Project

General

Profile

Activity

From 2009-06-07 to 2009-06-20

2009-06-19

13:16 Revision 0648a713 (github): Various fixes due to the change in behaviour between dd_snprintf and snprintf on the Windows platform.
danij
13:15 Revision 3a9e75e7 (github): Fixed: "Paused game while opened menu bug" (see here http://sourceforge.net/tracker/index.php?func=detail&aid=2808684&group_id=74815&atid=542099).
danij
12:54 Revision c003277c (github): Various changes working towards a common, game-side mobj spawn supersystem.
* Big clean up of player spawn routines in p_start.c (fewer logic paths, more code reuse and more flexible API(s)).
*... danij
12:22 Revision fc9246e0 (github): Changed my mind about how to fix the ripper spawned blood momentum issue in jHeretic/jHeretic. Implemented a kludge in PIT_CheckThing to always treat things of type MT_BLOOD as solid. Removed the mf_solid flag from thing BLOOD definitions.
I'm not particularly happy with either solution but for now, the kludge appears preferable. danij
11:37 Revision 6da975d5 (github): Various fixes due to the change in behaviour between dd_snprintf and snprintf on the Windows platform.
danij

2009-06-18

21:54 Bug #721 (Closed): Paused game while opened menu bug
If the player is standing on a floor with a height of -40 and you press Escape to get into the menu the game won't pa... xxmiltenxx
21:54 Bug #721: Paused game while opened menu bug
Doomsday out of Heretic and Doom
*Attachments:*
* http://sourceforge.net/p/deng/bugs/_discuss/thread/c4e1a73e/3e23/a... xxmiltenxx
08:45 Bug #720: HEXEN 32BIT Savegames not stored in linux
did'nt read your first comment.
This is definitely a 32bit build, with the
described problem. cteg
06:42 Bug #720: HEXEN 32BIT Savegames not stored in linux
dont know who and why someone changed the "hexen" tag to "64bit"?
its 32bit. cteg
03:57 Revision 2034f67a (github): Fixed translation tables not loading correctly (dd_snprintf behaviour).
danij

2009-06-17

21:44 Revision af1c8b76 (github): Various changes working towards a common, game-side mobj spawn supersystem.
* Commonized P_SpawnMapThing, moved to p_start.c and renamed spawnMapThing.
* Relocated translation logic applied to ... danij
19:21 Revision d94d6b5c (github): Optimize memory management in the deferred mobj spawn system. Allocate spawn queue nodes in batches of 32. Maintain a list of unused nodes to be recycled later.
danij
18:47 Revision d174d450 (github): Don't use mapspot_t as the representation of mobj-to-be-spawned in spawnqueuenode_t as this creates an unnecessary dependency.
danij
18:39 Revision 5de5081e (github): Added a void* context paramater to the deferred spawn callback.
danij
18:19 Revision 21ed3fe5 (github): Moved spawning of the item respawn fog and the accompanying sound out of the deferred mobj spawn system and into a caller-defined callback (thus making a P_DeferSpawnMobj* identical to to a P_SpawnMobj* other than the fact it may occur at some later point in time).
danij
14:49 Revision 95181052 (github): Fixed a couple of compilation warnings.
danij
14:37 Revision 16ee248a (github): Fixed paramater order in call in M_TranslatePath in G_LoadGame (VS8's "intellisense" db was out of date...).
Fixed save game log message. danij
14:17 Revision 708d1b9f (github): Implemented a new subsystem for deferred mobj spawns. Currently only used by jDoom to replace the item (re)spawn queue in deathmatch mode 2.
Changed (jDoom): Removed fixed limit ITEMQUESIZE (previous limit was 128).
Also added some more verbose status messa... danij
12:14 Bug #720: HEXEN 32BIT Savegames not stored in linux
This problem is looking likely to be limited to 64bit builds (not currently supported). In order to progress with thi... danij
07:19 Bug #720 (Closed): HEXEN 32BIT Savegames not stored in linux
seems like the game is "doing" something but no savegames are stored.
and the offset of the selector arrow is wrong.... cteg

2009-06-16

08:31 Revision 52f64455 (github): Updated jDoom64 in line with the other plugins (re changes since 1.9.0-beta6.2).
danij
08:10 Revision d80edbdc (github): Fixed jDoom: Mobjs of type MT_PUFF and MT_ROCKETPUFF spawned with too much z momentum.
danij
07:52 Revision 426eb003 (github): Fixed jDoom: At Nightmare skill, enemies respawned where they died (rather than where they were first spawned).
danij
05:57 Revision e39d97e0 (github): Fixed jHeretic/jHexen: "Heretic/HeXen: MF_Ripper Blood".
Changed jHeretic/jHexen: Gave MT_BLOOD the MF_SOLID flag. The original Heretic/Hexen collision detection logic did no... danij
05:41 Revision eb7cb3b9 (github): Removed code introduced for debug purposes that was included accidentally in the previous commit.
danij
05:34 Revision 8e8a7e20 (github): Commonized P_RipperBlood (moved to p_actor.c).
danij
03:57 Revision 4a1584a3 (github): Added missing include for stdarg.h to dd_share.h
danij
02:00 Revision c266de1e (github): Fixed jHeretic: "Heretic Maulotaur attack behaviour incorrect" (see here http://sourceforge.net/tracker/?func=detail&aid=2759673&group_id=74815&atid=542099).
danij
00:57 Revision 7bdb14d0 (github): Fixed jHeretic/jHexen: Minotaur floor fire moving too fast.
danij
00:23 Bug #693: Keybord and mouse controlls delayed
Newer versions of Doomsday simply "play nicer" with the underlying operating system and by that I mean; Doomsday will... danij

2009-06-14

22:45 Revision e2be2056 (github): Further fix to the player message log (messages truncated one character short).
danij
22:17 Revision 23ecab92 (github): Fixed jDoom: "A_SkelMissile not taking vertical offset into account" (see here http://sourceforge.net/tracker/?func=detail&aid=2796729&group_id=74815&atid=542099).
danij
22:02 Revision ee3e0d8f (github): Fixed: "Position of particles spawned using the spherical method" (see here http://sourceforge.net/tracker/?func=detail&aid=2480726&group_id=74815&atid=542099).
Fixed: Inaccuracy in particle bounce vector calculation. danij
16:42 Revision b27f6279 (github): Fixed: "on screen text error" (see here http://sourceforge.net/tracker/?func=detail&aid=2794337&group_id=74815&atid=542099).
danij
16:13 Revision 0254b8f2 (github): Increment highest ordinal in doomsday.def
danij
16:11 Revision 9b9561da (github): Implemented dd_snprintf to standardize the differences between the MSVC and C99 implementations. The output buffer is always terminated. Return value is as per the MSVC specification: Number of characters written to the output buffer if lower than or equal to the size of the output buffer (minus the terminating NULL), else -1.
Replaced all instances of snprintf with dd_snprintf.
Exported both dd_vsnprintf and dd_snprintf via the public API.
A... danij
15:30 Revision 1e92cf8c (github): Implemented dd_vsnprintf to standardize the differences between the MSVC and C99 implementations. The output buffer is always terminated. Return value is as per the MSVC specification: Number of characters written to the output buffer if lower than or equal to the size of the output buffer, else -1.
Replaced all instances of vsnprintf with dd_vsnprintf. danij
14:25 Revision 5d5344a2 (github): Use an appropriate release name for the trunk.
skyjake
14:17 Revision 63b8483e (github): Allow configuring CPU archs with CMake variables.
skyjake
14:05 Revision e961d6b2 (github): Replacing the current trunk with the 1.9.0-beta6 branch.
skyjake
13:43 Bug #691: on screen text error
Fixed in svn for 1.9.0-beta6.4 danij
13:04 Bug #691: on screen text error
> I have one comment about your commit... ...Since vsnprintf() returns the number of characters without the null, if ... danij
13:02 Bug #691: on screen text error
Oh, and you should be able to see the HUD log message last character truncation problem now on Windows as well. skyjake
13:00 Bug #691: on screen text error
An exported public implementation of dd_vsprintf() sounds fine.
I have one comment about your commit:
#ifdef WI... skyjake
12:51 Bug #691: on screen text error
There is one final decision to make regarding these; do we want an engine-side, exported, public implementation of dd... danij
12:21 Bug #691: on screen text error
Ok, will do. danij
12:16 Bug #691: on screen text error
I think we should go the dd_vsnprintf() route because otherwise the return value thing is a bit ugly and error prone ... skyjake
12:14 Bug #691: on screen text error
I think that is going to be our best option. Either that or sort out our headers... danij
12:12 Bug #691: on screen text error
Well, one option is to make it something like dd_vsnprintf() and use find/replace on all the instances of where it's ... skyjake
12:08 Bug #691: on screen text error
That is what I've been trying to do but I'm hitting a brick wall because of the way our headers are setup (there are ... danij
12:04 Bug #691: on screen text error
How about that earlier suggestion about naming it win32_vsnprintf() and using a #define from vsnprintf to that? skyjake
11:51 Bug #691: on screen text error
That sounds OK to me though we will naturally have to ensure we don't use the return value or where we do, we account... danij
11:13 Bug #691: on screen text error
On the bright side, I don't think the return value is used anywhere in the engine. It is vastly more important that t... skyjake
11:05 Bug #691: on screen text error
Clearly I need another coffee. On re-reading the definitions again, yes I agree that the Windows varient is indeed be... danij
10:51 Bug #691: on screen text error
I disagree. The difference is that the Windows version will return -1 if truncation occurs. The C99 returns -1 only i... skyjake
10:46 Bug #691: on screen text error
Here is the definition for vsnprintf from MSDN:
"vsnprintf, _vsnprintf, and _vsnwprintf return the number of chara... danij
10:42 Bug #691: on screen text error
Here's the definition from the C99 (draft) for vsnprintf(): [n is the size of the output buffer]
"The vsnprintf fu... skyjake
10:34 Bug #691: on screen text error
Well, from that man page I interpret that the C99 standard defines the behavior like that: the return value is the nu... skyjake
10:19 Bug #691: on screen text error
Looking at the function header comments you've added for snprintf and vsnprintf in dd_share.h I would like to clarify... danij
02:23 Revision 1e2e79b0 (github): Fixed jDoom: Graphic patch for five character '5' drawn instead of that for a six '6' when using FONTB.
danij

2009-06-13

21:37 Bug #691: on screen text error
Yeah I can finish this (I expect I'll be commiting tomorrow as I don't feel much like coding right now). danij
10:18 Bug #691: on screen text error
I committed the functions in rev 6668. Danij, I trust you can finish the implementation on Windows?
However, if vs... skyjake
08:55 Bug #691: on screen text error
Agreed. That sounds like the best course of action to me. danij
08:39 Bug #691: on screen text error
The problem seems to be that the (nonstandard) snprintf() function is defined a bit differently in Windows and Unix. ... skyjake
13:15 Revision e780e38a (github): Wrappers for (v)snprintf() on Windows.
Since I don't have a win32 build environment set up, I can't really implement this fully. Currently vsnprintf() just ... skyjake
13:11 Revision cb6403ae (github): Encoding issues.
skyjake
12:02 Revision 7a0633ca (github): Cleaned up s_cache.c
danij
08:16 Revision 256eb5a5 (github): Fixed a couple of comments.
danij
08:10 Revision 889d3e44 (github): Various interwoven fixes:
* Fixed several instances of lump cache abuse (writing back to cached lump memory).
* Fixed 64bit portability issues ... danij
07:36 Revision 107a7276 (github): Fixed: Particle generators not being spawned even if the maximum number of active generators had not been reached.
danij

2009-06-12

07:47 Revision e41e76a2 (github): Further clean up of r_extres.c Also, don't prepend a redundant delimiter at the end of the class data path in initClassDataPaths.
danij
06:39 Revision 97104bcb (github): Allow using the resource locator to finger files in the base path without returning their name.
danij
06:31 Revision ffe7e2d8 (github): Reworked tryLocateResource to represent paths with ddstring rather than filename_t.
danij
06:03 Revision e5e9f100 (github): Cleaned up redundant, garbage code in R_PrependDataPath.
danij
05:54 Revision 44ecd379 (github): Optimize: Pruned redundant code in R_RegisterSkin; we now use the resource locator for searching a model's local directory and will also handle trying each supported graphic file format in said location.
danij
04:31 Revision f75854e9 (github): Fixed: SIGSEGV when attempting to draw surface reflections if the textures for which had not been located.
danij
03:38 Revision 062300af (github): Further improved M_FindFileExtension to not be fooled by paths with trailing directory seperators or file names that end with a period '.' character.
danij
02:45 Revision 8b3d1cf9 (github): Various changes to the resource locator:
* Fixed: Resource locator unable to find the shinemaps and associated masks when using the mod "jDRP Reflections v1.2... danij

2009-06-11

06:09 Bug #719 (Closed): Windowed resolutions wider than primary monitor do not work
If Doomsday is launched in windowed mode with a resolution larger than that of the primary monitor, the window shrink... leapo

2009-06-10

12:04 Bug #718 (Closed): Heretic: Chicken auto movement
When the player is turned into a chicken, Dday seems to lack the random movement that occurs when you are stationary.... vermil
10:14 Bug #717 (Closed): Heretic/HeXen: MF_Ripper Blood
In the original Heretic and HeXen (i.e powered the up Dragon Claw or Mages wand) individual blood splats caused by mi... vermil
09:23 Bug #491: Tiny blood splat editing bug
I'll also throw in that the Doom players punch and chainsaw actions are similarly affected by this behaviour, except ... vermil
00:11 Bug #674: Segmentation violation: semi-transparent wall texture (bars)
I'm sorry to say this problem persists in version 1.9.0-beta6.3 Jun 8 2009 (DGL). mittens
00:10 Bug #674: Segmentation violation: semi-transparent wall texture (bars)
Output from version 1.9.0-beta6.3 Jun 8 2009 (DGL).
*Attachments:*
* http://sourceforge.net/p/deng/bugs/_discuss/th... mittens

2009-06-09

09:14 Revision 0fa4141e (github): Continued work on cleaning up the texture manager.
Redesigned management of color palettes for use with textures:
* DGL is no longer dependant upon a color palette (onl... danij

2009-06-08

15:31 Bug #716: Compatible Dehacked file not loading
The problem here is the completely bizarre way in which DeHackED choose to represent Text in its patch format. Rather... danij
14:37 Bug #716 (Closed): Compatible Dehacked file not loading
I was doing some more testing of Doomsday beta 6.3, and decided to load Alien Vendetta, along with its dehacked file ... sonicdoommario
03:21 Revision 7ccccf2c (github): Fixed: Model loading issues.
danij
03:13 Revision 6e9388fa (github): Fixed serious issue in FileHash_Create which could potentially result in heap corruption.
danij
02:34 Revision 70c87f13 (github): Ensure paths passed to R_AddClassDataPath are formed correctly (don't add empty paths or paths containing nothing but a directory separator character to the path list).
danij
01:34 Feature #1496 (Closed): Aspect correct scaling of UI and HUD elements/interfaces
Game HUDs stretch to the same aspect ratio as the current resolution, resulting in horizontally stretched HUD graphic... leapo
01:31 Feature #1495 (Closed): FOV cannot be set higher than 179
The FOV in Doomsday cannot be set higher than 179, making a proper FOV impossible to obtain on very wide aspect ratio... leapo

2009-06-07

22:52 Revision 1fad2991 (github): Copy skyjake's fix for the missing UI textures issue to the WinAPI window manager.
danij
20:14 Revision 55760c90 (github): Explicitly unload and reload UI textures when changing the video mode. Seems to fix the problem of the disappearing UI textures (OS X, jHeretic).
skyjake
19:36 Revision 8eaa5506 (github): OpenGL must be accessed only in the main thread. A crash was caused by OpenGL textures being deleted in the busy worker thread.
skyjake
14:09 Revision 0dd1cc78 (github): Fixed model path registration. It's not a particularly good idea to shuffle around function parameters, especially when they end up in the wrong order...
Restored original order of parameters. skyjake
14:07 Revision b38351e6 (github): Slight visual adjustment for busy indicator. The dashed ring is now in the middle of the two other rings.
skyjake
14:07 Revision 23166dd9 (github): Fixed Unix build.
There was a couple of strncpy() and other calls missing the len argument. Revised Dir_ExpandHome() to use ddstring. skyjake
 

Also available in: Atom