Icedtea-web splashscreen implementation
Pavel Tisnovsky
ptisnovs at redhat.com
Mon Mar 12 04:11:19 PDT 2012
Hi Jiri,
I'd like to comment the last patch with new test framework functionality:
1) the idea of adding some delay to see/test splashscreen is great, but is not
it better to add some annotation to such tests instead of using "magic" test name?
Similar functionality is used in JTreg tests (not "true" annotations, but similar approach).
I'm sure we'll need to add more per-test parameters in the future.
2) some notes to the patch:
tests/netx/jnlp_testsengine/net/sourceforge/jnlp/ResourcesTest.java:
- seems there's a bad indentation for the last 11 lines
tests/netx/jnlp_testsengine/net/sourceforge/jnlp/ServerAccess.java
- bad indentation in a method getIndependentInstance()
- -//- for executeJavawsHeadless
- -//- executeJavaws, executeBrowser
- -//- splitArray (it needs JavaDoc too, contains double ";;" on one line etc.)
I think it could be simplified to use one loop only and two index variables
(we can talk about this personally)
- I'm probably wrong but is seems that all applet/jnlp tests are using only "localhost" network interface.
If it is true I'm afraid we miss some iptables-related or IPv6 related issues.
Cheers,
Pavel
(Hint: Eclipse Ctrl+F)
Jiri Vanek wrote:
> Hi All!
>
> This patch is initial version of three and half connected things. Applet
> + javaws splashcreen and testing of applets in firefox (and testing of
> splascreen).
> I'm sorry for two days delay, which was caused by more testing and
> creating this "document" and its attending animation. I'm also really
> sorry for its length.
>
> People cc-ed (as patch is to big to pass to distro without approve)
> Deepak - as he will probably do the overall review as there nearly
> nothing left from his code and will definitely want to see the
> in-firefox tests concept.
> Roman - as he was promised (was persuaded O:) to check
> PluginAppletViewer class As is very similar to Deepak's original one
> Mark - as he is probably able to do an approve upon distro-pkg list.
> Pavel - who wanted to check graphics rendering and firefox-testing changes
> Omair - just to kept him in loop ;)
>
> No rebukes to "logo" will be now accepted (kdding ..but... O:)..to much
> effort to vectorization of the "agreed new one"!)
> I have kept the ICEDtea composition, because without this design (with
> IcedTea one) it does not look nice :(( Iced do not looks enough frozen
> and Tea does not looks like tea leaf stalk (But ICED looks frozen enough
> and tea looks like stalk ;) )
> I have decided to kept ICEDtea, because we _do not have logo_ :(( And
> because I consider it as art :) And way of art must be free in mind O:)
> However - if it will be comon wish or command, it is just change in two
> Strings so I can (so sad) adapt it. Horrible image with camelcase is
> also attached.
> Animation is rising watterlevel inside ICED and shinig metal web. When
>
> Splashscreen(s) are designed to be easily replaceable, and are providing
> set of utilities for lunching and disposing it. When ratio of WxH is to
> big or applet is to small then alternate - plain IcedTea web (plugin)
> [with little bit animated web (plugin)] is shown.
> Test framework was enriched for slow downloading. If the file have
> prefix "XslowX" then its downlaoding (even on localhost) will take
> 10seconds. It is useful for observing splashscreens. Also was enriched
> for lunching browser, which is necessary for testing applets.
> Several minor changes in authors, makefile, images and so on are also
> included.
>
> *Splashscreens*
> _integration_
> The gate for applet's splashscreen is PluginAppletViewer. It is
> responsible for show it when applet is to be initialized and for
> removing it when applet starts. Little bit more complicated is showing
> of error to user when some exception during initialization occurs. Those
> exceptions are mostly thrown from netx which do not have dependence on
> plugin.
> To forward error to applet is used SplashUtils.showError set of
> overloaded methods. SplashUtils have also factory method which handles
> which splash (type and purpose (javaws x applet)) is about to be shown
> (if any - see bonuses)
> There are actually three and half types of splashscreen. Applet one
> (which have error variation also), javaws one (forced by specification,
> which also forces title,vendor and homepage with description to be
> shown), and java's -spalsh which is showing image until control is
> forwarded to netx.
> All those four splashscreens are cooperating quite well and support of
> custom splashses (by -J-splash or by <icon kind="splash"...>) are not lost.
> _implementation_
> The splashscreen can be any class which implements interface
> SplashScreen. it must be able to paint itself, to add itself as
> component to show error, to be in composed of vectors.....and to be nice:)
> The factory method have parameter whether to provide splash for javaws
> (it is showing also information from information element (vendor,
> description,homepage,title..) - specification commanding) or for applet
> -it have also plugin above the tea-leaves.
> If javaws die, then splashscreen is hidden and netx is handling error on
> his own. When applet crashes, then it tries to show error screen. If
> showing of error screen is successful, spalsh turns dark and is
> providing click-able area ("button") which when clicked shows localized
> (whole the splash is localized) error dialog with some information about
> crash (most important exception) and with link to icetea-web wiki and
> with information how to post the bug
> The showing of error screen is very successfull, but one case I was
> unable to catch - when connection with server dies in middle of resource
> transfer and TODO - I have forgot to test signatures and stuff around:-/
> - but this should be handled n netx-way too.
> I'm showing also version on splash (and forwarding it to error mesage)
> (ok?) - Looks to me like good idea....
> _pitfalls_
> I have noticed that when more then one applets are presented in page,
> then splashes sometimes shows, sometimes no.... I have tried to
> synchronize (I can be wrong!) in Launcher and in PluginAppletViewer, and
> it looks like it helped but still this occurs (quait randomly :-/)
> Sometimes(less then 10% of lunches, randomly spread) freeing of applet
> takes quite long time, and it looks like C-error (Can not fetch applets
> id from java side), But it is rare and always exit at the end.
> _bonuses_
> I have added recognition of two environment variables -
> ICEDTEA_WEB_PLUGIN_SPLASH and ICEDTEA_WEB_SPLASH. Those variables alow
> user to change splahs screen before lunchtime and - what is real reason
> for those two variables - to disable splashscreen at all. I really can
> imagine some developer really tired of looking to same splash during his
> appelt/javaws_app testing phase.
> I have also kept inside Deepaks origianl implementation and my testing
> one. I like Deepak's one, and Pavel really likes my testing one which is
> really alternative :) There was left also dialog for testing
> splashscreen. I left it in for case That icedtea-web 3.0 will want new
> celebrating one;)
> However both secondary splashscreens can be switched by env.vars, I can
> remove them (although it will breake my hart;) without any harm.
> When "custom" (or "none") splash is defined, then javaws "java's"
> -splash is ignored too.
>
> *Testing*
> _splash tests_
> I was "facing really huge problem" When I wanted to see spalshscreen
> inside browser - it was always to quickly lost :) So Ihave added magic
> prefix XslowX to be rocgnized by internal server. When some resource
> starts with this prefix, its without-prefix-version is returned as
> resource. The difference is that its downloading is delayed to 10s
> (10%per seccond). It is done by spliting this resource's byte[] to 10
> peaces. This throws BrokenPipe Exception, but is working fine.
> Minor improvement inside ServerAcces.main() is tighter integration with
> testing framework - It can now be lunched instead of eg python
> smallHTTPserver and working directory is set in same way as during tests
> by -D property (so it can be lunched with same environment settings as
> in-ide tests)
> Second problem I was facing was automated testing of applets in firefox.
> The integration of "lunch html with browser, examine stdout/stderr,kill
> firefox" required just small changes and is working really well.
> But there is one big TODO - to configure (in make?) browser so it works
> with JUST BUILT iced-tea-web-plugin. Currently it expects to have
> firefox well configured (= to link
> /usr/lib/mozzila/libs/plugins/icedteaweb-plugin.so (or wherever firefox
> have plugins) with the one in --prefix path). But it is not the best
> way. To future I'm thinking about customised "embeded" firefox or to
> some link-plugin makefile target (run under root) which fill create
> this link (currently iced tea-web's install do not create this link, it
> is done in RPM post), or to have some local firefox which can be
> linkable without root and provide --with_browser.... - *any ideas
> welcomed* !
> And one small to do - in some future integrate awt-robot (+stdin/out)
> listeners
> _new tests_
> One new automated test is added to demonstrate the powers of testing of
> applets inside browser
> Bunch of new test which are using XslowX to allow everybody to see
> various cases of spalshscreens. Those tests are not automated and are
> testing custom splash, broken custom splash, my splashes... broken
> stuff..both for applets and javaws
>
>
> Atatchements:
> javaws_splash.png - "screenshot" of vector implementation, also used as
> java's -splash
> horribleIcedTea.png - nearly the same, but with camelcase to make your
> judge easier
> splasScreen-All2.diff - huge patch which si including all other together
> splasScreenImpls2.diff - implementation of current default splashscreen
> with all support classes
> splasScreenSplashs.diff - all supporting classes, interfaces utilities
> common for whole "splash sreens stuff"
> splasScreenImpls1.diff - implementation of old Deepaks one and
> alternative testing one
> splasScreen-integration.diff - all modified classes inside icedtea-web,
> which are calling my splash-screen implementation(s) Needs to be
> properly reviewed O:)
> splasScreenTests2.diff - all new tests
> http://download.eng.brq.redhat.com/scratch/jvanek/gif.gif - url to
> animation showing applet's (and javaws's) splash in work. have 90MB!!!
> desribed at the and of email - especially two sizes (full resizing at
> the end) of spalsh and error screen and error handlng and diference
> between javaws's one and plugin's one are worthy to be seen.
>
> Changelogs:
>
>
> splasScreen-integration.diff
> 2012-03-02 Jiri Vanek <jvanek at redhat.com>
>
> Integrated splashscreen, added authors, added bitmap splash,
> * AUTHORS: added Jan Kmetko as the person behind initial bitmap
> splash design
> * Makefile.am: (edit_launcher_script) enriched for replacement
> JAVAWS_SPLASH_LOCATION inside javaws.in by real
> datadir/PACKAGE_NAME/javaws_splash.png value
> (install-exec-local) enriched for installing javaws_splash.png from
> NETX_SRCDIR to datadir/PACKAGE_NAME
> *NEWS: mentioned splash screen
> * extra/net/sourceforge/javaws/about/resources/notes.html: added
> Danesh, Saad, me and Jan as missing developers and artist
> * launcher/javaws.in: added variable SPLASH_LOCATION, initiated
> during make by real path to javaws_splash.png location
> this splash is shown when application is lunched not-headless and
> if no custom spalsh is defined, -J-splash also overwrite this setting
> * netx/javaws_splash.png: new Binary file for primary bitmap loading
> * netx/net/sourceforge/jnlp/GuiLaunchHandler.java: implemented
> missing default splash screen with vector spalsh and with information's
> element content
> * netx/net/sourceforge/jnlp/JNLPSplashScreen.java: fixed to be able
> to use custom image or (if not defined or broken) our vector one
> * netx/net/sourceforge/jnlp/Launcher.java: added hooks which will
> force applet to show error screen in case of error
> change on lines @@ -696,22 +705,30 @@ probably does not work, and
> it is the case when connection with server is interrupted in middle of
> downloading jar
> (createApplet) and (createAppletObject) were made synchronized to
> prevent several applets lunched in one time to touch each other
> * netx/net/sourceforge/jnlp/NetxPanel.java: added field, getter and
> setter for its XEmbededFrame
> * plugin/icedteanp/java/sun/applet/PluginAppletViewer.java: main
> (createPanel)'s run is adding splash screen after start of
> onitialization and removing it before init.
> (framePanel) returns PluginAppletViewer and is creating sdplash
> (removeSplash) new method to handle spalsh removing
> (showErrorSplash) new method to be called by hooks to swap splash to
> error state
> * netx/net/sourceforge/jnlp/runtime/Boot.java): (name) and (version)
> made public
>
> splasScreenSplashs.diff
> 2012-03-02 Jiri Vanek <jvanek at redhat.com>
>
> added support classes to allow easy splashscreen implementation
> * netx/net/sourceforge/jnlp/resources/Messages.properties: added
> localization strings for JEditorPaneBasedExceptionDialog and
> DefaultSpalshScreen12 and InfoItem
> * netx/net/sourceforge/jnlp/splashscreen/SplashPanel.java: interface
> to control each splash screen implementation
> * netx/net/sourceforge/jnlp/splashscreen/SplashUtils.java: mainly
> factory and destructor for splashes
> There are two environment variables ICEDTEA_WEB_PLUGIN_SPLASH and
> ICEDTEA_WEB_SPLASH which can change before lunch time splashscreen and -
> what is their main reason to allow to disable it
> *
> netx/net/sourceforge/jnlp/splashscreen/parts/BasicComponentSplashScreen.java:
> is implementing interface and extending jcomponet and is handling most
> basic setters/getters. Ideal forefather for all splashes
> *
> netx/net/sourceforge/jnlp/splashscreen/parts/DescriptionInfoItem.java:
> extending InfIotem with kind. Is handling description value of
> information element
> * netx/net/sourceforge/jnlp/splashscreen/parts/InfoItem.java:
> object for handling items in information element - vendor, homepage and
> title. Is i18n.
> *
> netx/net/sourceforge/jnlp/splashscreen/parts/InformationElement.java:
> handling all InfoItems from information. Is selecting the proper
> description for each purpose as specification says
> *
> netx/net/sourceforge/jnlp/splashscreen/parts/JEditorPaneBasedExceptionDialog.java:
> dialog for showing exception and more information about crash in plugin.
> Can lead to icedtea-web homepage
>
> splasScreenImpls1.diff
> 2012-03-02 Jiri Vanek <jvanek at redhat.com>
>
> Adding optional splashscreens and testigne class for development
> *
> netx/net/sourceforge/jnlp/splashscreen/impls/CircleSplashScreen.java:
> Deepaks original implementation. I like it:) Have incorrect behaviour
> on pages with different then white background
> *
> netx/net/sourceforge/jnlp/splashscreen/impls/DummySpalshScreen.java:
> experimental splashscreen, worthy for testing purposes. And Pavel likes
> this alternative one!
> *
> netx/net/sourceforge/jnlp/splashscreen/impls/SpalshScreenTest.java:
> dialog window for testing splashes without icedtea-web
>
> splasScreenImpls2.diff
> 2012-03-02 Jiri Vanek <jvanek at redhat.com>
>
> Implementation of default splashscreen
> *
> netx/net/sourceforge/jnlp/splashscreen/impls/DefaultSpalshScreen12.java:
> default implementation of splash. The nicest one!, localized and most tuned
> *
> netx/net/sourceforge/jnlp/splashscreen/impls/defaultsutils/ControlCurve.java:
> base clsss for calculating vector curves
> *
> netx/net/sourceforge/jnlp/splashscreen/impls/defaultsutils/Cubic.java:
> helpful class for calculating cubic equation
> *
> netx/net/sourceforge/jnlp/splashscreen/impls/defaultsutils/ImageFontCutter.java:
> base class for cutting fonts from images
> *
> netx/net/sourceforge/jnlp/splashscreen/impls/defaultsutils/MoovingText.java:
> by extending ImageCutter is cutting text from gradient colors (shadow
> moving web)
> *
> netx/net/sourceforge/jnlp/splashscreen/impls/defaultsutils/NatCubic.java:
> base class for calculating natural cubic spline
> *
> netx/net/sourceforge/jnlp/splashscreen/impls/defaultsutils/NatCubicClosed.java:
> little bit adjusted NatCubic for having polygon instead of polyline
> *
> netx/net/sourceforge/jnlp/splashscreen/impls/defaultsutils/SplinesDefs.java:
> definitions of control points for NatCubicClosed splines on logo
> *
> netx/net/sourceforge/jnlp/splashscreen/impls/defaultsutils/TextWithWatterLevel.java:
> Extending ImageFontCutter is cutting from from basin-with-water like
> rendered image
>
> splasScreenTests2.diff
> 2012-03-02 Jiri Vanek <jvanek at redhat.com>
> *
> tests/jnlp_tests/signed/CacheReproducer/testcases/CacheReproducerTest.java:
> as javaws have now integrated splash, I have changed this test to lunch
> javaws -Xclearcache with -headless to skip this logo (although it is not
> fatal fr testrun itself)
> * tests/jnlp_tests/simple/AppletTest/resources/appletAutoTests.html:
> this html file is lunched during tests run in browser and stdout of
> lunched applet is examined. Is lunched with slow resources to test
> spalshscreen
> * tests/jnlp_tests/simple/AppletTest/resources/appletViewTest.html:
> this test html file is dedicated to manual lunch and let user to look
> how the applet (with slow loading) is loaded and how looks splashscreen
> in small mode and in large mode
> * tests/jnlp_tests/simple/AppletTest/testcases/AppletTestTests.java:
> test is enriched for lunching the html file with applet in browser and
> is examining output of this file. Browser must be always terminated as
> there is no way hoe to close from inside
> * tests/jnlp_tests/simple/simpletest1/resources/netxPlugin.png :
> image to let user observe that user-defined splashscreen is still
> working even when internal splashscreen is enabled
> *
> tests/jnlp_tests/simple/simpletest1/resources/simpletestCustomSplash.jnlp:
> this and all jnlp files below are just for manual lunching and for
> watching various lunches of splash screen - slow loading of resources
> and with custom splash
> *
> tests/jnlp_tests/simple/simpletest1/resources/simpletestMegaSlow.jnlp:
> slow loading of resource and jnlp
> * tests/jnlp_tests/simple/simpletest1/resources/simpletestSlow.jnlp:
> slow loading of resource
> *
> tests/jnlp_tests/simple/simpletest1/resources/simpletestSlowBrokenCustomSplash.jnlp:
> slow loading of resource with broken user's splash (our internal will be
> used)
> *
> tests/jnlp_tests/simple/simpletest1/resources/simpletestSlowSlowCustomSplash.jnlp:
> slow loading of custom splash screen and resource
> *
> tests/netx/jnlp_testsengine/net/sourceforge/jnlp/ServerAccess.java: Main
> server launcher was modified to support lunching of browser, stdout
> listteners and for slowing download of resources to provide time for
> watching splash screen
> (main) was rewritten to provide free port OR run server in-D
> specified directory on custom or default port - very useful for debuging
> reproducers
> (getIndependentInstance) can now run also on specified port and
> (or) directory
> (USED_BROWSER_COMMAND) new constant handling value of -D property
> to set browser = "used.browser.command";
> (getBrowserLocation) new method to provide specified (by
> used.browser.command -D property) or default browser location (firefox)
> (ensureServer) test is testing weather XslowXmodifier is working
> (executeBrowser) set of overloaded functions to lunch browser
> (TinyHttpdImpl) was enriched for XslowX modifier. When resource
> starts with this, is returned slowly - splited to 10 parts with 1s delay
> betwen sending each of them. Although it is throwing BrokenPipe
> exception, is working fine.
> (splitArray) new function to split array of byte to n arrays of
> bytes, which when concated do the same array
> (splitArrayTestN) set of tests for splitArray
> (ContentReader) now can also have lsteners for catching outputs n
> runtime.
> *
> tests/netx/jnlp_testsengine/net/sourceforge/jnlp/ContentReaderListener.java:
> Listener for catching chars and lines form processes outputs
> *
> tests/netx/jnlp_testsengine/net/sourceforge/jnlp/ResourcesTest.java:
> added (testListeners) to test listeners behaviour
>
>
> To much reading? To long "Patch" ? I'm terribly sorry for being so
> undisciplined to allow it to grow so much and not to commit t in smaller
> chunks. I hope it will be still review-able.
>
> The content of gif.gif - what you should see:
> LEFT RIGHT
> firefox
> applets started to be loaded
> (one small and one big)
> lunching javaws -Xclearcache -headless (no splash appears)
> lunching javaws -Xclearcache (bitmap splash appears with some
> artifact) small applet stars fine, big one detected crash and swapped
> to error screen
> lunching javaws -Xclearcache -headless (no splash appears)
> lunching javaws someUrl - lunching example jnlp which is at first loading
> very big user's spalsh, and then swap to it) clicking
> "show error" button on failed applet error screen
> Dialog with applets error message appears
> is resized, link is clicked, home is clicked, is closed
> firefox closed
> lunched loading of jnlp file with our custom (jumping through both
> screens, error during gif compositon :-/)
> resizing our default vector
> splash screen to show corner cases, the rising water is nicely to be
> seen during the resizing
>
> Best regards and good luck
> J.
>
>
>
> ------------------------------------------------------------------------
>
>
> ------------------------------------------------------------------------
>
More information about the distro-pkg-dev
mailing list