xg not checking correct line(s)?

There were numerous problems in your XG Line Type definitions in the DD_DEFNS lump in Test.wad. After fixing the prob... danij

22:48 Bug #554: Line_active xg not checking correct line(s)?

Fixed version of your Test.wad
*Attachments:*
* http://sourceforge.net/p/deng/bugs/_discuss/thread/3c2142a5/b5ca/att... danij

19:45 Bug #560: jDRP: Weapons point too far to the left

It is not the weapon models at fault here as they are positioned to match the original HUD sprites. The real issue he... danij

13:10 Bug #560 (Closed): jDRP: Weapons point too far to the left

If crosshairs are enabled, the weapon models all point a bit left of the crosshair. This becomes apparent when holdin... haarp

19:41 Bug #559: jDRP: Chaingun invisible on autoswitch while firing

Closing as duplicate. danij

13:03 Bug #559 (Closed): jDRP: Chaingun invisible on autoswitch while firing

If you're engaging a bunch of enemies, it may happen that you hold down the fire button and keep it pressed. When you... haarp

12:56 Revision 7b179649 (github): Seems I've been doing too much PHP lately.

danij

12:30 Revision ddc60b28 (github): Use a more robust method of formatting material names prior to hashing.

danij

12:09 Revision cc4c4345 (github): Fixed same silly bug but this time in R_MaterialCreate. Must have used a "replace all instances of term" or something...

danij

12:01 Bug #558: Segfaults when compiled with -O3

I forgot to mention, I'm using Yagisan's Launchpad sources on AMD64 haarp

12:00 Bug #558 (Closed): Segfaults when compiled with -O3

When Doomsday is compiled with CFLAGS containing -O3, it will become much more prone to segfaults. For instance, Hexe... haarp

11:51 Bug #544: jDRP models facing wrong direction?

Projectiles are also affected by this, like the fireballs of the baron of hell. They move in the proper direction, bu... haarp

11:43 Bug #544: jDRP models facing wrong direction?

I can confirm this, using the AMD64 version I compiled from Yagisan's Launchpad sources. haarp

11:30 Bug #557 (Closed): Conditional jump or move depends on uninitialised value(s)

SVN 6031 is causing valgrind to halt when roaming around on doom2 map1. It halts at this point.
==27647== Thread 1... yagisan

08:53 Revision cdff8e86 (github): Fixed a really silly bug in R_MaterialCheckNumForName.

danij

01:40 Revision c8ca51ef (github): Optimized sprite precaching in R_PrecacheLevel. Now only caches the materials of sprites used by states of mobjs actually present on the map. This is more conservative than the previous implementation which would attempt to precache every last sprite. Precaching of sprites is now enabled by default.

This still not ideal as sprites for mobjs created dynamically which weren't on the map at load time will not be preca... danij

2008-11-03

11:19 Revision 913402de (github): Fixed bug in LO_AddLuminousMobjs which could lead to a seg fault when trying to process segs of a degenerate subsector.

Added debug error message if a material for a given sprite frame can not be determined in LO_AddLuminous (trying to c... danij

11:00 Revision 0e00122b (github): Allocate new material texture instances only when needed and store in lists local to materialtex_t.

danij

07:40 Revision 227204c3 (github): Moved definition of materialtexinst_t to r_data.h temporarily to fix FTBFS issues with the new materials system with more standards compliant compilers (than msvc).

danij

07:25 Revision 4f4233ea (github): Materials system 2.0 (Part 6 of 6)

danij

07:24 Revision 6e81632e (github): Materials system 2.0 (Part 5 of 6)

danij

07:24 Revision 467287ca (github): Materials system 2.0 (Part 4 of 6)

danij

07:23 Revision 4961cb5a (github): Materials system 2.0 (Part 3 of 6)

danij

07:23 Revision 3f69e5a3 (github): Materials system 2.0 (Part 2 of 6)

danij

07:22 Revision acba5556 (github): Materials system 2.0 (Part 1 of 6)

Upon getting the initial implementation working well "enough" for a release I began about working on a revised versio... danij

06:48 Revision f71e8c95 (github): Added \fixme notes about the possibility that client/server materialnum_t identifiers might differ in the material change delta interpretation in cl_world.c. Todo: either send the client a material identifier dictionary or send material definitions to the client (which will then construct local materials).

danij

06:42 Revision 9e10f217 (github): Fixed broken interpretation of texture/flat change deltas in PSV_FRAME packets in cl_oldworld.c.

danij

06:40 Revision a6d02404 (github): Changed DD_SetConfigFile and DD_SetDefsFile to accept a filename_t and added bounds checking and limited character copys where needed to prevent potential overflows.

Todo: Remove 256 character limit in file paths! danij

06:36 Revision fb22c46d (github): Cleaned up dd_wad.c somewhat.

danij

06:36 Revision 35c28c62 (github): Removed unreferenced local variable in createDummyWindow.

danij

06:35 Revision e47b1b2f (github): Fixed a minor type mismatch in FR_PrepareGDIFont. Do we even use/need these still?

danij

06:33 Revision 1877c422 (github): Cleaned up GL_UploadTextureContent.

danij

06:32 Revision 1cb49ad9 (github): Removed a test in P_CheckPtcPlanes which would only allow particle generators to be used with flat textures.

danij

06:30 Revision 83c8723f (github): Fixed a potential segfault in P_Iteratep if traversal of the segs of degenerate subsector was attempted.

danij

06:28 Revision 941fddf1 (github): Changed a while(true) infinite loop to for(;;) in Huff_QueueExtract. Couldn't see an easy way to have it self-terminate.

danij

06:12 Revision 5a151e84 (github): Fixed bug in MPE_End where the map error reporting was being (attempted) after the sectors had been hardened (the error reports contain links to editable sectors).

danij

06:09 Revision 805ebf4c (github): Restructured jDoom's P_MobjMoveZ. It *should* be functionally equivalent but now much easier to follow.

Clean up. danij

05:34 Revision c3260073 (github): Added a typedef for the lump grouping tags enum (lumpgrouptag_t).

danij

05:33 Revision bf91debb (github): Added define DOOMSDAY_VERSION_NUMBER for use under win32 as the product number in the metadata of Doomsday.exe.

danij

05:29 Revision 85fc4f8e (github): Removed the old Splint for inconsistent size_t type.

Added a define for the max length of a filename_t. danij

05:24 Revision 113d2a5c (github): Removed unreferenced local variable in R_BeginWorldFrame.

danij

05:23 Revision 220ddaf0 (github): Removed unreferenced local variable in R_CornerNormalPoint.

danij

05:21 Revision 21077300 (github): Fixed incomplete initialization of a plane's surface_t in R_NewPlaneForSector.

danij

05:16 Revision 271c42e4 (github): Fixed a bug in SB_BeginFrame which could result in a segfault if a degenerate subsector is encountered.

danij

05:12 Revision a6e82f7a (github): Adding missing braces from my last commit (doh).

danij

04:58 Revision 376e5834 (github): Fixed an indirection issue in D_NetWorldEvent when dealing with DDWE_HANDSHAKEs.

danij

04:51 Revision 07d4456b (github): Fixed a debug log message in N_SendDataBufferReliably which was trying to print a size_t directly as an unsigned long without a cast.

danij

04:38 Revision 0916c8f2 (github): Fixed a subtle type mismatch in the WIN32 window manager.

danij

04:29 Revision a41a91e8 (github): All games: Removed the CLAMP macro as gcc was reporting it could lead to undefined behaviour. Use MINMAX_OF instead.

danij

02:06 Revision 3d28c33e (github): Fixed upper/lower case mix ups in vcproj file names in the doomsday.sln

danij

02:05 Revision ce69115b (github): Removed old addition include directories from the doomsday.vcproj

danij

01:59 Revision 831b4213 (github): Renamed remotely

danij

2008-11-02

05:14 Revision a9cac42a (github): Fixed missing multiplayer menu items due to another recent change.

danij

2008-10-31

21:22 Bug #556: Hexen Priest green Flechette not effect

Please do no modify the settings of tracker items, you are not a member of the Deng team.
This problem has been fixe... danij

20:41 Bug #556: Hexen Priest green Flechette not effect

This problem has since been fixed for the upcoming 1.9.0-beta6. Closing as duplicate. danij

20:02 Bug #556 (Closed): Hexen Priest green Flechette not effect

I play in hexen with priest character and then green flask is not effect. (purple flask is ok, heal me.)
But the gre... celtixia

2008-10-30

21:48 Revision b3ad4ae7 (github): Fixed surfaces with a "missing" material were being drawn with fakeradio edge shadows.

danij

21:45 Revision b9c2c6a3 (github): Changed warning messages in MPE_GameObjProperty to verbose level1.

danij

21:38 Revision 84651fdc (github): New manifests, resource scripts and updated vcprojs for all games. Proper vendor metadata is now added to the game libraries under Windows.

danij

13:15 Revision 1335dc97 (github): Fixed initialization error in the allocation of rendpolydata_t structures.

danij

13:10 Revision b041c407 (github): Fixed the various P_GetGMO<type> routines so that unknown properties and non-present property values arn't a critical error.

danij

13:03 Revision 856ee796 (github): Fixed Rend_DoesMidTextureFillGap; the return values from Rend_MidMaterialPos were not being used correctly.

danij

12:51 Revision 5ff7e4e6 (github): Made use of the HUD menu background feature to get rid of the old kludges used with the ReadThis/Credits screens. This has the added benefit that the player view(s) won't be drawn while viewing these menus if opaque.

danij

09:02 Revision e14a7229 (github): Fixed jHeretic: The Blaster Missile was not written to saved games. This was due to these mobjs using an alternative thinker (P_BlasterMobjThinker) and the save code has only ever checked for mobjs using P_MobjThinker (all the way back to Heretic.exe). I've gotten rid of P_BlasterMobjThinker and moved the logic into P_MobjThinker which is now used instead.

danij

08:58 Revision 0487d616 (github): Fixed jHexen: The Mage's Wand Missile and the Cleric's Flame Missile were not written to saved games. This was due to these mobjs using an alternative thinker (P_BlasterMobjThinker) and the save code has only ever checked for mobjs using P_MobjThinker (all the way back to Hexen.exe). I've gotten rid of P_BlasterMobjThinker and moved the logic into P_MobjThinker which is now used instead.

danij

08:49 Revision 2a8149b5 (github): Fixed missing initialization for var damageDone in jHexen's P_DamageMobj when the inflictor is a PoisonCloud.

danij

08:43 Revision e4bd5d08 (github): Fixed various issues with save games.

* jHexen - After map load the current weapon would lower and then raise again. (pendingWeapon was never saved and ins... danij

08:27 Revision 4760674f (github): Enhanced PIT_CheckThing to allow non-solid mobjs to move through solid mobjs (the inverse was already possible).

danij

2008-10-29

23:24 Revision aa1478de (github): Fixed jDoom: Gap in the statusbar when icarus.wad is played and statusbar is not opaque. This is due to the ST_ARMS patch being smaller than the IWAD version(s).

danij

23:17 Revision e1d2db59 (github)

danij

21:21 Bug #555: M_DOOM graphic being used on Option menu.

Fixed both issues in svn for 1.9.0-beta6.
To resolve the issue of the main menu items not aligning up, when PWADs cr... danij

17:21 Bug #555 (Closed): M_DOOM graphic being used on Option menu.

M_DOOM graphic being used on Option menu.
I wasn't sure whether to put this as a bug or feature request as it isn'... vermil

2008-10-28

15:09 Revision 3ba28a28 (github): Fixed linking problem with dsOpenAL under Windows due to differing calling convention for Z_Calloc.

danij

04:19 Revision d28a0954 (github): Fixed problem of light sequences not working in jHexen.

danij

2008-10-26

21:41 Bug #554: Line_active xg not checking correct line(s)?

*Attachments:*
* http://sourceforge.net/p/deng/bugs/_discuss/thread/3c2142a5/4e48/attachment/Test.wad
vermil

21:41 Bug #554 (Closed): Line_active xg not checking correct line(s)?

Tested under 1.8.6 and Beta5.x and consulted with Dani who deemed it something worthy of posting here.
Attached is... vermil

14:39 Revision eb0cf66c (github): Fixed potential segfault if R_MarkDependantSurfacesForDecorationUpdate is called for a plane of a sector with no linedefs.

danij

14:34 Revision feb141e8 (github): Fixed problem of not being able to hit the lower section of twosided linedefs if the ceiling in either the front or back sector is sky.

danij

00:17 Revision 9ef057df (github): Fixed: Return the materialnum_t rather than ofTypeID for the skymask material when queried via DD_GetVariable.

danij

2008-10-25

22:37 Revision c514873c (github): Fixed All Games: Due to the different approach taken by Doomsday and DOOM.exe in defining texture/flat animations and the numerous PWADS which exist that exploit DOOM.exe behaviour; it is not possible to use Group definitions to externalize the IWAD animations. All default Group definitions have been removed and instead the default animations are registered by each game during startup (and engine reset) in P_InitPicAnims. This fixes the long-standing problem of PWADS which abused the DOOM.exe behaviour not working correctly when played with jDoom/jHeretic.

Note: This fix depends upon an assumption that will need to be resolved; Materials for textures defined in TEXTURE(1)... danij

00:20 Bug #374: jDoom64: No monsters in E1M33, E1M35

Closing as out of date. Far too many changes have occurred within both the engine and jDoom64 to follow this up. Plea... danij

00:19 Bug #373: jDoom64: Broken behaviour in E1M12

Closing as out of date. Far too many changes have occurred within both the engine and jDoom64 to follow this up. Plea... danij

00:15 Bug #436: JHeretic: Hellstaff fire effect

Fixed in svn for 1.9.0-beta6 danij

00:11 Bug #551: TNT.wad MAP 07

Closing as duplicate. This problem has since been fixed for 1.9.0-beta6 danij

2008-10-24

19:24 Revision fa34c3fe (github): Fixed All Games: A kludge in P_MobjMoveXY prevented missiles from exploding against skyhack ceilings but they would still explode if they crossed a skyhack floor.

Handful of minor optimizations in p_map.c to avoid unnecessary DMU calls.
Clean up. danij

15:59 Revision 45b13d12 (github): Fixed compilation error with jDoom64.

danij

15:55 Revision 8d9e1164 (github): Fixed All Games: PTR_ShootTraverse would check for sky hack ceilings but not sky hack floors when looking to suppress hits.

danij

15:52 Revision bbbf010d (github): Fixed problem in calcSSecReverb which always considered flat #0 as an unknown material.

danij

15:51 Revision cc215191 (github): Fixed fakeradio bug which failed to determine that a side should be considered fully open if the neighbor linedef is twosided with a middle texture that fills the gap between floor and ceiling.

danij

15:48 Revision 235ada09 (github): Fixed bug in Rend_DoesMidTextureFillGap which used the surface material X offset in place of the Y offset.

Fixed problem of wall surfaces using flat #0 always being marked as potentially visible in Rend_MarkSegSectionsPVisible. danij

15:43 Revision b1654ef2 (github): Fixed problem of surfaces textured with flat #0 being misinterpreted as glowing surfaces (and thus fakeradio was disabled and other glitches).

danij

15:41 Revision e268e422 (github): Fixed problem of glowing wall surfaces not working.

danij

15:39 Revision b116dca9 (github): Fixed problem of rockets exploding and bullet puffs being spawned when a sky surface is hit. Caused by Doomsday returning of the material ofTypeID rather than the material num when querying DD_SKYMASKMATERIAL_NUM.

danij

14:00 Revision d5c0a4de (github): Fixed problem of the first flat being given the MATF_NO_DRAW flag. This flag was implemented to support the DOOM.exe bug where the first texture is never drawn. However the same problem is not present with flats.

danij

13:22 Feature #553 (Closed): Performance of many scrolling surface materials

Since the introduction of DMU, many scrolling surface materials on a map are currently causing significant slowdown. ... danij

11:52 Bug #552 (New): Ambient occlusion on middle surfaces of 2-sided lines

How best to handle this; Multitexture, split the poly into a 3x3 quadstrips (at most) and use vertex lighting?
*La... danij

04:53 Revision df9598f8 (github): Fixed jHexen problem of frozen things not dying either by timeout or when hit with a non-icedamage weapon.

danij

04:51 Revision 15f465ec (github): Fixed jHexen: HUD icons like the spinning Wings of Wrath are offset to the right to avoid overlaying the mana icons in the fullscreen HUD. However they were also offset when the view mode was statusbar-overlay.

danij

04:29 Bug #551 (Closed): TNT.wad MAP 07

Missing textures on lower linedefs in a few levels, especially MAP 07 in the first room. Usually around lifts / floo... mugrath

03:10 Revision 61f4f1ca (github): Renamed jHexen's SB_ChangePlayerClass to P_PlayerChangeClass and moved to plugins/common/src/p_player.c

Removed jHeretic's SB_ChangePlayerClass as it is currently unused.
Changed jHexen's version cheat to simply call the ... danij

00:17 Revision adcc9cdb (github): Fixed jHexen: Numerous problems with weapon piece display on the statusbar when semi-translucent.

Fixed jHexen: In deathmatch, the statusbar Kills label patch was not drawn if the statusbar was NOT semi-translucent.... danij

00:09 Revision 33a8e42a (github): Fixed jHeretic: When the statusbar is completely transparent the health marker chain would disappear.

danij