* Fix Rewrite/Latch Time
* dependencies can now be specified in the schema
Signed-off-by: Paulchen-Panther <Paulchen-Panter@protonmail.com>
* dependencies can now be specified in the schema
Signed-off-by: Paulchen-Panther <Paulchen-Panter@protonmail.com>
* correct default config
Signed-off-by: Paulchen-Panther <Paulchen-Panter@protonmail.com>
* dependencies part 2
Signed-off-by: Paulchen-Panther <Paulchen-Panter@protonmail.com>
* Update README.md
* build: Update Windows to Qt5.15 (#798)
* Update compile howto
- to qt5.15
- minimal install overhead
- sync with vscode and cmake
* Update docs with more input
gh actions windows: Qt5.15 and msvc2019
* fix: Azure build
* feat: Add Trapezoid to LED Layout (#791)
* Update schema-ledConfig.json
extend schema with corner points (vertcal and horizontal values in percent)
* Update conf_leds.html
input fields for corner points
* Update content_leds.js
add corner values and calculation of led rows
* Update de.json
german title for corners
* CHanges 26.12.2019 Correction and PowerOff add
* GrabberFix PiCam
* VType as QString
* Unused variable deletet
* Missing new paramneters for GrabberFix
* Error json
* JSon error by user
* Crosscompile SutnikElf frok
* Warning unused variable
* Revert "Crosscompile SutnikElf frok"
This reverts commit 443fe80090.
* Fixing a linting error
* Adding Grabber Override Names
* Adding more Grabber Fix translations
* Missed one
* remove V4L2 Grabber Fix and Power Button from Dashboard
Signed-off-by: Paulchen-Panther <Paulchen-Panter@protonmail.com>
* Setting some limits to the trapezoid
Co-authored-by: SputnikElf <58911932+SputnikElf@users.noreply.github.com>
Co-authored-by: SputnikElf <sputnikelf@gmx.de>
Co-authored-by: Paulchen-Panther <Paulchen-Panter@protonmail.com>
* add Hue EntertainmentAPI + Forwarder & other Fixes (#592)
* whitespaces + typo fixes
* JS / LGTM fixes
* SSDP Handler crash fix
* MessageForwarder handlePriorityChanges Slave fixes
* use aboutToQuit Signal
* complete rewriten Hue Entertainment API structure
combined Philips Hue and Entertainment API
with new MbedTLS based SSL UDP Provider
* add required cross-compile submodules
* logical rebuild fn: initLeds, setLights + new logs
-more detailed checks and error handling inside iniLeds and setLights
- logical script procedure before ProviderUdpSSL init
- first steps for multiple ProviderUdpSSL usage
- better fallback support to old RestAPI, if entertainment api is not supported
- just 4 u LordGrey: new log fn for cosmetic config outputs ;)
* add OSX CompileHowTo - undo from CrossCompileHowTo
* whitespace fixes
* lightID toString fix
* fix unsigned int E-API + debug output
* bugfixes, reworked black signal detection, wizard:
- change device config field light-ids from int to string -> real unsigned int fix
- add signal detection brightness minimum threshold value
0.0 for 0% brightness - 1.0 for 100% brightness to count for blacklight signal detection
reason: input may not 100% black, like mine - i have a deep dark gray input signal
-> my threshold value is set to 0.005 for 0.5% minimum brightness = 1 (from max 255) to count as black
- wizard optimations, with fallback without entertainment support (beta state)
- whitespace fixes
* cleanup + minor fixes
* change fixed Hue UPD SSL config to _devConfig paras
* Hotfix SSL Connection, new light models, wizard:
- Fix UPD SSL Connection failed Problems
- add new supported gamut C light models: LCG002, LCA001, LCA002, LCA003
- wizard: extend fallback support to classic mode + hints
* whitespace, typo fix
* uncheck useEntertainmentAPI, if noAPISupport detected + hint
* coredump fix -> add _blackLightsTimer nullptr init
* code cleanup / remove old debugs + whitespacefixes
* add gamut C LCP001, LCP002
* SSL UDP config made more flexible + remove qDebug
-> switch to hyerion.ng _log
-> replace logCommand with verbose
-> code cleanups etc...
* extended mbedtls debugging infos
* add adjustable ssl timeout settings
* error handling
* streamdebugger bugfixes
* UPDSSL psk / psk_identity bugfixes! + hue wizard fn typo fix +
- verbose option available without dependencies
- whitespace fixes
* Philips Hue Assistant now recognizes non-original bridges better...
+ Added note if no clientkey is set when using the entertainment API
+ User creation (+ clientkey) for non-original bridges can now also be used
+ Minor changes and bug fixes
* CMAKE mbedTLS detection
* minor bug fixes + code cleanups
* FindMbedTLS.cmake remove Path-Hints + wizard.js: ajax timeout handling
Test - content_grabber.js: run relevant code only, if V4L2_AVAIL is true:
conf_grabber don't displays other devices, if V4L2 is not available
* compile mbedtls via cmake as static lib
* remove libmbedtls-dev from compileHowto / scripts
* Fix Windows build
* Fix windows build (part 2)
* removed unnecessary osx x11 include directory path
* QTimer Shutdown bugfix
* cmake win32 fix + minor bugfixes
* cmake debug msg used mbedtls libs
* Bugfix: noSignalDetection wasn't switchedOn again
if no signal was previously detected
* Some code fixes based on alerts from lgtm.com
Co-authored-by: Paulchen Panther <16664240+Paulchen-Panther@users.noreply.github.com>
* Update devices allowing rewriteTime
* Fix typo
* Add missing "general_btn_saverestart" text to en-file
Co-authored-by: Paulchen-Panther <Paulchen-Panter@protonmail.com>
Co-authored-by: Paulchen Panther <16664240+Paulchen-Panther@users.noreply.github.com>
Co-authored-by: brindosch <edeltraud70@gmx.de>
Co-authored-by: Ben Williams <desertblade@gmail.com>
Co-authored-by: SputnikElf <58911932+SputnikElf@users.noreply.github.com>
Co-authored-by: SputnikElf <sputnikelf@gmx.de>
Co-authored-by: SJunkies <41368976+SJunkies@users.noreply.github.com>
* small collection of bugfixed debugger hints / warnings
* 'toStdVector' has been explicitly marked deprecated
* fixed double zip naming in artifacts
* V4L2 WebUI Fix
* Some code fixes based on alerts from lgtm.com
* only execute dynamic v4l2 enum code, if V4L2_AVAIL
* very high critical bugfix ;)
* merge fix
* some lgtm.com fixes
* lgtm fixes
* undo localtime_r fix
Co-authored-by: Paulchen Panther <16664240+Paulchen-Panther@users.noreply.github.com>
* whitespaces + typo fixes
* JS / LGTM fixes
* SSDP Handler crash fix
* MessageForwarder handlePriorityChanges Slave fixes
* use aboutToQuit Signal
* complete rewriten Hue Entertainment API structure
combined Philips Hue and Entertainment API
with new MbedTLS based SSL UDP Provider
* add required cross-compile submodules
* logical rebuild fn: initLeds, setLights + new logs
-more detailed checks and error handling inside iniLeds and setLights
- logical script procedure before ProviderUdpSSL init
- first steps for multiple ProviderUdpSSL usage
- better fallback support to old RestAPI, if entertainment api is not supported
- just 4 u LordGrey: new log fn for cosmetic config outputs ;)
* add OSX CompileHowTo - undo from CrossCompileHowTo
* whitespace fixes
* lightID toString fix
* fix unsigned int E-API + debug output
* bugfixes, reworked black signal detection, wizard:
- change device config field light-ids from int to string -> real unsigned int fix
- add signal detection brightness minimum threshold value
0.0 for 0% brightness - 1.0 for 100% brightness to count for blacklight signal detection
reason: input may not 100% black, like mine - i have a deep dark gray input signal
-> my threshold value is set to 0.005 for 0.5% minimum brightness = 1 (from max 255) to count as black
- wizard optimations, with fallback without entertainment support (beta state)
- whitespace fixes
* cleanup + minor fixes
* change fixed Hue UPD SSL config to _devConfig paras
* Hotfix SSL Connection, new light models, wizard:
- Fix UPD SSL Connection failed Problems
- add new supported gamut C light models: LCG002, LCA001, LCA002, LCA003
- wizard: extend fallback support to classic mode + hints
* whitespace, typo fix
* uncheck useEntertainmentAPI, if noAPISupport detected + hint
* coredump fix -> add _blackLightsTimer nullptr init
* code cleanup / remove old debugs + whitespacefixes
* add gamut C LCP001, LCP002
* SSL UDP config made more flexible + remove qDebug
-> switch to hyerion.ng _log
-> replace logCommand with verbose
-> code cleanups etc...
* extended mbedtls debugging infos
* add adjustable ssl timeout settings
* error handling
* streamdebugger bugfixes
* UPDSSL psk / psk_identity bugfixes! + hue wizard fn typo fix +
- verbose option available without dependencies
- whitespace fixes
* Philips Hue Assistant now recognizes non-original bridges better...
+ Added note if no clientkey is set when using the entertainment API
+ User creation (+ clientkey) for non-original bridges can now also be used
+ Minor changes and bug fixes
* CMAKE mbedTLS detection
* minor bug fixes + code cleanups
* FindMbedTLS.cmake remove Path-Hints + wizard.js: ajax timeout handling
Test - content_grabber.js: run relevant code only, if V4L2_AVAIL is true:
conf_grabber don't displays other devices, if V4L2 is not available
* compile mbedtls via cmake as static lib
* remove libmbedtls-dev from compileHowto / scripts
* Fix Windows build
* Fix windows build (part 2)
* removed unnecessary osx x11 include directory path
* QTimer Shutdown bugfix
* cmake win32 fix + minor bugfixes
* cmake debug msg used mbedtls libs
* Bugfix: noSignalDetection wasn't switchedOn again
if no signal was previously detected
* Some code fixes based on alerts from lgtm.com
Co-authored-by: Paulchen Panther <16664240+Paulchen-Panther@users.noreply.github.com>
* fix v4l2 standard
* ignore v4l2 meta devices
* added resolution, framerate and device dropdown list to WebUI (thx to @Lord-Grey & @b1rdhous3)
* Fix for kernels prior to v4.16
* Device names added & WebUI adapted
* fix: Resolve the enable state for v4l and screen capture
* Use instance index instead of pointer
* Second try
* fix(QtGrabber): QScreen ownership
* Remove v4l2 compState listener
* add SSDP name field
* YALL - yet another led layout
* led layout migration
* add initial vscode config
* merge clearAll with clear, rename Hyperion::compStateChange
* simpler components api
* Corrected code formatting
+ triggered PR build
* fix: regression from #636
* Support for color patterns
Co-authored-by: Paulchen Panther <16664240+Paulchen-Panther@users.noreply.github.com>
* Handle Exceptions in main & Pythoninit
* Have SSDPDiscover generic again
* Have SSDPDiscover generic again
* Change Info- to Debug logs as technical service messages
* Nanoleaf - When switched on, ensure UDP mode
* Include SQL Database in Cross-Compile instructions
* Fix Clazy (QT code checker) and clang Warnings
* Stop LedDevice:write for disabled device
* Nanoleaf: Fix uint printfs
* NanoLeaf: Fix indents to tabs
* NanoLeaf - Add debug verbosity switches
* Device switchability support, FileDevice with timestamp support
* Nanoleaf Light Panels now support External Control V2
* Enhance LedDeviceFile by Timestamp + fix readyness
* Stop color stream, if LedDevice disabled
* Nanoleaf - remove switchability
* Fix MultiColorAdjustment, if led-range is greater lednum
* Fix logging
* LedFileDevice/LedDevice - add testing support
* New "Led Test" effect
* LedDeviceFile - Add chrono include + Allow Led rewrites for testing
* Stabilize Effects for LedDevices where latchtime = 0
* Update LedDeviceFile, allow latchtime = 0
* Distangle LinearColorSmoothing and LEDDevice, Fix Effect configuration updates
* Updates LedDeviceFile - Initialize via Open
* Updates LedDeviceNanoleaf - Initialize via Open, Remove throwing exceptions
* Updates ProviderUDP - Remove throwing exceptions
* Framebuffer - Use precise timer
* TestSpi - Align to LedDevice updates
* Pretty Print CrossCompileHowTo as markdown-file
* Ensure that output is only written when LedDevice is ready
* Align APA102 Device to new device staging
* Logger - Remove clang warnings on extra semicolon
* Devices SPI - Align to Device stages and methods
* Fix cppcheck and clang findings
* Add Code-Template for new Devices
* Align devices to stages and methods, clean-up some code
* Allow to reopen LedDevice without restart
* Revert change "Remove Connect (PriorityMuxer::visiblePriorityChanged -> Hyperion::update) due to double writes"
* Remove visiblePriorityChanged from LedDevice to decouple LedDevice from hyperion logic
* Expose LedDevice getLedCount and align signedness
* Handle Exceptions in main & Pythoninit
* Have SSDPDiscover generic again
* Have SSDPDiscover generic again
* Change Info- to Debug logs as technical service messages
* Nanoleaf - When switched on, ensure UDP mode
* Include SQL Database in Cross-Compile instructions
* Fix Clazy (QT code checker) and clang Warnings
* Stop LedDevice:write for disabled device
* Nanoleaf: Fix uint printfs
* NanoLeaf: Fix indents to tabs
* NanoLeaf - Add debug verbosity switches
* Device switchability support, FileDevice with timestamp support
* Nanoleaf Light Panels now support External Control V2
* Enhance LedDeviceFile by Timestamp + fix readyness
* Stop color stream, if LedDevice disabled
* Nanoleaf - remove switchability
* Push progress
TODO: rework RESET, probably to main.cpp again
* resetPassword rework
* enable administration restriction
* add short cmd for userdata
* Js apis
* Refactor JsonCB class
* Add userToken Auth
* Feat: Close connection if ext clients when def pw is set
* Feat: Protect db against pw/token tests
* WebUi PW Support (#9)
* Initial WebUi Password Support
* Small changes
* Initial WebUi Password Support
* Small changes
* Basic WebUi Token support
* added "removeStorage", added uiLock, updated login page
* Small improvments
* Small change
* Fix: prevent downgrade of authorization
* Add translation for localAdminAuth
* Feat: Show always save button in led layout
* Revert "Feat: Show always save button in led layout"
This reverts commit caad1dfcde.
* Feat: Password change link in notification
* Fix: body padding modal overlap
* Feat: Add instance index to response on switch
* prevent schema error
Signed-off-by: Paulchen-Panther <Paulchen-Panter@protonmail.com>
* Feat: add pw save
* Feat: callout settings/pw replaced with notification
* Remove clone & index in ledConfig
* Additional index correction
* Rename ledsConfig maximum and minimum in max and min
* Rename ledsConfig hscan and vscan with h and v
* Optimize ledColorsStream
* tiny correction
* Added new restricted ip api checkbox, initial instance localStorage support
* Added instance notification, added correct instance switch after page refresh (to last current instance)
* Optimize intention in wizard.js
* ":" from i18n json to code
* Changes for Responsive about site
* Feat: Add image sender to webui
This PR adds a new image sending feature to the webui and extends the api accordingly. In javascript the processing of images is horrible slow (without WASM), so the solution is at the API side with out-of-the-box power of Qt.
- The image cmd accepts now a raw image that is encoded with base64. Supported are all image formats that qt supports (enough)
- There is no real size limit. It will be automatically scaled down to max 2000px width or height according to aspect ratio
- It's possible to scale down through a new "scale" property if developer wants that.
- Test were successfull until 3MP pictues, 4k+ closes the websocket on browser side, so 2k is a decent value
- Adds a new image streaming feature from browser (tabs/applications/complete desktop as target). This works just if used with HTTPS PR#612 AND with a recent version of Firefox/Chrome
* Feat: Add SSL support for webserver + websocket
Finally, Hyperion reaches the SSL century!
- Uses by default a internal key and certificate to provide working HTTPS out-of-the-box
- Your browser won't like that, for a green ssl seal next to the browser address bar you need to use Let's Encrypt with a own legit domain. This is out of the scope of Hyperion
* Moved Token Management to Network Settings
* Fixed alignment of Token Settings Div if Help Tables are switched off
* Added Chars counter, new format of the div panel (footer), change i18n
* Added Explanation for Token Management
* Moved Token Management to Network Settings
* Fixed alignment of Token Settings Div if Help Tables are switched off
* Added Chars counter, new format of the div panel (footer), change i18n
* Added Explanation for Token Management
* i18n Fix
* Revert "i18n Fix"
This reverts commit 0186290689d8b140310d5376617a3d43429bb632.
* i18n Fix
libqt5sql5-sqlite appended to the dependency list
lgtm alerts fixed
added message forwarder to global settings
subscribe type leds-update appended (thanks @Brindosch)
Boblight server port check instead of error message
Race Condition of different priorities are prevented at startup
Signed-off-by: Paulchen-Panther <Paulchen-Panter@protonmail.com>
- Added CodeDocs config file for customization
- Fixing LGTM alerts
- LGTM bug fixed again
- added token option to hyperion-remote
- fix DBManager::getDB()
- next bugfix
- correct broken signal from SettingManager to Hyperion
- Token list is created after the schema is fetched
Signed-off-by: Paulchen-Panther <Paulchen-Panter@protonmail.com>
- versionBranch was renamed to watchedVersionBranch (including all variables and strings. even config)
- Enhancements for JSON Editor. Detection of advanced was not so good. Now you can see advaced if advanced or expert is selected. expert is the highest and should also make advanced visible.
- The levels default, advanced, expert are now also respected in the information tabs. Same logic as for the forms itself. Expert is highest and will also show all advanced settings.
- Build problems:
- Qt 5.13 obsoleted some calls that were used in our Qt library.
- The PhilipsHueLight and PhilipsHueBridge classes did not use the private logger class
- Undo changes:
- In Commit e6c2e7e, I made changes that were not covered. An apology goes to @b1rdhous3
Signed-off-by: Paulchen-Panther <Paulchen-Panter@protonmail.com>
- added native RPi docker image instruction to CompileHowto.md
- version channel is received from JsonAPI instead of reading from the configuration file
Signed-off-by: Paulchen-Panther <Paulchen-Panter@protonmail.com>
Update Check:
- New Property which allows you to select stable or beta branch checks.
- This would rely on current versions.json. It selects the first version in list with you branch (stable, beta). If it is newer hyperion tells you that a update is available
- For this we should keep our deployment pipeline in mind to format the json this way.
Cleanup:
- I cleaned up the WebUI from dead links and the support page with the affiliate links and donation adresses as we don't have them under our control. Currently only disabled, not deleated
Switcher:
- If the address which is published via bonjour serve is IPv6 the Switcher was bugged. Fixed that.
- Increases linker performance under Linux builds by using Gold linker, if available
- ccache is used if available
- removed statistic class (Stats.cpp) from project due to the missing result (sorry @Brindosch)
- add LGTM bandges for code analysis overview
Signed-off-by: Paulchen-Panther <Paulchen-Panter@protonmail.com>
- Video format MJPEG implemented (libjpeg/qimage)
- Inactive priorities are now skipped correctly (PriorityMuxer.cpp line 297)
- v4l configuration section replaced with an object (preparation for #542)
Protocol Buffer reimplemented to receive image data from third-party apps
The status of the component "UDPListener" is now displayed correctly in WebUI
Global signal names for WebUI added
fix ge2d mode
added ge2d single/combined ioctl call
speed up imageresampler
speed up amvideocap0 mode
added config to choose if amvideocap0 or ge2d grabber should be used
- More i18n
- Easy use of mutual exclusion in JsonAPI with QMutexLocker
- Smoothing type "linear" hidden in the WebUI, because there is currently only one
- Message forwarding implemented again
- For compatibility to home assistants and other remote controls, "activeEffects" and "activeLedColor" has been added to the JSON-RPC
- FlatBuffer clear now the Priority on disconnect
- The information "available V4L2 devices" is now only displayed if the device list is not empty
- LED device "PiBlaster" excluded from OSX build
* - grabber auto off when not set as active prio
- join aml and fb - on aml platform both grabbers are needed, so they joind in one module and share one prio. user don't the the nasty magic behind
- aml: preparation for direct ge2d access
* just save it, in the middle of ge2d impl
* fix compile issues
* now grabbing works basicly
* add 3d support for ge2d
* next step, we got some video from aml
* switch back to rgba
* remove unfinished ge2d stuff
* commit missing changes
* some urgent fixes, needs some beautifying, but it works now
* fixes and refctoring
* move setvideomode to common place
* implement more croping and 3d support
* more api unification
* more refactoring
* osx fix
* next step
* add a mock for osx grabber. Now it is possible to test compile on none osx platforms.
* more unifications ...
* remove obsolete includes and grabbers are not dyn allocated. dispmanx needs rework an probaly not work atm
* first version of dispmanx mock. it compiles, but outputs a black image
* now dispmanx mock works!
* activate mocks in travis linux build
prepare dispmanx to rgb image out
* dispmanx now with image rgb output
fix deadlock with w/h -1 in grabber v4l
cleanups
* fix json
* fix some runtime stuff
* Update FramebufferWrapper.cpp
fix missing code
* unify grabframe
* 3d and croping for amlogic
* fix setimage not working
* make use of templates
save some codelines
* save more code lines
* implement set videomode via json api
* refactor grabbers:
- new base class
- move shared code to base class
* fix osx
* rework all cmakelist files with auto file collection. except leddevices (need further restructuring)
* store current video and grabbing mode
* add json stuff
* remove grabbingmode - we do not want to expose it
* add dynamic smoothing first step
* extend prio muxer to hold smoothing preset id
* add icons for systray
* fix missing changes in prio muxer
* implement specific smoothing params for effects
* refactoring: std::min/max to qMin/Max
* some code optimization
* fix schema and translation
* revoke change of python include order
* fix eol in effect shemas
* optimize random,candle and fadecandy json schemas
* add deployment
* add correct api key
* fix cmake lists and add heroku app name
* Update .gitmodules
sync modules with upstream
* add possibility to start hyperiond as systray app
* cleanup
* - new command line options: --desktop --service to set desired mode (systray icon / console only)
- auto detect x server - if avail run in gui mode
- on osx always run in gui mode
- use existing icon from webconfig, instead of own icon
- add ability to no gice a config file name. If config not given, default config file will be set (home dir, or hyperiond dir, depending on writable state)
* fix warnings and compile error
* use own icon for systray purpose
* use new logo
* - set application properties
- fix force service mode
* update
* resolve qt5w
* test
* Go down
cmake python search won the price for crazy lib searches!
* 2.7 python forced, RPATH
* upstream
* ...
* update
* ...
* 2.7.12 py it picks random versions...
* max 7.12.5....
* Test 14.04 tests
* revoke schema split
* add "getAutoCorrectedConfig" function
* revoke schema split
* revoke schema split
* revoke schema split
* Prevent compiler error if none grabber is available
* revoke schema split
* add "getAutoCorrectedConfig" function
* revoke schema split
* remove "configMigrator"
* remove "configMigrator"
* Change TestConfigFile to show how the autocorrection works
* revoke schema split
* revoke schema split
* remove "ConfigMigrator"
* remove "ConfigMigrator"
* remove "ConfigMigratorBase"
* remove "ConfigMigratorBase"
* Add QJsonUtils.h
* added ability "ignore-required"
It has been added the ability to correct the configuration without having to pay attention to the keyword "required" in the hyperion schema
* Allow Comments in Hyperion Schema
* add ability to ignore the "required" keyword in hyperion schema
* add ability to ignore the "required" keyword in hyperion schema
* add ability to ignore the "required" keyword in hyperion schema
* //Allow Comments in Hyperion Schema
* Update jsonschema.py to version 0.8.0 to support ...
references in json schema
* add RefResolver from jsonschema.py to resolve
references in hyperion schema
* remove dupe code
* split the hyperion schema in separatly files
For better readability
* add function "resolveReferences" to resolve
references in hyperion schema.
* remove CURRENT_CONFIG_VERSION
* remove CURRENT_CONFIG_VERSION
* split the hyperion schema in separatly files
For better readability
* Create schema-backgroundEffect.json
* Add the rest of the Hyperion schema via upload
* Remove Comments in config file
* Add return variable to function writeJson().
fix function resolveReferences().
edit function load() to handle QPair result from schemaChecker.
* edit function validate() to return QPair variable
* fit function loadEffectDefinition()
* fit function checkJson()
* Expand error check by dividing
"_error" variable in "_error" and "_schemaError".
Replace variable "bool" in validate() in QPair
* Extend function "cmd_cfg_set" to handle auto correction
* Extend function "loadConfig" to handle auto correction
* fix function loadConfig()
* add limit for write to leddevice
* add to default config
* add i18n
* extend xmas effect
* fix indention
* add check for minimum brightness
* adapt effects to fading and new minWriteTime
* remove old latchTime
rename minimumWriteTime to latchTime
make it as dev specific option
* set default for rewriteTime to 1s
pause smoothing on color too
* reenable smoothing for color - it looks nicer :-)
* fix timeout timer
