Event-based gamepad functionality#1106
Conversation
src/src.pro
Outdated
| macx:QMAKE_MACOSX_DEPLOYMENT_TARGET = 10.7 | ||
|
|
||
| QT += network opengl uitools multimedia gui | ||
| QT += network opengl uitools multimedia gui gamepad |
There was a problem hiding this comment.
Include this conditionally like before - I'll build all installers with Qt 5.9 but we have min 5.6 ATM for Linux distros.
Should be safe to compile on versions of Qt < 5.8
Probably another instance of Windows security which does not let user level code have access to USB HID devices that are identifiable as keyboards or mice (some of the latter can send the equivalent of keystrokes under some circumstances) to try to prevent the installation of key-logging software... 😈 Also, can you pin down when |
|
On an initial glance, I see a post on a Qt forum from 4 years ago that mentions qtHaveModule as new at that time. |
|
Apparently there are slot related problems that don't compile in one of the tests Travis is doing. I see a fix was applied later on. I'll see if that fixes the issue. |
|
I think if you change it back and put a |
| macx:QMAKE_MACOSX_DEPLOYMENT_TARGET = 10.7 | ||
|
|
||
| QT += network opengl uitools multimedia gui | ||
| qtHaveModule(gamepad): QT += gamepad |
|
I do happen to own a Logitech F310 which supports Direct input and XInput modes. Currently waiting for the Qt 5.9 stuff to finish installing and then I'll be able to test with it. :) |
There was a problem hiding this comment.
This works out nicely, all the expected buttons and axis work. I only have an issue with connection and disconnection as they don't seem to be raised at all for my setup. Qt's debug console prints Gamepad: Could not read from input device (No such device) when I unplug my controller and unless it is plugged in when Mudlet starts the controller isn't noticed or connected later.
I half suspect either the age of my controller or perhaps driver configuration is mishandled by Qt. This is all in XInput mode, just to clarify. I'm not worried about Direct Input since it has been deprecated for some time.
I've had similar issues with controllers in Unity so this isn't totally unexpected behavior, imo.
[Edit]
Did a bit of digging on this and the issue I have is already reported in the bug tracker here:
https://bugreports.qt.io/browse/QTBUG-57078
|
For reference, here's where I got the impression Windows support isn't going to really improve for Gamepad over time: https://bugreports.qt.io/browse/QTBUG-57389?focusedCommentId=346127&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-346127 I'd really prefer access to DInput as well for a number of reasons, especially since a lot of accessibility devices will work with DInput but not XInput. Seems it's not meant to be though. |
|
I did test the Direct input mode and the current code does seem to register two axis (one joystick) and one of the buttons on my specific controller. With that I gather some extremely basic joystick input devices can work. |
|
The only part that's really hard after getting that going would be getting analog sticks handled properly. I've not dealt with the sort of math required to implement a proper deadzone in a long time, but I might be able to figure it out over time. The fact that it can't be made to work with a DS4 or similar controller just connected via bluetooth without a third party program like DS4Windows running to emulate an XInput device is what made it feel so incomplete to me. |
|
I've come to expect the need for special drivers or applications to get most third-party stuff to work nice on windows. At least there are applications like http://www.x360ce.com/ that could help with with emulating XInput for older DirectInput controllers on windows. As for the deadzone math, that shouldn't be too hard. From my experience in Unity, input axes are usually able to return values from -1.0 to 1.0, where 0.0 is "centered" but since joysticks may not always come to rest exactly at (0.0, 0.0) after they are moved Unity recommends a deadzone between about -0.2 and 0.2. I think for now the deadzone math can be carried out in the lua scripts, so as to keep the most raw information available to Lua while we learn about the quirks of QGamepad and other conditions that might effect this cool new feature. :) |
|
I'll merge this PR in as it's been sitting ready for over a week now and has been reviewed by most people on the team. If anything is wrong with this, we can always fix it up in another PR! |
|
@dicene this needs some pretty good documentation now else it'll be inaccessible save for the hardcore scripters - could you document all of the events + make a sample demo package to demonstrate how to use this? Something like I did to demo coroutines or one-line installs. |
* Removed duplicate clear event handlers call * Revert back to dev version. 3.3 here we go! * Adds saveWindowLayout() call to AutoSaveOnExit processing. * Edbee editor implementation (#985) * Added a frame to edbee so it looks nicer * Got utf8 module to load (#1074) (update packaging!) * Modified tooltip for Save and Save Profile to clearly indicate their keyboard shortcuts * Added: shortcut for delete action in editor (#1084) * shortcut added for delete action * Removed redundant edbee config related code (#1091) * Delete TScriptEditorManager.h * Delete TScriptEditorManager.cpp * BugFix: import some fixes by "Zolder" from Mudlet (Realms of Legends) fork import: some Zoilder fixes from the MudletRL fork: * in TBuffer::translateToPlainText(...): make some MXP related code execute only if Host::mFORCE_MXP_NEGOTIATION_OFF is NOT set as well as another condition. * in TBuffer::wrap(...): change a for(...) {...} loop termination condition be a "<" rather than a "<=" test - which sounds like an "off-by-one" error fix-up. * In TTextEdit:updateScreenView(): revise code that seems to prevent the width of a "main" console being set too small and NOT written to the "parent" Host::mScreenWidth variable - the original modification removed such a restriction completely, but I have just lowered the minimum from 100 to 40 and ensured it is always stored both in the local "main" console and to the parent record in the Host class for that console. * In TTextEdit::drawForeground(...): revise the behaviour *I think* when at the point where a console "split" becomes unnecessary... Signed-off-by: Stephen Lyons <slysven@virginmedia.com> * Cut down on the length of some variable names (#1086) * Clarified use of QLatin1String * Adding GCC 5 for C++14 requirement (#1098) * Fixed "reset map colors" button to work (#1049) * Fixed "reset map colors" button to work That's a 5 year old bug out of the way! * Change travis configuration to use installed gcc-5 (#1108) * Change travis configuration to use installed gcc-5 * Fixes #1104 (#1109) Thanks to Rick for his prompt response on this * Fixed typo in GeyserLabel.lua Per #1112 * Fix locale to C Enforce uniform locale so scripts don't get tripped up on number representation differences (. vs ,) * IRC client update of UI and logic (#1072) * Add Communi Model & Util libs * Full refactor of IRC Client code & UI This retains the primary functions of the old IRC client but adds many more new IRC commands by default. The UI is slightly more organized and interactive now. * Adds QPointer to IRC Dialog pointer variable. * Update Lua `sendIrc()` slightly. * Added /msg command, Input History, and updated display colors. Input History stores a total of 8 previously entered commands the user can cycle through using the Up Arrow key. The previously include /msg command has been added again. Colors of some messages have been updated for (relatively) easier distinction of message lines. * Updated channel handling. * Fix CMake builds Things needed to fix CMake builds: - communi CMake builds needed `include_directories` explicitly specified - more MOC calls in communi - more explicit dependencies to MOC files in communi - ircmessageformatter must be compiled into mudlet * IRC3 with more Lua features (#1087) Changes IRC client to run with settings given per-profile, making the IRC client more visible and configurable within Lua. This adds configuration options to the settings window in the "Special Options" tab for configuring an IRC client on the host profile which opens the settings window. See the PR summary comment for details on available Lua functions and IRC Client commands. * Add a Code of Conduct (#1110) * Create CODE_OF_CONDUCT.md * Fix for the QMake warnings caused by the IRC library includes. * Finished formatting all filed & added braces (#1115) * Finished formatting all filed & added braces * Re-enabled map load on profile load (#1118) * Re-enabled map load on profile load - disabling it broke scripts that relied on map data w/o the map open * updated assignments. * Event-based gamepad functionality (#1106) * Add theme switching to code editor (#1095) * Fix more Toolbar & Button bugs (#1080) (#1090) * Fix bugs with Buttons & Toolbars. - Fixed change detection for location, orienation, and CSS data. - Buttons are not rendered unless they are part of a group, Toolbar/Menu. - ToolBars whos locations are changed now hide their previous TToolBar or TEasyButtonBar instance as needed. * Fix Buttons being left behind when parent toolbar is deleted. Also allows TAction to use QPointer<T> * Update old sourceforge page links (#1132) (#1134) * Fixed blank command separator to not separate as people expect * Replaced isTempKey, isTempAlias, etc with isTemporary (#1102) * also pushed isFolder() and setIsFolder() to the parent Tree class * Give meaning to encodings (#1107) * Delete dev_README Deleted outdated dev_README file. It also didn't seem to be used all that often, else people would have raised the fact that it's outdated... * Update travis to use Qt 5.9 for compiling (#1141) * Re-add single QT 5.6 build job Since we want to keep sure that the minimal Qt version is still compilable (currently we need at least 5.6), we add a single job for that version. * Refactor: rename QMake project/sub-project files (#1103) * refactor: rename QMake project/sub-project files It is down right confusing having several project files in the Qt IDE which are all labelled "src" - this commit renames the main project file from "src.pro" to "mudlet.pro" and also renames: * the lua_yajl module from src.pri to lua_yajl.pri * the luazip module from src.pri to luazip.pri it also: * adds a new communi.pri file (see below) I also took the liberty of renaming the project resource file from: * mudlet_alpha.qrc to mudlet.qrc (we are well past the alpha stage now!) and sorting a couple of items into the right order. By creating a local project file 3rdparty/communi/communi.pri it is possible to group the three separate but related communi libraries into one - as it happens the effective content of the new file is identical to that of the communi libraries own libcommuni/src/src.pri but it has some notes on how the portion of the upstream library that we are using was picked out of there; how to identify the effect version and how to (hopefully) recreate things if upstream gets revised in the future! By not using the upstream "src.pri"/"src.pro" files we avoid the confusion of having a generic "src" project in the Mudlet project as a whole and it is clearly identify-able as a file from the Mudlet project rather than the sub-project that it pulls into the whole thing... There is an minor nuisance side-effect in that having the main project src.pro file open in a Qt Creator IDE or stored in a session will get confused when this commit takes effect as the file will disappear from the repository - it will be necessary to close the project and open the new src/mudlet.pro project file (and possibly re-enter the build steps) - OTOH the main project will now be identified as "mudlet" in Creator rather than the previous "src". 😄 Signed-off-by: Stephen Lyons <slysven@virginmedia.com> * Removed accidental debug echo left in (#1151) * initialize int nIrcPort with dlgIRC::DefaultHostPort value. (#1153) * Removes false from sendIrc() return, ignores command filtering status (#1154) * Removes false from sendIrc() return, ignores command filtering status (not used). * Remove punctuation from lua messages * Improved layout of IRC preferences (#1152) * Fix get map events (#1120) * Adapt addMapEvent error messages to current standard * Fix getMapEvents() The function now returns a proper indexed table of objects that each describe a map event. * Add missing call to set Editor as editable. (#1159) * Fixed tab indent to be two spaces again * Fixed getComputerEncoding in case no key found It'll now return the original value passed - it's way better than returning the wrong value. * Add linux ci deployment (#1128) * Modify files needed for CI deployment of linux builds * Syntax fix in .travis.yml * tar the CI AppImages as well * Add lua sqlite runtime dependency * Unset LD variable for linuxdeployqt * Bugfix: unset LD_LIBRARY_PATH in every codepath * Test: install appmenu-qt5 for global menus * Revert 781ef97 (Install appmenu-qt5) The issue was that the first AppImage used Qt 5.6.2 instead of Qt 5.9 (which is how 3.2 was built and how it is build now after merging development in) * Remove manager again (#1163) * Improved map load output to be condensed down if it's all successful (#1117) OK! Would be happy to see that in as well for maps that take a while to load (>1s). * Re-enable search highlighting (#1096) * Remove mainwindow statusbar (#1162) * Replaced statustips with tooltips in the mapper menu * Removed statusbar * Tagging 3.3.0 * Initial QTextToSpeech additions * Further TTS additions and bugfixes * Wrapping TTS functions in #ifdefs for QT_TTS_LIB * Missing #ifdefs * More misplaced #ifdef lines for TTS functions * More QT_TTS_LIB ifs * Adding QTextToSpeech to cmake lists * Various TTS updates * Adding GetState to TLuaInterpreter.h * Minor fixes * Tts temporary fixes * Exchanging toLatin1 with toUtf8 * Fix cmake TTS module includesion * Install Qt 5.9 TTS module on linux * Use Qt provided definition for enabled modules We don't need (and possibly don't want) to provide our own definition for when a module is enabled. Let's use the Qt provided definition instead. * Output confirmation messages of enabled Qt modules * Use upper case "Gamepad" in output * ttsClearQueue crash fix * Fix for global speech events * Whitespace replacement * Various minor fixes * More minor fixes. * Cleanup: make CMakeList.txt have uniform indenting and adjust CMake message The minimum versions of the Qt TextToSpeech and Gamepad reported in the CMake project files needed tweaking. I have also tried to regularise the indentation in the Mudlet CMake project files to standardise on 2 spaces. Signed-off-by: Stephen Lyons <slysven@virginmedia.com> * Reformat: clean up code-loyout of new TTS TLuaInterpreter code Signed-off-by: Stephen Lyons <slysven@virginmedia.com> * BugFix: fixup a switch() in (int) TLuaInterpreter::ttsGetState(lua_State*) This switch was missing "break;" on all the cases and must have been defective without them - for all but the last case extra strings would have been pushed onto the lua stack but as the return value was 1 they may have been discarded... Signed-off-by: Stephen Lyons <slysven@virginmedia.com> * Enable Qt 5.9 Windows builds for testing * Deploy Qt 5.9 builds for testing * Fixing double-events for queueing purposes * Added missing symbol * Added missing symbol * Updating tts events to globals and adding voice * Create TTSValues.lua * Adding TTSValues to LuaGlobal * Adding TTSValues to Mudlet project * Update CMakeLists.txt * Update TTSValues.lua * Update Qt to 5.10 for trusty This is needed to hopefully have some TTS backend plugins in Linux * Queueing and crash fixes * Renaming tts funcs * Renaming internal functions * Adding functions for retrieving settings * Fixing typo * Making ttsGetState public * Improve error messages for ttsGetCurrentLine() It can be incredibly frustrating to just get a boolean back as an error message when you don't know what is going on * Update error message to be consistent with the new style * Aligned error messages to be consistent with the new style * Fix ttsClearQueue to return nil+error, not actually error on invalid index * Minor code formatting * Floats to doubles * Update documentation links
Should only be sending the events to whichever window is active, but I could be mistaken. Needs some testing. Note that for Windows users, you're limited to XInput devices, like Xbox 360/One controllers, some 3P controllers, and PS3 and PS4 controllers if you use the right software to spoof XInput with them. Other OS's should have access to all gamepads, although the QGamepad class is only really meant to handle Gamepads ala: https://en.wikipedia.org/wiki/Gamepad and might not function with joysticks/etc