Re: [FlexJS] Working to fix a compiler bug - problem with the Problems
I believe MXML is supposed to treat more than one child in a child tag as an array. And thus, the equivalent AS code for: is: initialView = [ new MyInitialView, new MyOtherInitialView]; I understood the same, but I had thought it was supposed to be a bit smarter than that and only assume an implicit wrapper if it is assigning to an Array typed property (maybe even 'Arraylike') property. I will check the behavior using Flex 4 - I assume we want to keep whatever 'smarts' were implemented there, or at least as close as makes sense? I'll also double check the compiler error for this same scenario in Flex 4, which I did not do yet. If you code that in AS, you will get an error (I hope), and I think the compiler should report the same error for this MXML scenario, since initialView is of type IApplicationView or something like that. There might arguably be some justification for having an 'mxml version' of an error, describing things more in mxml terms, particularly for new users as they come on board and perhaps play with mxml examples first before learning actionscript, but I will check both the actionscript error in FlexJS and the mxml error in Flex 4 (assuming there is one) for this scenario and report back here for consensus. Is the equivalent of: var initialView:IApplicationView = new SimpleCSSValuesImpl(). Exactly, and in the fix, I have this currently being described as: In initializer for 'js:initialView', type org.apache.flex.core.SimpleCSSValuesImpl is not assignable to target type 'org.apache.flex.core.IApplicationView'. I will check the FlexJS actionscript error for this also, but I definitely pulled that one above as-is from Flex 4 for the same mxml problem (which is also correctly treated as a compile time error) On Fri, Mar 3, 2017 at 6:33 PM, Alex Haruiwrote: > Hi Greg, > > Thanks for digging into this. > > Without having dug into it myself, I would like to suggest returning a > type mismatch error of some sort. IIRC, the DuplicateChildTagProblem is > for the following: > > > > > > > > > > The child tag (js:initialView) is really in there twice. > > > I believe MXML is supposed to treat more than one child in a child tag as > an array. And thus, the equivalent AS code for: > > > > > > > is: > > initialView = [ new MyInitialView, new MyOtherInitialView]; > > If you code that in AS, you will get an error (I hope), and I think the > compiler should report the same error for this MXML scenario, since > initialView is of type IApplicationView or something like that. > > And same for the second scenario: > > > > > > Is the equivalent of: > > > var initialView:IApplicationView = new SimpleCSSValuesImpl(). > > My 2 cents, > -Alex > > On 3/2/17, 9:09 PM, "Greg Dove" wrote: > > >I have been looking into FLEX-35273 [1]. > > > >This is a compiler bug where it is possible to do things that don't make > >sense, like: > > > > > > > > > > > > > >or even this : > > > > > > > > > > > >Neither of the above caused compile time errors. > > > >I have a fix for both the above scenarios. > > > >For the first one, I used MXMLDuplicateChildTagProblem which seems 'close > >enough' > > > >"Child tag '${childTag}' bound to namespace '${childNamespace}' is > >already specified for element '${element}'. It will be ignored."; > > > >This renders as : > >Child tag 'MyInitialView' bound to namespace '*' is already specified for > >element 'js:initialView'. It will be ignored. > > > >^ > > > >in the first example above. The text does not feel entirely right, but > >seems close enough. "It will be ignored" sounds more like a warning (as > >opposed to an error/failure), which I think a) it should not be and b) it > >is not. > > > >For the second one, I could not find any relevant 'Problem' class (I may > >have missed one perhaps?) So I just made a new one. I don't really know > >what the rules are here. > >These Problems seem to include an error code so I am unclear what to do > >if I add a new one. i.e. it is not clear what new error code I should > >apply (or even if the error code is important for something). > >For now I just added an arbitrary errorCode = 1999 on > >MXMLBadChildTagPropertyAssignmentProblem. > > > >The new problem renders out to : > >In initializer for 'js:initialView', type > >org.apache.flex.core.SimpleCSSValuesImpl is not assignable to target type > >'org.apache.flex.core.IApplicationView'. > > > >Alex, you may be able to advise here are there any rules I need to > >follow for the CompilerProblem classes (in particular, when adding a new > >class). > > > >Assuming all is well above, I will commit this fix tomorrow. I was also > >trying to see if I could figure out how to get checkintests running, but > >I so far I did not. However the regular build tests are all fine with the > >changes I made for this. > > > > > >1. https://issues.apache.org/jira/browse/FLEX-35273 > >
Re: [FlexJS] Any gude to getting checkintests working
Excellent! Thanks for sticking with it. -Alex On 3/3/17, 11:05 PM, "Greg Dove"wrote: >That was it. Downgrading firefox worked. Thanks! > >I saw the browser launch and quick test from ant checkintests, > >I did not see it from the maven build: >mvn clean install -P build-examples -Drat.skip=true > >...then I remembered I had an unrelated issue with using >distributionTargetFolder defined from a locally set environment variable >(this is Windows, might be windows-specific) for maven > >so... for posterity, this is the maven version that works for me on >windows: > >mvn clean install -P build-examples -Drat.skip=true >-Dwebdriver.gecko.driver=%cd%/../testing/selenium/firefox/geckodriver.exe > >where 'testing' in > >testing/selenium/firefox/geckodriver.exe > >is a sibling folder to flex-asjs that I just set up arbitrarily in my >local >flexjs related directories > >So I now (finally) have the browser tests working. :) > > > > > >On Sat, Mar 4, 2017 at 7:04 PM, Alex Harui wrote: > >> >> >> On 3/3/17, 9:53 PM, "Greg Dove" wrote: >> >> >Thanks, I'm running 51.0.1, I'll see if I can downgrade that. Suffice >>to >> >say that "If you are using Ant, it should just work" was not quite >> >applicable in my case :). >> >A lof of this is very likely due to me being unfamiliar with the >>territory >> >here, though. >> >> You're definitely right that it didn't just work. Thanks for helping >>see >> where the problems might be. >> >> If you can't downgrade FF, it might be time to try upgrading Selenium. >>We >> upgrade Selenium every so often. >> >> Thanks, >> -Alex >> >>
Re: [FlexJS] Any gude to getting checkintests working
That was it. Downgrading firefox worked. Thanks! I saw the browser launch and quick test from ant checkintests, I did not see it from the maven build: mvn clean install -P build-examples -Drat.skip=true ...then I remembered I had an unrelated issue with using distributionTargetFolder defined from a locally set environment variable (this is Windows, might be windows-specific) for maven so... for posterity, this is the maven version that works for me on windows: mvn clean install -P build-examples -Drat.skip=true -Dwebdriver.gecko.driver=%cd%/../testing/selenium/firefox/geckodriver.exe where 'testing' in testing/selenium/firefox/geckodriver.exe is a sibling folder to flex-asjs that I just set up arbitrarily in my local flexjs related directories So I now (finally) have the browser tests working. :) On Sat, Mar 4, 2017 at 7:04 PM, Alex Haruiwrote: > > > On 3/3/17, 9:53 PM, "Greg Dove" wrote: > > >Thanks, I'm running 51.0.1, I'll see if I can downgrade that. Suffice to > >say that "If you are using Ant, it should just work" was not quite > >applicable in my case :). > >A lof of this is very likely due to me being unfamiliar with the territory > >here, though. > > You're definitely right that it didn't just work. Thanks for helping see > where the problems might be. > > If you can't downgrade FF, it might be time to try upgrading Selenium. We > upgrade Selenium every so often. > > Thanks, > -Alex > >
Re: [FlexJS] Any gude to getting checkintests working
On 3/3/17, 9:53 PM, "Greg Dove"wrote: >Thanks, I'm running 51.0.1, I'll see if I can downgrade that. Suffice to >say that "If you are using Ant, it should just work" was not quite >applicable in my case :). >A lof of this is very likely due to me being unfamiliar with the territory >here, though. You're definitely right that it didn't just work. Thanks for helping see where the problems might be. If you can't downgrade FF, it might be time to try upgrading Selenium. We upgrade Selenium every so often. Thanks, -Alex
Re: [FlexJS] Any gude to getting checkintests working
Thanks, I'm running 51.0.1, I'll see if I can downgrade that. Suffice to say that "If you are using Ant, it should just work" was not quite applicable in my case :). A lof of this is very likely due to me being unfamiliar with the territory here, though. On Sat, Mar 4, 2017 at 6:45 PM, Alex Haruiwrote: > What version of Firefox are you using? > > We are currently using Selenium 2.53.1. I think it likes FF 47.0.1. > > HTH, > -Alex > > On 3/3/17, 6:48 PM, "Greg Dove" wrote: > > >Thanks Chris, and Alex, > > > > > >I am still working on getting the browser test support to work via either > >maven or ant checkintests. > > > >Thus far I get a firefox launch when running ant checkintests but nothing > >displays and I eventually get > >org.openqa.selenium.firefox.NotConnectedException: Unable to connect to > >host 127.0.0.1 on port 7055 after 45000 ms. > > > >With maven I set the environment variable and also added geckodriver.exe > >containing directory to the PATH, but I don't see any browser launch from > >the build-examples. > > > >So clearly I have something wrong with my setup for Selenium/Firefox, will > >try to figure it out. > > > > > > > > > >I have not run the selenium > > > >On Sat, Mar 4, 2017 at 2:02 AM, Christofer Dutz > > > >wrote: > > > >> Hi Greg, > >> > >> With Maven it should also work. > >> > >> For the compiler: > >> If you do a “mvn clean install” it should run almost all unit-tests and > >> integration-tests. > >> If you define an Evironment variable FLASHPLAYER_DEBUGGER (Same as for > >>the > >> ANT build) pointing to the flash debug player the integration-tests with > >> Flash should work. > >> > >> For the typedefs: > >> They don’t have tests (at least in Maven) > >> > >> For the framework: > >> A simple “mvn clean install” compiles just the framework libs, but > >>doesn’t > >> build the examples. > >> If you do a “mvn clean install –P build-examples” it will also build the > >> examples (Examples will run a small unit test that tests, the build > >> artifacts are correctly created) > >> If you do a “mvn clean install –P build-distribution” it will also build > >> the distribution (but not the examples) > >> If you do a “mvn clean install –P build-examples,build-distribution” it > >> will build the libs, examples and distribution. > >> And If you also define an environment variable “webdriver.gecko.driver” > >>to > >> the selenium gecko-driver and run “mvn clean install –P build-examples” > >>the > >> build will also run some browser-tests. > >> > >> Chris > >> > >> > >> > >> > >> Am 03.03.17, 05:59 schrieb "Greg Dove" : > >> > >> There are a few things I have not yet gotten working so far, and the > >> ant checkintests is one of them. > >> > >> If I understand correctly I should be running this before any > >>commit? > >> I tried to make sense of this today, but it is not so easy (for me) > >>because > >> I am unfamiliar with this stuff. > >> > >> As I have never done this so far, is it ok if I just continue to > >> commit based on the regular ant and maven builds (with their various > >>tests) > >> completing successfully? Otherwise if I need to get this set up, is > >>there a > >> guide to doing so somewhere that I can follow? > >> > >> > >> > >> > >> > >> > >
Re: [FlexJS] Any gude to getting checkintests working
What version of Firefox are you using? We are currently using Selenium 2.53.1. I think it likes FF 47.0.1. HTH, -Alex On 3/3/17, 6:48 PM, "Greg Dove"wrote: >Thanks Chris, and Alex, > > >I am still working on getting the browser test support to work via either >maven or ant checkintests. > >Thus far I get a firefox launch when running ant checkintests but nothing >displays and I eventually get >org.openqa.selenium.firefox.NotConnectedException: Unable to connect to >host 127.0.0.1 on port 7055 after 45000 ms. > >With maven I set the environment variable and also added geckodriver.exe >containing directory to the PATH, but I don't see any browser launch from >the build-examples. > >So clearly I have something wrong with my setup for Selenium/Firefox, will >try to figure it out. > > > > >I have not run the selenium > >On Sat, Mar 4, 2017 at 2:02 AM, Christofer Dutz > >wrote: > >> Hi Greg, >> >> With Maven it should also work. >> >> For the compiler: >> If you do a “mvn clean install” it should run almost all unit-tests and >> integration-tests. >> If you define an Evironment variable FLASHPLAYER_DEBUGGER (Same as for >>the >> ANT build) pointing to the flash debug player the integration-tests with >> Flash should work. >> >> For the typedefs: >> They don’t have tests (at least in Maven) >> >> For the framework: >> A simple “mvn clean install” compiles just the framework libs, but >>doesn’t >> build the examples. >> If you do a “mvn clean install –P build-examples” it will also build the >> examples (Examples will run a small unit test that tests, the build >> artifacts are correctly created) >> If you do a “mvn clean install –P build-distribution” it will also build >> the distribution (but not the examples) >> If you do a “mvn clean install –P build-examples,build-distribution” it >> will build the libs, examples and distribution. >> And If you also define an environment variable “webdriver.gecko.driver” >>to >> the selenium gecko-driver and run “mvn clean install –P build-examples” >>the >> build will also run some browser-tests. >> >> Chris >> >> >> >> >> Am 03.03.17, 05:59 schrieb "Greg Dove" : >> >> There are a few things I have not yet gotten working so far, and the >> ant checkintests is one of them. >> >> If I understand correctly I should be running this before any >>commit? >> I tried to make sense of this today, but it is not so easy (for me) >>because >> I am unfamiliar with this stuff. >> >> As I have never done this so far, is it ok if I just continue to >> commit based on the regular ant and maven builds (with their various >>tests) >> completing successfully? Otherwise if I need to get this set up, is >>there a >> guide to doing so somewhere that I can follow? >> >> >> >> >> >>
Re: [FlexJS] Any gude to getting checkintests working
Thanks Chris, and Alex, I am still working on getting the browser test support to work via either maven or ant checkintests. Thus far I get a firefox launch when running ant checkintests but nothing displays and I eventually get org.openqa.selenium.firefox.NotConnectedException: Unable to connect to host 127.0.0.1 on port 7055 after 45000 ms. With maven I set the environment variable and also added geckodriver.exe containing directory to the PATH, but I don't see any browser launch from the build-examples. So clearly I have something wrong with my setup for Selenium/Firefox, will try to figure it out. I have not run the selenium On Sat, Mar 4, 2017 at 2:02 AM, Christofer Dutzwrote: > Hi Greg, > > With Maven it should also work. > > For the compiler: > If you do a “mvn clean install” it should run almost all unit-tests and > integration-tests. > If you define an Evironment variable FLASHPLAYER_DEBUGGER (Same as for the > ANT build) pointing to the flash debug player the integration-tests with > Flash should work. > > For the typedefs: > They don’t have tests (at least in Maven) > > For the framework: > A simple “mvn clean install” compiles just the framework libs, but doesn’t > build the examples. > If you do a “mvn clean install –P build-examples” it will also build the > examples (Examples will run a small unit test that tests, the build > artifacts are correctly created) > If you do a “mvn clean install –P build-distribution” it will also build > the distribution (but not the examples) > If you do a “mvn clean install –P build-examples,build-distribution” it > will build the libs, examples and distribution. > And If you also define an environment variable “webdriver.gecko.driver” to > the selenium gecko-driver and run “mvn clean install –P build-examples” the > build will also run some browser-tests. > > Chris > > > > > Am 03.03.17, 05:59 schrieb "Greg Dove" : > > There are a few things I have not yet gotten working so far, and the > ant checkintests is one of them. > > If I understand correctly I should be running this before any commit? > I tried to make sense of this today, but it is not so easy (for me) because > I am unfamiliar with this stuff. > > As I have never done this so far, is it ok if I just continue to > commit based on the regular ant and maven builds (with their various tests) > completing successfully? Otherwise if I need to get this set up, is there a > guide to doing so somewhere that I can follow? > > > > > >
Re: [FlexJS] Any gude to getting checkintests working
HI, > For the framework: > A simple “mvn clean install” compiles just the framework libs, but doesn’t > build the examples. > If you do a “mvn clean install –P build-examples” it will also build the > examples (Examples will run a small unit test that tests, the build artifacts > are correctly created) Just tried this: mvn clean install –P build-examples and I got: [ERROR] Unknown lifecycle phase "–P". You must specify a valid lifecycle phase or a goal in the format : or :[:]:. Available lifecycle phases are: validate, initialize, generate-sources, process-sources, generate-resources, process-resources, compile, process-classes, generate-test-sources, process-test-sources, generate-test-resources, process-test-resources, test-compile, process-test-classes, test, prepare-package, package, pre-integration-test, integration-test, post-integration-test, verify, install, deploy, pre-clean, clean, post-clean, pre-site, site, post-site, site-deploy. -> [Help 1] What am I doing wrong? This may help: mvn -version Apache Maven 3.3.9 (bb52d8502b132ec0a5a3f4c09453c07478323dc5; 2015-11-11T03:41:47+11:00) Maven home: /usr/local/Cellar/maven/3.3.9/libexec Java version: 1.8.0_73, vendor: Oracle Corporation Java home: /Library/Java/JavaVirtualMachines/jdk1.8.0_73.jdk/Contents/Home/jre Default locale: en_AU, platform encoding: UTF-8 OS name: "mac os x", version: "10.12.3", arch: "x86_64", family: "mac" Thanks, Justin
[4.16.0] # Apache Flex SDK 4.16.0 nightly build 247: Fixed
flex-sdk_release-candidate - Build #247 - Fixed Changes since last build: No changes For more information, check the console output at http://apacheflexbuild.cloudapp.net:8080/job/flex-sdk_release-candidate/247/.
[FlexJS][Maven] JS-only
Hey Maven Experts, Over in the flex-asjs branch named "dual", I am trying to get the projects in examples/native to build. The two projects that build in Ant are ButtonExample and USStatesMap. These projects do not produce a SWF, only HTML/JS/CSS. In Maven, ButtonExample now compiles cleanly, but then fails with: [ERROR] Failed to execute goal org.apache.maven.plugins:maven-install-plugin:2.5.2:install (default-install) on project ButtonExample: The packaging for this project did not assign a file to the build artifact -> [Help 1] The AppMojo seems to make the SWF the artifact, but in this case there won't be one. What is the right thing to do here? I've disabled the native folder in the pom for now. You should be able to grab the flex-falcon dual branch and flex-asjs dual branch and use Maven to reproduce this. Thanks, -Alex
Re: massive memory leak in Flex
On 3/3/17, 1:23 PM, "Jason Taylor"wrote: >sounds good, this is all stemming from a memory I discovered in our >SparkTree component which is based off >https://github.com/kachurovskiy/Spark-Tree which I'm sure a lot of people >use. In the middle of redesigning a solid tree component based of >SaturnBoy's work that is FAR more performant with a lot less code >complexity, and hopefully dosen't exhibit the memory leak issue. OK. I'm not clear if you still think there is a memory issue or just saw some aliasing from Scout. The System.totalMemory technique is how we used to test for memory leaks before there were FB Profiler and Scout and is less likely to cause aliasing. I haven't used that technique in a long time, but I think maybe I actually used code like: Var tm:int = System.totalMemory; Math.max(lastTotalMemory, tm) lastTotalMemory = tm; so you are observing the high water mark. HTH, -Alex
RE: massive memory leak in Flex
sounds good, this is all stemming from a memory I discovered in our SparkTree component which is based off https://github.com/kachurovskiy/Spark-Tree which I'm sure a lot of people use. In the middle of redesigning a solid tree component based of SaturnBoy's work that is FAR more performant with a lot less code complexity, and hopefully dosen't exhibit the memory leak issue. ~ JT -Original Message- From: Alex Harui [mailto:aha...@adobe.com] Sent: Friday, March 03, 2017 11:26 AM To: dev@flex.apache.org Subject: Re: massive memory leak in Flex What I sometimes do as add a Label to the app that displays System.totalMemory. It reach a max after several minutes. I run the app for several minutes before going to bed, write down the System.totalMemory number, then when I wake up I see if it has climbed or the app has crashed. -Alex On 3/3/17, 11:04 AM, "Jason Taylor"wrote: >player 22, windows 7 64, when scout is running it appears no objects >get released to the gc, so this appears to be a bug with scout. > >-Original Message- >From: Alex Harui [mailto:aha...@adobe.com] >Sent: Thursday, March 02, 2017 8:45 PM >To: dev@flex.apache.org >Subject: Re: massive memory leak in Flex > >So what OS and version, and what kind of player and its version are you >using? > >Scout will turn on sampling in the app so that should result in an >increase in memory as well. > >-Alex > >On 3/2/17, 4:48 PM, "Jason Taylor" wrote: > >>wonder if that might be osx / windows difference in scout? as I only >>show this leak atm when scout is ran- Scout 1.1.3.345121 on Scout >> >>-Original Message- >>From: Jason Taylor >>Sent: Thursday, March 02, 2017 4:48 PM >>To: dev@flex.apache.org >>Subject: RE: massive memory leak in Flex >> >>what the hell? Mine looks nothing like that in scout, no advanced >>telemetry >> >>-Original Message- >>From: Clint M [mailto:cmod...@gmail.com] >>Sent: Thursday, March 02, 2017 4:38 PM >>To: dev@flex.apache.org >>Subject: Re: massive memory leak in Flex >> >>it didn't leak using advanced-telemetry either… it didn't leak with a >>custom renderer with curly binding either… probably because spark >>lists recycle item renderers by default >> >>On Thu, Mar 2, 2017 at 4:29 PM, Clint M wrote: >> >>> here's a video of it running in scout >>> http://www.screencast.com/t/S5SHGLiskRN >>> >>> didn't leak… are you using advanced-telemetry ? >>> >>> On Thu, Mar 2, 2017 at 4:24 PM, Jason Taylor wrote: >>> Yeah, it looks like it may be an artifact of scout somehow -Original Message- From: Clint M [mailto:cmod...@gmail.com] Sent: Thursday, March 02, 2017 4:24 PM To: dev@flex.apache.org Subject: Re: massive memory leak in Flex Can't repro with the code in the original post… here's a video of it running for 60 seconds. http://www.screencast.com/t/mPFMIdKQMK On Thu, Mar 2, 2017 at 4:17 PM, Jason Taylor wrote: > Yeap task manager keeps climbing as well > > -Original Message- > From: Alex Harui [mailto:aha...@adobe.com] > Sent: Thursday, March 02, 2017 4:07 PM > To: dev@flex.apache.org > Subject: Re: massive memory leak in Flex > > Scout only reports certain player memory zones. Hopefully > TaskManager has a higher number? If you got past 1GB then it probably isn't capped. > > -Alex > > On 3/2/17, 3:56 PM, "Jason Taylor" wrote: > > >Tried 4.13, 14, and 15, but I have a weird feeling scout is a > >dirty Liar > > > >-Original Message- > >From: Alex Harui [mailto:aha...@adobe.com] > >Sent: Thursday, March 02, 2017 3:54 PM > >To: dev@flex.apache.org > >Subject: Re: massive memory leak in Flex > > > >OK Thanks. Sounds like Justin may be digging into it, so I'll > >wait until he reports in. Which version of Apache Flex are you using? > >Have you tried Adobe Flex 4.6? > > > >-Alex > > > >On 3/2/17, 3:50 PM, "Jason Taylor" wrote: > > > >>This is all the code needed to cause it, and no it does not > >>appear to stop, though I will need to run it for awhile to blow > >>all my memory in this comp. > >> > >> > >> > >>http://ns.adobe.com/mxml/2009; > >>xmlns:s="library://ns.adobe.com/flex/spark" > >>creationComplete="OnCreationComplete(event)"> > >> > >> > >> > >> > >> > >> > >>-Original Message- > >>From: Alex Harui [mailto:aha...@adobe.com] > >>Sent: Thursday, March 02, 2017 3:47 PM > >>To: dev@flex.apache.org > >>Subject: Re: massive memory leak in Flex > >> > >>Jason, we aren't seeing all of the code. Can you put all of > >>the code
Re: [FlexJS, MDL] Add example "MDLDynamicTabsExample"
Hi Olaf, Glad to hear that! :) Thank you! - Apache Flex PMC piotrzarzyck...@gmail.com -- View this message in context: http://apache-flex-development.247.n4.nabble.com/FlexJS-MDL-Add-example-MDLDynamicTabsExample-tp59888p60051.html Sent from the Apache Flex Development mailing list archive at Nabble.com.
Re: [FlexJS, MDL] Add example "MDLDynamicTabsExample"
piotrz wrote > Hi All, > > I just added "MDLDynamicTabsExample" example which illustrate how to > create "Tab" dynamically in our MDL library. [1] > > [1] > https://github.com/apache/flex-asjs/tree/develop/examples/flexjs/MDLDynamicTabsExample > > Piotr Hi Piotrz, I've just checked it out and it works like a charme :-) Thank you for working on this! Olaf -- View this message in context: http://apache-flex-development.247.n4.nabble.com/FlexJS-MDL-Add-example-MDLDynamicTabsExample-tp59888p60050.html Sent from the Apache Flex Development mailing list archive at Nabble.com.
Re: massive memory leak in Flex
What I sometimes do as add a Label to the app that displays System.totalMemory. It reach a max after several minutes. I run the app for several minutes before going to bed, write down the System.totalMemory number, then when I wake up I see if it has climbed or the app has crashed. -Alex On 3/3/17, 11:04 AM, "Jason Taylor"wrote: >player 22, windows 7 64, when scout is running it appears no objects get >released to the gc, so this appears to be a bug with scout. > >-Original Message- >From: Alex Harui [mailto:aha...@adobe.com] >Sent: Thursday, March 02, 2017 8:45 PM >To: dev@flex.apache.org >Subject: Re: massive memory leak in Flex > >So what OS and version, and what kind of player and its version are you >using? > >Scout will turn on sampling in the app so that should result in an >increase in memory as well. > >-Alex > >On 3/2/17, 4:48 PM, "Jason Taylor" wrote: > >>wonder if that might be osx / windows difference in scout? as I only >>show this leak atm when scout is ran- Scout 1.1.3.345121 on Scout >> >>-Original Message- >>From: Jason Taylor >>Sent: Thursday, March 02, 2017 4:48 PM >>To: dev@flex.apache.org >>Subject: RE: massive memory leak in Flex >> >>what the hell? Mine looks nothing like that in scout, no advanced >>telemetry >> >>-Original Message- >>From: Clint M [mailto:cmod...@gmail.com] >>Sent: Thursday, March 02, 2017 4:38 PM >>To: dev@flex.apache.org >>Subject: Re: massive memory leak in Flex >> >>it didn't leak using advanced-telemetry either… it didn't leak with a >>custom renderer with curly binding either… probably because spark lists >>recycle item renderers by default >> >>On Thu, Mar 2, 2017 at 4:29 PM, Clint M wrote: >> >>> here's a video of it running in scout >>> http://www.screencast.com/t/S5SHGLiskRN >>> >>> didn't leak… are you using advanced-telemetry ? >>> >>> On Thu, Mar 2, 2017 at 4:24 PM, Jason Taylor wrote: >>> Yeah, it looks like it may be an artifact of scout somehow -Original Message- From: Clint M [mailto:cmod...@gmail.com] Sent: Thursday, March 02, 2017 4:24 PM To: dev@flex.apache.org Subject: Re: massive memory leak in Flex Can't repro with the code in the original post… here's a video of it running for 60 seconds. http://www.screencast.com/t/mPFMIdKQMK On Thu, Mar 2, 2017 at 4:17 PM, Jason Taylor wrote: > Yeap task manager keeps climbing as well > > -Original Message- > From: Alex Harui [mailto:aha...@adobe.com] > Sent: Thursday, March 02, 2017 4:07 PM > To: dev@flex.apache.org > Subject: Re: massive memory leak in Flex > > Scout only reports certain player memory zones. Hopefully > TaskManager has a higher number? If you got past 1GB then it probably isn't capped. > > -Alex > > On 3/2/17, 3:56 PM, "Jason Taylor" wrote: > > >Tried 4.13, 14, and 15, but I have a weird feeling scout is a > >dirty Liar > > > >-Original Message- > >From: Alex Harui [mailto:aha...@adobe.com] > >Sent: Thursday, March 02, 2017 3:54 PM > >To: dev@flex.apache.org > >Subject: Re: massive memory leak in Flex > > > >OK Thanks. Sounds like Justin may be digging into it, so I'll > >wait until he reports in. Which version of Apache Flex are you using? > >Have you tried Adobe Flex 4.6? > > > >-Alex > > > >On 3/2/17, 3:50 PM, "Jason Taylor" wrote: > > > >>This is all the code needed to cause it, and no it does not > >>appear to stop, though I will need to run it for awhile to blow > >>all my memory in this comp. > >> > >> > >> > >>http://ns.adobe.com/mxml/2009; > >>xmlns:s="library://ns.adobe.com/flex/spark" > >>creationComplete="OnCreationComplete(event)"> > >> > >> > >> > >> > >> > >> > >>-Original Message- > >>From: Alex Harui [mailto:aha...@adobe.com] > >>Sent: Thursday, March 02, 2017 3:47 PM > >>To: dev@flex.apache.org > >>Subject: Re: massive memory leak in Flex > >> > >>Jason, we aren't seeing all of the code. Can you put all of the > >>code in a JIRA issue? > >> > >>There is a notion of outrunning GC. And also memory fragmentation. > >>How high does memory go? Does it cap out or will it blow up for > >>lack of memory at some point? > >> > >>Thanks, > >>-Alex > >> > >>On 3/2/17, 3:39 PM, "Jason Taylor" wrote: > >> > >>>yeah you can remove that, right now mine is just > >>> > >>> > >>>http://ns.adobe.com/mxml/2009; > >>>xmlns:s="library://ns.adobe.com/flex/spark" > >>>xmlns:Core="CoreClasses.*">
RE: massive memory leak in Flex
player 22, windows 7 64, when scout is running it appears no objects get released to the gc, so this appears to be a bug with scout. -Original Message- From: Alex Harui [mailto:aha...@adobe.com] Sent: Thursday, March 02, 2017 8:45 PM To: dev@flex.apache.org Subject: Re: massive memory leak in Flex So what OS and version, and what kind of player and its version are you using? Scout will turn on sampling in the app so that should result in an increase in memory as well. -Alex On 3/2/17, 4:48 PM, "Jason Taylor"wrote: >wonder if that might be osx / windows difference in scout? as I only >show this leak atm when scout is ran- Scout 1.1.3.345121 on Scout > >-Original Message- >From: Jason Taylor >Sent: Thursday, March 02, 2017 4:48 PM >To: dev@flex.apache.org >Subject: RE: massive memory leak in Flex > >what the hell? Mine looks nothing like that in scout, no advanced >telemetry > >-Original Message- >From: Clint M [mailto:cmod...@gmail.com] >Sent: Thursday, March 02, 2017 4:38 PM >To: dev@flex.apache.org >Subject: Re: massive memory leak in Flex > >it didn't leak using advanced-telemetry either… it didn't leak with a >custom renderer with curly binding either… probably because spark lists >recycle item renderers by default > >On Thu, Mar 2, 2017 at 4:29 PM, Clint M wrote: > >> here's a video of it running in scout >> http://www.screencast.com/t/S5SHGLiskRN >> >> didn't leak… are you using advanced-telemetry ? >> >> On Thu, Mar 2, 2017 at 4:24 PM, Jason Taylor wrote: >> >>> Yeah, it looks like it may be an artifact of scout somehow >>> >>> -Original Message- >>> From: Clint M [mailto:cmod...@gmail.com] >>> Sent: Thursday, March 02, 2017 4:24 PM >>> To: dev@flex.apache.org >>> Subject: Re: massive memory leak in Flex >>> >>> Can't repro with the code in the original post… here's a video of it >>> running for 60 seconds. >>> >>> http://www.screencast.com/t/mPFMIdKQMK >>> >>> On Thu, Mar 2, 2017 at 4:17 PM, Jason Taylor wrote: >>> >>> > Yeap task manager keeps climbing as well >>> > >>> > -Original Message- >>> > From: Alex Harui [mailto:aha...@adobe.com] >>> > Sent: Thursday, March 02, 2017 4:07 PM >>> > To: dev@flex.apache.org >>> > Subject: Re: massive memory leak in Flex >>> > >>> > Scout only reports certain player memory zones. Hopefully >>> > TaskManager has a higher number? If you got past 1GB then it >>>probably isn't capped. >>> > >>> > -Alex >>> > >>> > On 3/2/17, 3:56 PM, "Jason Taylor" wrote: >>> > >>> > >Tried 4.13, 14, and 15, but I have a weird feeling scout is a >>> > >dirty Liar >>> > > >>> > >-Original Message- >>> > >From: Alex Harui [mailto:aha...@adobe.com] >>> > >Sent: Thursday, March 02, 2017 3:54 PM >>> > >To: dev@flex.apache.org >>> > >Subject: Re: massive memory leak in Flex >>> > > >>> > >OK Thanks. Sounds like Justin may be digging into it, so I'll >>> > >wait until he reports in. Which version of Apache Flex are you >>>using? >>> > >Have you tried Adobe Flex 4.6? >>> > > >>> > >-Alex >>> > > >>> > >On 3/2/17, 3:50 PM, "Jason Taylor" wrote: >>> > > >>> > >>This is all the code needed to cause it, and no it does not >>> > >>appear to stop, though I will need to run it for awhile to blow >>> > >>all my memory in this comp. >>> > >> >>> > >> >>> > >> >>> > >>http://ns.adobe.com/mxml/2009; >>> > >>xmlns:s="library://ns.adobe.com/flex/spark" >>> > >>creationComplete="OnCreationComplete(event)"> >>> > >> >>> > >> >>> > >> >>> > >> >>> > >> >>> > >> >>> > >>-Original Message- >>> > >>From: Alex Harui [mailto:aha...@adobe.com] >>> > >>Sent: Thursday, March 02, 2017 3:47 PM >>> > >>To: dev@flex.apache.org >>> > >>Subject: Re: massive memory leak in Flex >>> > >> >>> > >>Jason, we aren't seeing all of the code. Can you put all of the >>> > >>code in a JIRA issue? >>> > >> >>> > >>There is a notion of outrunning GC. And also memory fragmentation. >>> > >>How high does memory go? Does it cap out or will it blow up for >>> > >>lack of memory at some point? >>> > >> >>> > >>Thanks, >>> > >>-Alex >>> > >> >>> > >>On 3/2/17, 3:39 PM, "Jason Taylor" wrote: >>> > >> >>> > >>>yeah you can remove that, right now mine is just >>> > >>> >>> > >>> >>> > >>>http://ns.adobe.com/mxml/2009; >>> > >>>xmlns:s="library://ns.adobe.com/flex/spark" >>> > >>>xmlns:Core="CoreClasses.*"> >>> > >>> >>> > >>> >>> > >>> >>> > >>>where MyLabel is a copy of s:label so I can make changes and >>> > >>>try to figure out where the hell this is coming from, appears >>> > >>>to be deep in TLF >>> > >>> >>> > >>>-Original Message- >>> > >>>From: Justin Mclean [mailto:jus...@classsoftware.com] >>> > >>>Sent: Thursday, March 02, 2017 3:38 PM >>> > >>>To: dev@flex.apache.org >>> > >>>Subject: Re: massive memory leak in Flex >>> > >>> >>> > >>>Hi, >>> > >>> >>> >
Re: [FlexJS] Coordinate Space
Hi, I think I understand this now. When you set x, y on a component, you will be setting the left and top styles on the HTML element. Setting these properties will no longer set the position style (which will remain unchanged it has been set). The Container classes will no longer have a default layout. This means you can use Container and place items in there and style or layout everything using CSS if you want to go that route. Some of that may not be supported by the SWF version, however. At least not right away. The BasicLayout will set the container's position to "relative" if it is not already set. Then it will go through the children and set their position styles to "absolute". This way the left and top (as well as bottom and right) styles will work. Without setting a position of "absolute" those styles will not locate the components (using a position of "relative" means that items will be placed in relation to each other, which you can now do if you do not use the BasicLayout and set position to "relative" yourself). NOTE: unlike normal HTML/CSS, FlexJS containers are sized by their contents: if you do not give a container a size, the layout will determine the size of that container. In HTML/CSS, you would probably see the container/div span the entire page by default or have a zero height by default depending on the placement of the children. You need to be aware of how margin affects placement. When using absolute position along with margin, the item will be offset by the margin value(s). For example, if you set left:10 (or item.x = 10) and add a margin-left:10, then item's content position will be 20. This will be the value the x property returns. If you are migrating from Flex 4, just leave margin settings out. Also, for BasicLayout (aka, a container with position:relative|absolute and children with position:absolute), container's padding has no effect. That is, a container with padding:10 will not change the location of any of the children placed with BasicLayout. The padding style will affect other types of layouts. For the Flexbox layouts, as long as you do not set the left, top (x, y) style (or property) the items will be placed correctly. Pixel and percentage dimensions will be honored as well. When you examine the HTML DOM, you should not see any position or placement styles on the items in a Flexbox type layout. Essentially, you should be able to do a lot more of the layout in CSS if you are targeting HTML/CSS/JS as your main platform. The layout beads are helpful to do programmatic placement and are how the SWF version will know what to do. Finally, there will probably be some new container components. Not sure what just yet, but we need to have very light ones and others that handle scrolling and chrome. Stay tuned. I now need to make these changes, which I will do in a separate branch. Thank you for your input and suggestions. I think this will solve our layout issues once complete. Regards, Peter On 3/2/17, 9:37 AM, "Peter Ent"wrote: >We still have to have FlexJS work on both JS and SWF sides with some >compatibility. We could do this: > >x,y sets "native" values. Reading them back on SWF vs JS might yield >different results if padding, border, and margins are set on the parent >element. If you don't want your coordinates messed with, then nest your >containers and apply padding, border, background, etc to the outer >container. And definitely do not use margins. > >If you want to use CSS to position items, use the top, left, bottom, right >style properties (and margin on the children). On the SWF side, these >styles will look at the parent's padding and border values and position >the elements accordingly (it will use x, y). If you intend to read the >values back, they will not necessarily be what you set since the values >are determined by the parent's padding and border as well as the child's >own margin values. > >You need to specify layouts inside containers so that the SWF side knows >what to do. If you don't intend on using the SWF output, you can just set >the style on the container and let HTML/CSS/JS take care of it. If you >supply a layout bead, the bead will set the display, positioning, and >other styles on the container as necessary and may even impose styles on >the children for the JS side. The SWF side is purely programmatic to mimic >the JS side and it will be as close as possible but may need to be >multiple passes. > >Scrolling provides more challenges for the SWF side as nested containers >need to be used with the outer container used to mask the inner which is >then repositioned to simulate scrolling. When you want scrolling, use >ScrollableContainer. This class simply sets overflow:auto on the JS side. >If you do not want or care about the SWF output, then just set the >overflow style on the container. > >‹peter > >On 3/1/17, 4:12 PM, "Justin Mclean" wrote: > >>Hi, >> >>> I think we have confusion over
Re: [MDL Dialogs] was: Re: [FlexJS] MDL - Dynamic Child Problem
I have found in MDL library that process of upgrade won't occur if component is already upgraded. In this case I could get rid off this isDynamic property and just add upgrade ability as default. What do you think Carlos ? Piotr - Apache Flex PMC piotrzarzyck...@gmail.com -- View this message in context: http://apache-flex-development.247.n4.nabble.com/FlexJS-MDL-Dynamic-Child-Problem-tp59595p60046.html Sent from the Apache Flex Development mailing list archive at Nabble.com.
Re: [MDL Dialogs] was: Re: [FlexJS] MDL - Dynamic Child Problem
In order to refresh dynamically created element in some cases I have to upgrade also subcomponent. In each scenario it can be something else. I've created UpgradeElement which can upgrade only simple component. In case of Button this will be working: var btn:Button = new Button(); btn.addBead(new UpgradeElement); But in case of other components like TextField it won't. So I've decided add isDynamic property which add ability to refresh on demand - instead doing this every time. Piotr - Apache Flex PMC piotrzarzyck...@gmail.com -- View this message in context: http://apache-flex-development.247.n4.nabble.com/FlexJS-MDL-Dynamic-Child-Problem-tp59595p60045.html Sent from the Apache Flex Development mailing list archive at Nabble.com.
Re: [MDL Dialogs] was: Re: [FlexJS] MDL - Dynamic Child Problem
On 3/3/17, 8:00 AM, "piotrz"wrote: >I didn't think about such scenario, so it look like we should expose >isDynamic also to MXML. That might work, although it can't be a constructor parameter. MXML doesn't allow constructor parameters. I would rather explore why whatever isDynamic does can't be built into the lifecycle so no isDynamic flag is ever needed. Are we just trying to detect if addElement is being called after the onLoad event? Thanks, -Alex
Re: [MDL Dialogs] was: Re: [FlexJS] MDL - Dynamic Child Problem
I didn't think about such scenario, so it look like we should expose isDynamic also to MXML. Piotr - Apache Flex PMC piotrzarzyck...@gmail.com -- View this message in context: http://apache-flex-development.247.n4.nabble.com/FlexJS-MDL-Dynamic-Child-Problem-tp59595p60043.html Sent from the Apache Flex Development mailing list archive at Nabble.com.
Re: [FlexJS] Any gude to getting checkintests working
And if you create a feature-branch “feature/{mybranchname}” in all three repos, Jenkins will automatically create a job to build that branch (currently it doesn’t work if you do this only in one of the repos) Chris Am 03.03.17, 05:59 schrieb "Greg Dove": There are a few things I have not yet gotten working so far, and the ant checkintests is one of them. If I understand correctly I should be running this before any commit? I tried to make sense of this today, but it is not so easy (for me) because I am unfamiliar with this stuff. As I have never done this so far, is it ok if I just continue to commit based on the regular ant and maven builds (with their various tests) completing successfully? Otherwise if I need to get this set up, is there a guide to doing so somewhere that I can follow?
Re: [FlexJS] Any gude to getting checkintests working
Hi Greg, With Maven it should also work. For the compiler: If you do a “mvn clean install” it should run almost all unit-tests and integration-tests. If you define an Evironment variable FLASHPLAYER_DEBUGGER (Same as for the ANT build) pointing to the flash debug player the integration-tests with Flash should work. For the typedefs: They don’t have tests (at least in Maven) For the framework: A simple “mvn clean install” compiles just the framework libs, but doesn’t build the examples. If you do a “mvn clean install –P build-examples” it will also build the examples (Examples will run a small unit test that tests, the build artifacts are correctly created) If you do a “mvn clean install –P build-distribution” it will also build the distribution (but not the examples) If you do a “mvn clean install –P build-examples,build-distribution” it will build the libs, examples and distribution. And If you also define an environment variable “webdriver.gecko.driver” to the selenium gecko-driver and run “mvn clean install –P build-examples” the build will also run some browser-tests. Chris Am 03.03.17, 05:59 schrieb "Greg Dove": There are a few things I have not yet gotten working so far, and the ant checkintests is one of them. If I understand correctly I should be running this before any commit? I tried to make sense of this today, but it is not so easy (for me) because I am unfamiliar with this stuff. As I have never done this so far, is it ok if I just continue to commit based on the regular ant and maven builds (with their various tests) completing successfully? Otherwise if I need to get this set up, is there a guide to doing so somewhere that I can follow?
Re: [MDL Dialogs] was: Re: [FlexJS] MDL - Dynamic Child Problem
Yeah. Debugging these things are not fun. Chrome also allows you to set breakpoints on DOM elements. I’ve found this useful in breaking at a specific place in the lifecycle to see what’s going on. > On Mar 3, 2017, at 7:24 AM, Alex Haruiwrote: > > > > On 3/2/17, 8:51 PM, "piotrz" wrote: > >> I will try to fight with this. The problem is that in this case I don't >> have >> any exception in the console. :) > > Yuk. I think I would next examine the DOM. Compare it against the > js-debug version. If DOM objects are missing in js-release, then I would > set breakpoints where they are created. > > HTH, > -Alex >
Re: [FlexJS] Error with Alert.show
Thanks Greg! > On Mar 3, 2017, at 9:48 AM, Greg Dovewrote: > > Quick update here, for anyone who may come across this thread: the > Reflection data bug mentioned earlier in the thread is fixed in recent > builds. > > On Sat, Feb 18, 2017 at 8:50 PM, Alex Harui wrote: > >> >> >> On 2/17/17, 7:50 PM, "Justin Mclean" wrote: >> >>> Hi, >>> >>> Currently when using Alert we’re getting this when compiling: >>> >>> SEVERE: .../javascript/bin/js-debug/org/apache/flex/html/Alert.js:254: >>> ERROR - Object literal contains illegal duplicate key "show", disallowed >>> in ES5 strict mode >>> 'show': { type: 'void', declaredBy: 'org.apache.flex.html.Alert', >>> parameters: function () { return [ { index: 1, type: 'Object', optional: >>> false } ]; }} >>> ^^ >> >> That's a bug in the Reflection data. For I worked around it in >> SimpleAlert by renaming show to showAlert so the static show calls >> showAlert. Looks like I should have applied the same hack to Alert. So >> maybe try that same change to get past this error. >> >> HTH, >> -Alex >> >>