Alex-
It's actually mainly just the compc target-player in the build.xml of the CI
project. So the issue would only appear if trying to launch the a flash or air
instance to run tests against.
Cyril actually asked me this question (as noted about 5 years ago) and I
completely missed it until t
I could be wrong but I suspect the core of flexunit doesn't have sdk
dependencies. The CI functions can run without a gui.
>Correct. The FlexUnit core is just AS but there are hooks for the GUI. Its
>honestly something I wanted to refactor more. Right now it's possible to build
>a version with
Question for the group:
The FlexUnit dependencies could be a lot cleaner but right now we have static
classes that reach out to try to figure out if it's an AS only application or a
Flex application among other things. The reason we do this is that Flash
Builder has a wrapper for the FlexUnit c
get this correct: With "we" you are refering to you and the company you work
for? Would It be possible to rearange things? Sort of having a nice and tidy
core library and have one fb-compatability lib to make FB work with FlexUnit?
If you could provide a little more details, p
>- FlexUnit (If none else wants to)
Chris,
I can't commit to doing the talk but, if you or anyone else wants to do so, I
would be happy to provide you with any/all of the FlexUnit slide decks and
talks I have done in the past as well as content around a bunch of really
useful but under-documen
>It looks to me like the intention of the person who originally created the new
>ArrayList (Michael A. Labriola) was for it to be tested with the rest of the
>framework and then to replace the original ArrayList:
Nope, I don't think that was me. I was just changing the itemUp
In the FlexUnit repo right now there is only a master.
I am creating a develop branch presently and pushing some changes we made to
the flexunit tasks to support a URL instead of just a local file when executing
tests.
We should just make sure this gets integrated into master before the next
>Great! .. I pushed my changes always to "makeFlexUnitApacheReady". Shall I
>merge those changes also to develop branch and continue to work on this one?
I think that's a good idea and we can keep merging back to develop. I always
work in my own local branches in either case.
I have a bunch of
>Our FlexUnit tests have started to fail lately. The JVM on the build server
>was updated... or at least changed not sure which.
>The error we are getting is:
>[flexunit] java.lang.NoSuchMethodError: java.lang.String.isEmpty()
>I'm pretty sure we have using the latest version of FlexUnit, can you
>It was 6 now it's 7 still failing :-(
Okay, the only thing that makes sense though is a mismatch. Possible there are
two JDKs on the box, one that is in your JAVA_HOME but another being used for
the compilation? I ask because this error is fairly well documented in the Java
community for other
>Does flexUnit plan on making a mocking framework built in? The documentation
>on Mockolate and other frameworks seems to have just died mid >devleopment,
>and the code doesn't match the docs or the forums anymore. :( brought to you
>by the letters A, V, and I and the number 47
FlexUnit belongs
>Are any of the mocking frameworks being donated to apache then? The save work
>great, it's just documentation and getting the stack working together >that I
>feel could be improved.
I do not believe so. You can contact the authors about donations if others are
interested as well. It is not my
>I will check the patches and see if any of the names cross-reference with
>those who had signed agreements, but I highly doubt it or their patches would
>have likely already been integrated.
None have agreements. I can post a comment on each github issue asking them to
approve the integration
>FWIW, Adobe did not need to contact any of its prior patch providers because
>the terms and conditions of providing the patch gave Adobe ownership and
>therefore the right to donate.
We did have an agreement, in fact the Adobe one since FlexUnit .9 was under the
Adobe banner. This had to be si
>Okay. So the tag is wrong. FB is creating a spark application as it would
>create an mx one I guess?
It's not that it's wrong, you can actually use layout as a property, even in
the tag context to specify a layout class.
It's just that a string isn't valid.
Mike
>"He may be also able to provide a more recent version of Mockolate."
This is good. FlexUnit may not be built against the latest right now as I
believe there were some method signature changes, so it's possible upgrading to
the latest will cause some minor library changes. I am also going to com
How should we specify bugs for things like FlexUnit when entering them in JIRA
to keep things a bit separate.
Maybe there is something obvious I missed... if not, ideas?
Mike
>Maybe an additional JIRA Component named FlexUnit would be helpful.
I think that's the right idea. Else you are immediately tying things to a
version and that's not the point of reporting a bug
>IIRC, it would require a player change. If you dispatch an event with the
>target already >filled out, the player clones a new instance and dispatches
>that instead.
And by spec, target is immutable so you would be going against specification as
well.
Mike
>Oh - I suppose that as a Flex Developer I should have known to connect the
>dots. In my opinion though, the "stand - alone" Open Spoon Foundation's
>website comes off as looking >abandon. So if that is the case - I would
>recommend asking them / us to take it down.
Jack,
You're not missing
>Or even longer?
I just renewed the domain for another year about a month ago, so we will
forward it until it expires and I will bring it up on the list before its gone
in case we want to renew again, etc.
Mike
>Phew. Worst case, the infallible WaybackMachine comes to the rescue (for
>content, anyway)...
>http://web.archive.org/web/20120819004707/http://www.spoon.as/ecosystem/
We still have all of the content from the site and, if anyone here is
interested, they can have it.
Mike
>this is a new feature request to have timeless date for Flash Player and
>blazeDS.
> can you please join the vote to make this high priority for Adobe to
> implement it?
Several years ago, (June of 2011) Adobe did at least a test implementation of
such a feature (although I was not particularl
A while back I asked if I could take down the FlexUnit repo and website.
At the time I was asked to wait until progress had been made here. Any chance
we are ready now as I am receiving patches on the FlexUnit github account and
then need to ask the people who already donated their time to now r
>This topic seems to come up repeatedly. I don't know what anyone on the
>runtime team tried or why Labriola didn't like it, and I'm no expert on the
>topic, but maybe we can collect some information in this thread that will help
>create a solution.
They created a SimpleDate class which extend
>Besides of the "Date without time" issue, there is also the issue handling
>regular Date+Time in different TZ.
That's a lot more difficult, especially, as you mention because of DST as it
requires data tables to make those decisions historically.
I think the only way you can achieve this is c
>What about them? For Doug and Mike, was it really easier to do custom AMF and
>timezone math vs just creating a custom DateWithoutTime class and transfer the
>Date portion to/from the Date values in the components?
It wasn't an OR for us but an AND.
In some cases where we couldn't use a compl
>I guess the first thing to sort out would be who actually wrote the code in
>that pull request (it's unclear to me) and have they all signed an ICLA?
They have not yet. That was actually the origin of my previous comment on
wondering if I can take the github repo and site offline and start redi
>If the patch is small, they should just submit the patch in a JIRA and an ICLA
>is not required.
It will eventually come in via jira but there are other hurdles to pass first.
It is using work from others, likely binaries and a ton of other issues that
all need to be resolved before the patch
>for "historical" dates we've used calls back to the server (coldfusion) or
>during data retrieval where we could access the whole tz database. but the
>issue then became what's >"historical"? which ended up as "whatever the client
>said it was", not the best approach.
Problem with the server c
>I will have a look at icu4j.
Icu4j (actually icu4c) is what we used as the basis for all of our work too. It
has a ton of great stuff in there. In particular we were also dealing with
non-gregorian calendaring and they have some reasonably helpful code throughout.
We ended up taking the transi
>Before we switch to it, we have to be absolutely sure of when it was first
>supported. I wouldn't want to break our Flash version compatibility range.
>But otherwise, I have no objections if you want to switch to it.
Internally describeType now calls describeTypeJSON and then serializes the
>Before we switch to it, we have to be absolutely sure of when it was first
>supported. I wouldn't want to break our Flash version compatibility range.
>But otherwise, I have no objections if you want to switch to it.
If it helps, in my code, I am checking for FP10.1 or higher which is when I
>i have a very vague recollection that somebody partially ported icu4j/icu4c to
AS3 (maybe just the calenders). wondering what happened to it?
You may be recalling my work. Not sure.
>could the flasCC or whatever its called now be used to wrapper/import the
>icu4c bits???
Tried that first, it p
>Shumway is a plugin.
It's not a plugin. You probably need to take a look at the project and the code.
Yes, they say in that article is a browser extension, but effectively it is a
core JS library that acts as a Flash Player virtual machine but executing
inside of the JavaScript VM. I am not s
If Shumway ever gets completed and officially supported, it could extend swf
format lifetime but it would also introduce a platform fragmentation.
>Think about it this way. Right now Flash Player is implemented in a
>combination of low-level programming languages and compiled for native code
>s
>Search for "Puffin Web Browser" on the App Store.
>https://itunes.apple.com/fr/app/puffin-web-browser-free/id472937654?mt=8
>It has support for Flash.
>But it's so slow, even on simple flash pages...
The real key to this is in the detail. This is from the apple licensing
agreement:
"3.3.2 An
>wBox=vbox.rawChildren.getChildByName("lstControl");
>lstControl= new List();
>vbox.addChild(lstControl);
This probably belongs on the user list. So, continued the discussion there.
getChildByName() works based on the child's name, not the name of a variable
that you used to instantiate and add
>No idea. I was trying to think of any danger of manipulating the returned
>string if it is the original and not always a copy, but I can't think of
>anything off-hand.
>Is it much faster to add the check and return the original?
Aren't AS strings immutable anyway? So, the method got a copy t
>If class B extends class A and class A implements interface IA, does class B
>also implement IA? In other words, in this scenario, does (B is IA) return
>true?
Yep
>And the label displays "!Hi there" is that expected?
Yes, that is correct in this case. The Hi There are Latin characters which
carry information in Unicode to indicate they should be displayed LTR. However,
the punctuation mark is a weak character (can be used in both and doesn't have
this ty
>>And the label displays "!Hi there" is that expected?
>Yes, that is correct in this case. The Hi There are Latin characters which
>carry information in Unicode to indicate they should be displayed LTR.
>However, the punctuation mark is a weak character (can be used in both >and
>doesn't have t
>I know "in" can be expensive but so is try catch. I've run each though scout
>and it looks like the in operator is orders of magnitude faster and caused no
>garbage collection both when the field exists and when it >doesn't. Anyone
>have any insight/something to add?
Any chance data can be nu
>And the same code is called when using an index loop (which works as expected)
>so far as I can tell it's a "side effect" of using "in".
It is. The issue exists with XMLList as well and also Dictionary (although less
reproducible with the latter)
In all cases using the in while modifying prop
>And the same code is called when using an index loop (which works as expected)
>so far as I can tell it's a "side effect" of using "in".
Let me take some of that back, I was describing a different issue. I think this
might be related to parenting issues within XMLList. It can produce very
unex
>Thanks for all that. The code isn't adding anything to the original list ,
>it's looping though the nodes of a list then and adding them to a new
>collection.
>So why does using a loop index work, but using in doesn't?
I will write it all up tonight if possible. I can provide example of each st
>I will write it all up tonight if possible. I can provide example of each step
>and what's going on.
In XMLListAdapter, addItemAt() we see this after the first addItem()
source[0] = length > 0 ? item + source[0] : item;
this effectively sets source[0] to the unchanged item... and the parent do
>mainXML.item.length() which shows the same problem.
Actually, It shows a worse problem... and infinite loop, cause length() keeps
growing so the loop will never terminate. If you save the invariant to a local
variable first, obviously it works better... but the XMLList still grows.
Mike
Here is my proposed fix in XMLListAdapter:: addItemAt() :
...
if (length > 0) {
var localLength:uint = source.length();
//Adjust all indexes by 1
for ( var i:uint = localLength; i>index; i-- ) {
source[i] = source[ i - 1 ];
}
}
source[ index ] = item;
...
We may be able to optim
>Here is the thing though -- it's not a "us vs. them" type of situation. We
>are all Apache (including the community). The trademark group within the
>Apache Foundation is there to protect us. Please see it that way. >We live
>in a litigious world and we all need to realize that.
I don't wa
>Thanks for the contribution - perhaps we should make you a committer? :-)
Let's not rush into anything :)
Mike
>Anyone have a clue where skinParts is getting populated from? Its not
>overridden in a derived class, so really not sure how/whats populating it?
It's compiler magic. The compiler creates a static variable called _skinParts
on each class with the skin parts for the class in it. It then override
>so we don't have do this sort of "mystery meat" guesswork whilst debugging.
Totally agree.
>I know you can find what code the compiler outputs by using the
>"keep-generated-actionscript" and doing a diff with the uncompiled code, but
>that's just such a pain to look up something which should t
>Done for that method - any other methods need the same treatment?
UIComponent::initialize() is probably the best example. Its overwritten and
mangled quite successfully by the compiler.
Mike
Does apache have any interest in taking over FlexUnit.org? I brought it up a
while back and I don't think I heard a response.
I get pinged a few times a week on people still looking for the docs at this
URL?
Mike
>Maybe set up a redirect to here?
>https://cwiki.apache.org/confluence/display/FLEX/FlexUnit
I am willing to turn it over to Apache or someone else. I would prefer not to
continue to host it.
Mike
>Are you saying you want Apache to take over the cost of owning that domain?
>And become the domain administrator, etc? Or just have Apache host the
>content and have the current FlexUnit.org redirect to the same >content on
>Apache? Or something else?
Good questions. I was looking to renew
>FWIW, the default software grant really just gives Apache a license. The
>original copyrights still make you the "owner". Anyway, I guess we should ask
>infra about what it takes to have flexunit.org point to apache servers to make
>sure it is "possible" then have a lazy vote on the dev list.
>TypeError: Error #1034: Type Coercion failed: cannot convert
>mx.collections::ArrayList@d6724d9 to
>mx.controls.dataGridClasses.DataGridColumn.
Based on your error message it looks like the way you are trying to provide the
columns to the DataGrid.
I didn't see you post that code at all. If you
If anyone is looking for a way to help, the FlexUnit wiki pages were all moved
over but some of the internal links still point back to FlexUnit.org instead of
self-referencing and that seems to orphan some content unless you know the URL.
It will be a while before I get to it, but a good simple
> Execute my FlexUnit4 Tests and write the results in a log
Daniel,
For the FlexUnit question you may want to look at the build scripts that Apache
is using to build FlexUnit and run the tests itself from Jenkins.
Mike
>BTW, Infra did say they can host flexunit.org, but they did not answer my
>question about what form the files need to be. It won't be any fun if we have
>to convert it to mdtext.
So agree.
Mike
>So, were the files on flexunit.org in the original FlexUnit donation? If not,
>what is the licensing on the site content? You may need to do another
>donation.
They likely weren't but I will look back at the language I wrote in the grant.
No problem if I need to do another.
As far as the si
>My bad for not actually looking at the site. I just tried though and it isn't
>responding. What do you mean by "domain staying intact?" Are you saying we
>should just map flexunit.org to a brand new site on Apache hardware and not
>take any of the old content?
Probably. And if there is some
>Hi Mike do you if anything has been done to solve the external dependencies
>issue so it could be donated to FlexUnit?
Nothing has been done.
>>That pattern works for mx components in MX DataGrid, but Spark components do
>>not "drop-in" as item renderers.
>Oleg says he is using mx components only.
>"(we use Flex3 MX and datagrids/ADGs throughout that app, so switching to
>Flex4 DGs/ADGs is not an option)."
"Now I am able to render C
A quick reminder in case anyone has cycles:
https://issues.apache.org/jira/browse/FLEX-33490
We donated the entire FlexUnit courseware and did the work to strip the HTML
down and make it easier to add to Apache. It does still need some decisions on
final binaries, etc. If anyone finds the cycl
>private function updateCompositionShapes():void
> {
> for each (var controller:ContainerController in
> controller)
> controller.updateCompositionShapes();
> }
Hmm. That's fun. It will end up doing nothing. Since th
Alex,
Since this seems to have 'passed' any idea of next steps? I am happy to do the
transfer paperwork but need the destination information... i.e. who will own it
and DNS server info.
Mike
Alex,
>Did you decide the content is already listed in a Grant doc? Otherwise it
>might be a good idea to do that.
Any content that I think you would actually like to have was included in the
original grant.
The rest is blog posts and fluff I don't think you need/want.
Mike
Going to check one last time before I just down this but I would like to take
down the FlexUnit github repo. People continue to comment and attempt to work
there and are frustrated because it's not active.
Any final objections or can I just do this...
Mike
>What need to be done to complete the donation of Flex Unit.
Yes, it just needs a release.
Other outstanding issues on FlexUnit which I don't have answers too... just in
case anyone cares to help:
1) The FlexUnit.org domain name. I am willing to pay for 10 years and point it
to Apache servers
>Mike, I'm pretty sure in past threads I've asked you to work with Infra
>directly by filing a JIRA. Can you do that? I >think it would go more
>smoothly if you worked directly with them.
Please feel free to send me that thread. I was fairly certain you were
discussing this with infra.
I kno
> Labriola or some other FlexUnit expert can tell us if they are for
> required features?
I will do my best to monitor this thread. I think all relevant issues have been
discussed in threads before but, if for any reason, you are waiting on a
response here from me or need technical info and don
>Open a JIRA ticket at https://issues.apache.org/jira/secure/Dashboard.jspa
>for the Infrastructure project. Tell them what you want to do and we'll see
>what happens.
Thanks.
I will give it a whirl,
Mike
>What does UIListener do and why does it require embedded fonts?
Our good friends at Adobe Consulting wrote and contributed it many moons ago.
It just provides a user interface of test progress. It could certainly be
tweaked to not use those fonts and instead just use the system defaults.
Mike
>Question is what should we do with this ? Should we change it to FlexUnit ?
That is more commonly known as FlexUnit 1. FlexUnit 4 should already run those
tests internally.
>More over I think library was modified in some way for TLF tests purposes. for
>ex: Couple of properties from TestCase
>I found when trying to get FlexUnit to work with FlexJS that there is some
>work needed there as well since the FlexUnit demos use mx or Spark
>applications around the runners.
Alex, this isn't completely true and I can help you with it. The runner itself
is entirely AS only, no Flex depende
I put together a presentation years ago on FlexUnit for contributors that
explained the architecture. I will find it and get that posted somewhere.
Additionally, if anyone is interested, I would be happy to do a session
sometime on the internals of it as well as talk about what was in the future
>Should there be mention of how to swap out Flash Builder's flexunit for this
>one in the readme?
Side not on this. The way to swamp out Flash Builder's version is convoluted.
Good reference:
http://probertson.com/articles/2011/05/03/flash-builder-4-5-tip-use-a-custom-flexunit-sdk/
As Paul ind
major pieces
fit together.
Mike
-Original Message-
From: Michael A. Labriola [mailto:labri...@digitalprimates.net]
Sent: Sunday, March 23, 2014 11:57 AM
To: dev@flex.apache.org
Subject: FlexUnit was Unit tests in TLF
I put together a presentation years ago on FlexUnit for contributors
Piotr,
>For ex: http://bit.ly/1jjjhJ8
>They do not have any tags [Test] above the methods and so on.
These should still be able to be run through FlexUnit 4.x FlexUnit 4
understands how to run FlexUnit 4.x, FlexUnit 1.x and Fluint 1.x tests.
However, it may still be worthwhile to update them s
>BTW, was one of the enhancements automated test generation? Maybe from
>metadata on the class?
There was a bigger project on that, yes, although we didn't get particularly
far before the Apache switch so most of it was put on hold. Also, Flash Builder
would create stubs for you based on the c
>OK, I didn't see any non-Flex apps around the runners. The deal is the mx and
>Spark apps can't wrap the runners and load in FlexJS classes at the same time.
> I think I'll just have to replace the Spark app with a FlexJS app.
No problem. We even got FlexUnit working in Flash pro at one point.
>I'd like Labriola to verify that DP should be in the NOTICE and whether anyone
>else like Adobe should also be in there.
I will make time to review but may have some questions as I feel a bit
mal-informed on procedure on these fronts.
Members of Adobe Consulting did write and donate the UIRunn
>FlexJS will keep metadata on the JS side, probably as simple properties on the
>classes and/or functions. I imagine we'll write a library that abstracts the
>fetching of metadata and returns it as an array of strings or maybe as JSON.
>If we could easily retrofit the metadata fetching in Flex
>BTW, a key motivating factor in creating a set of FlexUnit tests is to try to
>conform to that government standard you told me about. Do you have a 5
>sentence summary of what a test suite needs to cover? Is it measured purely
>by code coverage or do you have to have boundary tests, negative
>Are we talking about a single file here?
>https://github.com/apache/flex-flexunit/blob/master/FlexUnit4Test/src/UIRunner.mxml
No, we are talking about the code in the tag below:
Which resides here:
https://github.com/apache/flex-flexunit/tree/master/FlexUnit4UIListener
It was originally writ
>2) The code was modified by DP folks.
It was modified significantly as the original code was intended to work with
FlexUnit .9 alone and we modified it to work with both projects.
>It is important to verify that the copyright license text is not MPL.
It's a BSD-2 header.
Mike
>Sadly I can't even get at the original source code to compare now vs then as
>it look like Adobe removed it from source forge. There's only a binary
>download.
The plot thickens.
I went through old emails like an auditor and found a bit more detail. The
original code was committed by Xavi Beu
>Thicker still, I found this:
>http://opensource.adobe.com/svn//opensource/flexunit/trunk/FlexUnitLib/
Good find Alex. I didn't realize the UI was checked in here. This looks like
the original version checked in by Alistair on August 15th, 2008. I think we
can use this repo as part of our refere
>And in it I found AbstractRowData.as which looks exactly like the one on
>github for FlexUnit4 that Mike linked to earlier in this thread. I
haven't done any forensics on who checked it in and when, but keep in mind that
I am assuming that this code base on Adobe svn is the source for FlexUnit
>I checkout out the code from fluint.googlecode.com. An examination of the
>logs indicates that Xavi contributed files in FlexUnit4CIRunner as well as
>FlexUnit4UIRunner. Does that match your understanding? Were >those files
>brought into GitHub?
The UIRunner yes, it morphed into the UI List
And these 2 files were missing from the list which I assume should be Adobe
licensed.
>/FlexUnit4UIListener/src/org/flexunit/flexui/data/filter/IgnoredTestFunctionStatus.as
No, we wrote this one
>/FlexUnit4UIListener/src/org/flexunit/listeners/UIListener.as
No, this one was ours
>I would change the notice to say that some of the code (instead of original,
>since original is Adobe) was developed by DP.
It was not originally Adobe code. FlexUnit 4 is a complete rewrite. We included
FlexUnit1 for backward compatibility. FlexUnit 4 will run without a single line
of Adobe c
>Sorry about that - changed to Apache header and checked in.
No worries. Easy mistake as they were in similar packages. I just want Apache
headers on everything that can have them.
Mike
>OK, I just wanted to make sure. If the agreement was similar to the one Adobe
>had you sign when submitting patches then I think we're set on LICENSE and
>NOTICE. If the agreement was more Apache-like, then Brian has copyright and
>DP has license and then we may need to ask him.
Yep, we basi
>My understanding is that Adobe donated stuff to fluint under Adobe copyright
>and BSD. That's why you put the Adobe header back in a bunch of files. My
>view says that two of these Adobe files are in CIListener. I think we need
>Mike to verify or provide a more detailed explanation of how th
>Do anyone have any any objections to the current RC (copyright or otherwise)
>before I go to the effort of making another one? Has anyone actually tested
>the current RC for things that actually matter? Like that the signed fields
>are correct, it can compile and it actually works? I would like
>I used to pronounce them SWIK and SWIF ... is that the officially correct form?
Yep, that is correct.
Mike
1 - 100 of 259 matches
Mail list logo
