Doomsday Engine

2014-11-12

16:07 Revision d79d3a8e (github): Typos

danij

15:57 Revision 9ef8a1cb (github): Refactor|PlayerImpulse: Separate Impulse, PlayerImpulse and ImpulseAccumulator components

PlayerImpulse is Record-like (todo) description of a logical player
impulse. Each local player has their own set of I... danij

13:44 Revision 06a85dd2 (github): Debug|BindContext|Client: Include the parsed descriptor in new Binding announcements

Composing the Binding descriptor enables one to quickly check whether
the parser interpretation is correct. danij

11:41 Bug #1897 (New): Slightly different sprite clipping in Dday versus vanilla

This is a very small thing, but it is a difference versus Vanilla.
Here is a pwad made to test clipping of sprites... vermil

10:27 Revision f8b06c1c (github): Builder|OS X|Fixed: Detecting platform version number

skyjake

08:01 Revision b6ebb234 (github): Refactor|Binding|Client: Moved Binding (Accessor) conditions into the data Record

danij

05:35 Revision c5757cd9 (github): Documentation|CommandBinding|ImpulseBinding|Client: Cleanup

danij

05:32 Revision f493c6c2 (github): Refactor|Bindings|Client: Moved Binding configuration to [Command/Impulse]Binding

danij

04:57 Revision a65cf533 (github): Refactor|Bindings|Client: Internalized more BindContext functionality

danij

04:18 Revision 5dec006c (github): BindContext|Client: Cleanup

danij

04:12 Revision de01bffe (github): Documentation|CommandBinding|Client: Cleanup

danij

04:06 Revision 711247ac (github): Refactor|Bindings|Client: Moved commandActionFor() to CommandBinding::makeAction()

danij

03:51 Revision 492d750c (github): Refactor|Bindings|Client: Moved new binding components to new source files

danij

02:31 Revision 17b8ebcf (github): Merge branch 'master' of github.com:skyjake/Doomsday-Engine

danij

02:31 Revision 2df6253c (github): Refactor|Bindings|Client: Bindings are now de::Record based

CommandBinding and ImpulseBinding are now de::RecordAccessors based
and provide some handy utility methods.
At this ... danij

2014-11-11

12:59 Revision 571d8b58 (github): Fixed|Client|64-bit: Build failure

skyjake

11:01 Bug #1832: Vsync inconsistently applied in different games

I will remove any remaining game-specific vsync config and only use @de::Config@ for this setting. skyjake

11:00 Bug #1832 (In Progress): Vsync inconsistently applied in different games

skyjake

10:58 Feature #1852: Support for Oculus Rift DK2 (extended desktop mode, LibOVR 0.4.3)

Closed this issue in the scope of 1.15. skyjake

10:58 Feature #1852 (Closed): Support for Oculus Rift DK2 (extended desktop mode, LibOVR 0.4.3)

skyjake

2014-11-10

17:41 Revision f595d143 (github): BindContext: Announce new bindings before removing the overridden ones

danij

17:19 Revision 7254297e (github): Fixed|InputSystem|Client: In symbolic echo mode, event bindings should not be active

danij

17:15 Revision 0499a756 (github): InputSystem|Client: Cleanup

danij

2014-11-09

20:15 Revision 92bcf169 (github): Merge branch 'master' of github.com:skyjake/Doomsday-Engine

danij

20:14 Revision b79c0b73 (github): Fixed|Bindings|InputSystem|Client: ImpulseBinding id generation for multiple local players

Previously, the bindings for all local players of any given impulse
would share the same binding Id.
Todo: The logic... danij

14:58 Revision 344a8468 (github): libcore|String|TokenBuffer: Case sensitive copy-free string equality method

Avoids calling deep-copying QString constructors. skyjake

2014-11-08

21:15 Revision f1a86810 (github): Fixed|BindContext|PlayerImpulse: PlayerImpulse double-click binding activation

Symbolic events have no device identifier, they are referenced by
name and context only. Also fixed a recent refactor... danij

20:27 Revision 389cdc51 (github): Merge branch 'master' of github.com:skyjake/Doomsday-Engine

danij

20:27 Revision 397d9d29 (github): Fixed|PlayerImpulse: PlayerImpulse double-click symbolic event synthesis

Double click impulses are still not working, however. The (bug) hunt
continues... danij

19:40 Bug #1892: Window incorrectly sized when display mode changes during startup (random)

I can definitely reproduce this. It is random in nature, so it most likely relates to the behavior of the Canvas duri... skyjake

19:36 Bug #1879: [Windows] Doomsday randomly fails to start when/after loading shader definitions

Just pushed another fix for a bug that was causing random crashes during startup. Hopefully this was the only one. skyjake

19:36 Bug #1879 (In Progress): [Windows] Doomsday randomly fails to start when/after loading shader definitions

skyjake

19:32 Revision fdba3b8b (github): Merge branch 'master' of https://github.com/skyjake/Doomsday-Engine

skyjake

19:32 Revision 0586c008 (github): libcore|TokenBuffer: Checking for token equality

The sizes must match for an exact match. skyjake

19:30 Revision 95d20512 (github): Fixed|libcore|String: Out-of-bounds memory access (leading to crash)

String::compareWithCase() was calling the QString constructor with a
specific size with the intention of limiting the... skyjake

19:26 Revision 0310fec6 (github): Refactor|libcore|Script: Apply PIMPL in the Script class

skyjake

19:25 Revision 24906588 (github): Fixed|Client: Build failure in release build

skyjake

17:45 Revision d46165fe (github): Refactor|PlayerImpulse: Continued PlayerImpulse remodeling

Evidently double-clicks still aren't working correctly. Hmm... danij

15:04 Revision e992beda (github): Refactor|InputSystem|Client: Moved all binding .cfg file writing to con_config.cpp

danij

02:36 Revision 4bc92be1 (github): Documentation|InputSystem|Client: Basic Binding model theory, plus some cleanup

danij

01:13 Revision f8c92a17 (github): BindContext|InputSystem|Client: Tweaked formatting of bind command/impulse announcements

danij

01:03 Revision 273e3a39 (github): Merge branch 'master' of github.com:skyjake/Doomsday-Engine

danij

01:02 Revision 5eca3dbd (github): Windows|Joystick|DirectInput|InputSystem|Client: Don't log errors regarding nonexistent joystick controls

In this case the errors are due to our basic DirectInput init not
taking the time to enumerate the available controls... danij

00:52 Revision 674a1c94 (github): Merge branch 'master' of https://github.com/skyjake/Doomsday-Engine

skyjake

00:52 Revision 607a18c4 (github): Fixed|Client: Crash during main window construction

It was possible that the log was flushed during construction of the
main window, which lead to the log alerter attemp... skyjake

00:24 Revision ce398127 (github): InputSystem|Client: Cleaned up Input domain log output

danij

2014-11-07

23:42 Revision 433e36e0 (github): InputSystem|Client: Try an event with all BindContexts if none is named

danij

23:37 Revision 5fd578c0 (github): Typos

danij

23:04 Revision 7dc33f45 (github): Typos

danij

22:56 Revision dd107f1c (github): Refactor|InputSystem|Client: Combined BindContext and WidgetActions

Keeping track of who had authority and/or ownership over the relevant
input system components was becoming hard work,... danij

19:29 Revision 202162e0 (github): Debug|InputSystem|Client: Cleanup

danij

19:26 Revision 5ce5f977 (github): Documentation|InputSystem|Client: Todo comment wrt future BindContext decentralization

danij

19:21 Revision d6b2c282 (github): Refactor|InputSystem|Client: Renamed input related source files following organization changes

danij

12:56 Revision c05210a6 (github): Fixed|InputSystem|InputDeviceButtonControl|Client: Expiring InputDeviceButtonControls on context change (typo)

danij

08:33 Revision aca5ba29 (github): InputSystem|BindContext|Client: Revert statecondition_t equality test (refactoring oversight)

danij

08:17 Revision 408ee13f (github): InputSystem|BindContext|Client: Clean up BindContext acquired device usage

Evidently the "stuck control" problem occurs when the binding context
for bound control is changes and then back agai... danij

08:13 Revision 0f6cd491 (github): InputSystem|Client: Maintain id and name maps for PlayerImpulse lookup

danij

05:11 Revision 00bd5a33 (github): Console|InputSystem|Client: Applied rich formatting to input system command output

danij

02:31 Revision 8a4f15fd (github): InputSystem|Bindings|Client: Minor cleanup

danij

01:09 Revision 6615a548 (github): ClientApp: WidgetActions can now be destroyed whenever

danij

01:07 Revision 728d8c51 (github): Merge branch 'master' of github.com:skyjake/Doomsday-Engine

danij

01:07 Revision 280bcf17 (github): Refactor|InputSystem: Completed binding system reimplementation in OO C++

Todo:
- Fix the BindContext management regression.
- Cleanup. danij

2014-11-06

21:56 Revision 812e4c43 (github): UI|Client|Fixed: Layout of the close button of Renderer Appearance sidebar

skyjake

21:52 Revision 24c4d61e (github): Oculus Rift|libappfw: Providing a projection when HMD is not connected

de::OculusRift should not be asked to provide a projection matrix
if no HMD is connected. skyjake

21:10 Revision e701599d (github): UI|Client: Added an icon for a "Close" action

Currently used in the Renderer Appearance editor.
Todo: Evaluate use of this elsewhere in the UI and tweak appearance. skyjake

20:51 Revision b949a247 (github): Oculus Rift|libappfw: Updated for LibOVR 0.4.3

Version 0.4.3 has some API changes, so updated the use of LibOVR
accordingly. Still no luck with direct mode renderin... skyjake

20:41 Revision f4c5c4e9 (github): Fixed|Release: Compiler warning (unused argument)

skyjake

20:40 Revision 9b8f7ed6 (github): Fixed|Client: Crash at shutdown

~WidgetActions accesses the singleton ClientApp's InputSystem, so
WidgetActions must be deleted first. skyjake

15:04 Bug #1896: Validating Group def tic durations (value -1 is mishandled/unsupported)

Something I'd like to do in a future version is redesign how the sky-glow is sampled. Presently this works by evaluat... danij

14:00 Bug #1896: Validating Group def tic durations (value -1 is mishandled/unsupported)

Apologies for arguably reporting several things in one report (I got carried away writing). vermil

13:59 Bug #1896 (Closed): Validating Group def tic durations (value -1 is mishandled/unsupported)

With state def's, one can set a length of -1 to make the mobj never leave that state, but -1 doesn't appear to work w... vermil

00:38 Bug #1895 (Closed): [Windows] Pressing Alt+F4 during engine startup causes Dday to crash

Pretty much what the subject states.
If one presses Alt+F4 during gameplay, the game's quit dialog comes up (I ass... vermil

2014-11-05

16:06 Revision dc7d51a0 (github): Refactor|InputSystem|Bindings|Client: Renamed source files for CommandBinding, ImpulseBinding

danij

13:46 Bug #1879: [Windows] Doomsday randomly fails to start when/after loading shader definitions

Touch wood, I've also not yet had any issues with Dday failing to start anymore since upgrading to 1402 from 1399 (fr... vermil

13:24 Bug #1879: [Windows] Doomsday randomly fails to start when/after loading shader definitions

skyjake wrote:
> vermil wrote:
> > I don't know if it matters, but since 1.15 I've noticed Dday will hang for a cou... vermil

11:39 Bug #1879: [Windows] Doomsday randomly fails to start when/after loading shader definitions

vermil wrote:
> I don't know if it matters, but since 1.15 I've noticed Dday will hang for a couple of seconds when ... skyjake

11:20 Bug #1879: [Windows] Doomsday randomly fails to start when/after loading shader definitions

It is still random yes, though a lot more common with 1.15 I believe.
Upon pressing play, I get a loading cursor a... vermil

11:02 Bug #1879: [Windows] Doomsday randomly fails to start when/after loading shader definitions

vermil wrote:
> Dday will very often not start through Snowberry
Do you mean the behaviour is still random, so th... skyjake

10:38 Bug #1879: [Windows] Doomsday randomly fails to start when/after loading shader definitions

I'm afraid that with build 1399 I'm personally still finding Dday will very often not start through Snowberry or even... vermil

10:23 Bug #1894: [Heretic] Map that relies on original footclipping

danij wrote:
> Thanks for the clarification. That explains a lot (why it works in vanilla, is fixed in dday when not... vermil

00:08 Bug #1894: [Heretic] Map that relies on original footclipping

Thanks for the clarification. That explains a lot (why it works in vanilla, is fixed in dday when not loading a save,... danij

02:27 Revision 33f20e81 (github): InputSystem|Client: Further cleanup

danij

01:41 Revision 501b1fdc (github): InputSystem|Client: Cleanup

danij

01:27 Revision 73c84538 (github): Refactor|InputSystem|Bindings: Renamed *Binding id members for clarity

danij

01:07 Revision e2aa18a5 (github): Refactor|InputSystem|Bindings: Disambiguate binding data struct names

Refer to logical player controls as "impulses", to avoid confusion
input device controls. danij

2014-11-04

23:56 Bug #1894: [Heretic] Map that relies on original footclipping

danij wrote:
> I'm not sure what you mean by "vanilla bug that upon loading a save game, all mobjs are considered in... vermil

13:03 Bug #1894: [Heretic] Map that relies on original footclipping

I'm not sure what you mean by "vanilla bug that upon loading a save game, all mobjs are considered in the sector thei... danij

12:10 Bug #1894: [Heretic] Map that relies on original footclipping

GZDoom also appears to have the same issue with the blue key on this map. Dday 1.8.6 can't actually load this map for... vermil

11:07 Bug #1894: [Heretic] Map that relies on original footclipping

Foot-clipping should not actually matter here, as its a visual effect only. Needs investigating further. danij

00:29 Bug #1894 (New): [Heretic] Map that relies on original footclipping

E1M2 of this Heretic pwad relies on the original games buggy foot clipping in order for the player to be able to pick... vermil

21:12 Revision dd3aa4cf (github): InputSystem: Internalized updateAllDeviceStateAssociations()

Observe BindContext changes to determine when to perform the update. danij

18:59 Revision 7cc0f59e (github): InputSystem|Bindings|Client: BindContext (de)activation is now observable

danij

18:15 Revision d552ba68 (github): InputSystem|Bindings|Client: Cleanup

danij

17:44 Revision 2d7c2183 (github): Refactor|InputSystem|Bindings|Client: Continued splitting up b_main.h/cpp

danij

16:03 Revision 5b79ff5e (github): Refactor|InputSystem|Bindings|Client: Renamed b_context.h as bindcontext.h

danij

15:34 Revision 687e432c (github): Refactor|InputSystem|Bindings|Client: InputSystem has ownership of BindContexts

Todo: Cleanup danij

14:08 Revision cdcd0149 (github): Cleanup|InputSystem|Bindings: Clarify ambiguous/misleading names/terms

danij

12:50 Revision ea99c83a (github): Client|InputSystem|Bindings: Cleanup

danij

00:17 Bug #1893 (New): Automap drawing difference to original games

Doomsday will only draw lines behind a mid texture, if there is a see through element on the graphic being used as a ... vermil

2014-11-03

22:12 Bug #1888: doomsday 1393 segfault ubuntu 14.10 radeon drivers

still a crash with build 1402... eunbolt

14:22 Bug #1888: doomsday 1393 segfault ubuntu 14.10 radeon drivers

Is this reproducible with newer builds? 1402 in particular fixes many memory handling issues. skyjake

14:26 Bug #1854 (Closed): libsdl2 14.04

skyjake

14:26 Bug #1854: libsdl2 14.04

Prebuilt packages are now for Ubuntu 14.04, so closing this issue. skyjake

14:19 Feature #1891: Notification area animations

Marking 10% progress because the notification area ownership model has been revised to work via relays. This gives th... skyjake

12:55 Revision 27c4a598 (github): Refactor|Input|Bindings|Client: Began remodeling bindings into OO C++

danij

2014-11-02

22:25 Revision 3ec56280 (github): Debug|Input: Fixed labels for input device control state visuals

danij

19:21 Revision 2027a13e (github): Performance|Oculus Rift: Detecting connected HMDs

Use the correct LibOVR API function for detecting connected HMDs
rather than trying to blindly create an ovrHmd. At l... skyjake

19:20 Revision ba04c59b (github): Performance|libcore: Avoid unnecessary mutex locking

theLogs() does not need locking after Logs exists. skyjake

17:33 Revision 5c986c2a (github): Cleanup|Input|Bindings: InputDevice IDs are signed ints

skyjake

15:00 Revision 67e5955c (github): Cleanup|UI|Bindings: InputDevice IDs are signed ints

danij

14:35 Revision 41daa95d (github): Merge branch 'master' of github.com:skyjake/Doomsday-Engine

danij

14:35 Revision 101a949a (github): InputSystem|Client: Cleanup

danij

14:03 Revision 388b79cb (github): Cleanup

skyjake

14:02 Revision 9f78e710 (github): Cleanup

skyjake

12:24 Revision 618ec44c (github): Fixed|MSVC: Compilation issue

skyjake

12:04 Revision 0b3276ce (github): Fixed|libcommon|Client|64-bit: Crash when binding a control

The bit shifts for forming a 64-bit pointer were constrained to 32-bit
values, producing an invalid value.
Also fixe... skyjake

10:59 Revision a92b2dd8 (github): Merge branch 'master' of ssh://github.com/skyjake/Doomsday-Engine

skyjake

10:59 Revision 82a0a536 (github): Refactor|UI|Client: Ownership of notification widgets

Now using the UniqueWidgetPtr template. skyjake

10:46 Revision edcaf6e6 (github): libappfw: Revised notification area, added RelayWidget

The notification area's ownership model was changed so that it does not
own the notification widgets, just relays the... skyjake

10:38 Revision 6ae3357b (github): libappfw|GuiWidget: Destroying a GUI widget at a later time

Particularly when handling audience notifications about object deletion,
it is necessary to schedule object deletion ... skyjake

10:37 Revision fecc36c4 (github): Fixed|libappfw|GuiWidgetPrivate: Observe root atlas for deletion

Fixes the problem where the atlas has already been deleted when the
widget is being deleted. skyjake

10:36 Revision 56683229 (github): Fixed|libappfw|ChildWidgetOrganizer: Potential memory access issue

Issue found by Guard Malloc. Since we have an iterator already, it
is better to use it to remove the map entry rather... skyjake

10:31 Revision 557ff2a7 (github): libcore|Unix|Debug: Macro to generate backtrace as a string

skyjake

10:29 Revision e066eb53 (github): Cleanup

skyjake

10:28 Revision 209a3cbf (github): libcore|Log: Multithreading robustness

Theoretically, a race condition might occur in theLogs(), if it were
called from multiple threads for the first time. skyjake

04:02 Revision a4132cff (github): InputSystem|Client: Re-initialize InputDevices during a game change

Todo: Could observe App::GameChange? danij

03:47 Revision 3cc8faed (github): Refactor|InputSystem|Client: Moved InputDevices and ddevent_t queue into InputSystem

danij

01:37 Revision 1526ba0a (github): Refactor|InputSystem|Client: Moved InputDevice components to new source files

danij

00:30 Revision 49e868b5 (github): InputSystem|Client: InputDevice (de)activation is now observable

danij

2014-11-01

23:28 Revision 5780aabd (github): InputSystem|Client: InputDevice initialization cleanup

danij

20:41 Revision 03873742 (github): Refactor|InputSystem|Client: Hide the meaning of InputDevice(Control) IDs at API level

Users of the InputDevices should not assume anything from the IDs
assigned to devices and/or their controls. danij

13:08 Revision 2af910da (github): libappfw: Added a smart pointer for owning a GuiWidget

Automatically calls the correct destructor on the widget. skyjake

13:06 Revision c91f70dc (github): libcore|Widget: Setting a widget's root manually

Normally the root widget is found by tracing all the parents. skyjake

2014-10-31

16:47 Revision 7ecd1053 (github): Refactor|InputSystem|Client: Reimplemented InputDevice models in OO C++

danij

11:08 Revision fbeb0abb (github): UI|Client|Console: Restored the [>] button appearance

skyjake

11:05 Revision 333dc49c (github): Fixed|Client: Memory leak (notification widgets not deleted)

The Alert and Updater notifications were not deleted if they were
hidden when the app was shut down.
Todo: A better ... skyjake

11:04 Revision 2e8f3346 (github): Debug|Client: Assert that all reference-counted objects get deleted

skyjake

11:03 Revision df4b2168 (github): Fixed|libappfw|FoldPanelWidget: Memory leak

The content was not deleted if the fold was closed when the widget was
deleted. skyjake

11:01 Revision 0e216a3f (github): Cleanup

skyjake

08:50 Revision cbfb157a (github): Fixed|libcore|Scripting: Memory leak in expression evaluator

The scopes for the evaluated results were not always deleted. skyjake

2014-10-30

21:34 Revision 8f563889 (github): UI|Client|AlertDialog: Fixed layout problem when dialog is narrow

Now space is reserved for the auto-hide setting. skyjake

21:32 Revision 1d32f4e8 (github): libappfw|DialogWidget: Specifying a minimum width for a dialog

skyjake

21:32 Revision 6554d2fc (github): libcore: Cleanup

skyjake

21:31 Revision ad88641a (github): Fixed|libcore|ArrayExpression: Memory leak

The array expression's arguments were not deleted when the expression
was deleted. skyjake

21:30 Revision ec8017ed (github): Fixed|libgui|Font: Memory leak (release font metrics rules)

skyjake

21:29 Revision 82832a80 (github): Fixed|libappfw: Memory leak in LogWidget (at shutdown)

The cached entries were not deleted when the widget was deleted. skyjake

16:16 Revision 23396b2b (github): Merge branch 'master' into common-episode-definition

Conflicts:
doomsday/plugins/common/include/hu_lib.h
doomsday/plugins/common/include/m_ctrl.h
doomsday/plugins/common/... danij

11:06 Revision 1bc12137 (github): Default Style|Client: Tweaked subheading style

The subheading style on Windows and Unix was changed to use a small,
bold, normal case font. On OS X, they remain lig... skyjake

11:05 Revision 54713b8d (github): Fixed|Client: Missing header

skyjake

01:56 Revision 6ed86819 (github): Menu|InputBindingWidget|libcommon: Cleanup

danij