Exceptions escape

2 views
Skip to first unread message

Jesse Warden

unread,
Sep 30, 2009, 11:17:09 AM9/30/09
to fluint-di...@googlegroups.com
I know this was discussed before but forgot the resolution.  I'm having exceptions escape, mainly when I call "fail".  Specifically, fluint doesn't catch them, and they popup as the Exception box in Flash Player causing my tests to be in a pending status for years.

I'm looking around branches, thinking maybe I can merge some stuff that's newer?  Joel Hooks mentioned I just should use the FlexUnit4 branch which has all the metatag goodness, but I couldn't get it to compile in Flex 3.

Suggestions?

Jesse Warden

unread,
Sep 30, 2009, 11:35:54 AM9/30/09
to fluint-di...@googlegroups.com
In addition, my fail's fail to stop the timeout for asynchronous calls... so if it fails, it still waits for the timeout... UGH!

Michael Labriola

unread,
Sep 30, 2009, 11:39:43 AM9/30/09
to Fluint Discussions

Jesse,

Something seems to be going seriously wrong. If you can send me a
sample, I will debug.

I think Joel was likely recommending the FlexUnit 4 branch here:

http://opensource.adobe.com/svn/opensource/flexunit/branches/4.x/

Is that what you were trying to use?
Mike

On Sep 30, 10:35 am, Jesse Warden <jesse.war...@gmail.com> wrote:
> In addition, my fail's fail to stop the timeout for asynchronous calls... so
> if it fails, it still waits for the timeout... UGH!
>
> On Wed, Sep 30, 2009 at 11:17 AM, Jesse Warden <jesse.war...@gmail.com>wrote:
>
>
>
> > I know this was discussed before but forgot the resolution.  I'm having
> > exceptions escape, mainly when I call "fail".  Specifically, fluint doesn't
> > catch them, and they popup as the Exception box in Flash Player causing my
> > tests to be in a pending status for years.
> > I'm looking around branches, thinking maybe I can merge some stuff that's
> > newer?  Joel Hooks mentioned I just should use the FlexUnit4 branch which
> > has all the metatag goodness, but I couldn't get it to compile in Flex 3.
>
> > Suggestions?- Hide quoted text -
>
> - Show quoted text -

Jesse Warden

unread,
Sep 30, 2009, 11:42:03 AM9/30/09
to fluint-di...@googlegroups.com
Yeah, that's what I downloaded.  I already read the docs from your post 10 years ago about how to use it and such; it's just getting it to compile was rough.

I can't really send you a sample since it's proprietary HBO code; doing a connect session would take a lot of your precious time, so..... how about instead you just (A) confirm/deny that FlexUnit4 is the future while fluint is not and (B) it CAN be used in Flex 3 using Flex 3.3 SDK...?

Jesse Warden

unread,
Sep 30, 2009, 11:44:26 AM9/30/09
to fluint-di...@googlegroups.com
So Joel just pointed me here:


Apparently the SWC's in there work with Flex 3, not sure about SDK 3.3; trying now...

Michael Labriola

unread,
Sep 30, 2009, 11:46:20 AM9/30/09
to Fluint Discussions
I am just suprised that getting it to compile was rough, hence the
reason I am asking.

Glad to do a connect if that helps, but let me answer your questions
first:

a) FlexUnit 4 is the future for the unit testing piece of what is
Fluint today. However, fluint features (and libraries) will basically
morph into a plugin to FlexUnit4 to allow the fancier integration
testing pieces... so both projects are the future, but each with
separate focus. Ultimately though, fluint will be more of the behind
the scenes piece and FlexUnit 4 will likely be the project most people
sync and use directly

b) FlexUnit 4 works with all versions of Flex 3 and 4. It can also be
compiled to work with Flex 2 and even ASOnly projects

Mike



On Sep 30, 10:42 am, Jesse Warden <jesse.war...@gmail.com> wrote:
> Yeah, that's what I downloaded.  I already read the docs from your post 10
> years ago about how to use it and such; it's just getting it to compile was
> rough.
> I can't really send you a sample since it's proprietary HBO code; doing a
> connect session would take a lot of your precious time, so..... how about
> instead you just (A) confirm/deny that FlexUnit4 is the future while fluint
> is not and (B) it CAN be used in Flex 3 using Flex 3.3 SDK...?
>
> > > - Show quoted text -- Hide quoted text -

Michael Labriola

unread,
Sep 30, 2009, 12:00:10 PM9/30/09
to Fluint Discussions

Side note: If you can get me a reproducible case outside of your code,
I can also likely figure out the issue or if we are missing a patch
pretty quickly

On Sep 30, 10:42 am, Jesse Warden <jesse.war...@gmail.com> wrote:
> Yeah, that's what I downloaded.  I already read the docs from your post 10
> years ago about how to use it and such; it's just getting it to compile was
> rough.
> I can't really send you a sample since it's proprietary HBO code; doing a
> connect session would take a lot of your precious time, so..... how about
> instead you just (A) confirm/deny that FlexUnit4 is the future while fluint
> is not and (B) it CAN be used in Flex 3 using Flex 3.3 SDK...?
>
> > > - Show quoted text -- Hide quoted text -

Jesse Warden

unread,
Sep 30, 2009, 1:43:02 PM9/30/09
to fluint-di...@googlegroups.com
Cool, last question: where are y'all discussing this stuff?  On the forums Confluence links to?  Got an email list instead?

Michael Labriola

unread,
Sep 30, 2009, 2:01:58 PM9/30/09
to Fluint Discussions

Some of this is being discussed on the forums but this will likely be
evolving in the coming weeks. Behind the scenes we have been working
dilligently to get systems in place to open things up a bit more. We
are going to start actively discussing architecture and educating
about the internals of the framework as well as have open meeting
about our iterations with to-do lists where we can recruit help from
the community.

Over the coming weeks to a month, you should see more and more of
this. Right now there are just a few of us pulling everything we can
together, but I am hoping to expand that dramtically in even the short
term.

Cheers,
Mike

On Sep 30, 12:43 pm, Jesse Warden <jesse.war...@gmail.com> wrote:
> Cool, last question: where are y'all discussing this stuff?  On the forums
> Confluence links to?  Got an email list instead?
>

Jesse Warden

unread,
Sep 30, 2009, 3:06:50 PM9/30/09
to fluint-di...@googlegroups.com
Ok, so I'll ask here then since technically fluint is backwards compatible with the RunWith.

Why use [Test(async,timeout="500")] when you:
- have to use the var response:Function = Async.asyncHandler thing anyway
- and you have to set the timeout in that above function as well; I don't get what the tag is doing...

Jesse Warden

unread,
Sep 30, 2009, 3:46:31 PM9/30/09
to fluint-di...@googlegroups.com
CRAP, same thing in FlexUnit4... it's gotta be me doing a fail in an error method that is indepedent on of the asyncHandler...

Like:

- setup creates service + error handler
- later on, I use an asyncHandler for the complete and timeout
- error runs, exception isn't caught
- timeout fires later, and my tests continue

Make sense?

Michael Labriola

unread,
Sep 30, 2009, 4:03:59 PM9/30/09
to Fluint Discussions

Jesse,

If you are writing Fluint tests, then FlexUnit 4 is literally
instantiating Fluint in the background to handle this, so it isn't a
suprise that you see the same behavior as it is the same code.

Can you give me anything that I can use to reproduce. I can look at it
today if you can.

Mike

On Sep 30, 2:46 pm, Jesse Warden <jesse.war...@gmail.com> wrote:
> CRAP, same thing in FlexUnit4... it's gotta be me doing a fail in an error
> method that is indepedent on of the asyncHandler...
> Like:
>
> - setup creates service + error handler
> - later on, I use an asyncHandler for the complete and timeout
> - error runs, exception isn't caught
> - timeout fires later, and my tests continue
>
> Make sense?
>

Michael Labriola

unread,
Sep 30, 2009, 4:53:52 PM9/30/09
to Fluint Discussions

To answer these questions:

First, the async property actually causes this test to take a separate
path internal to the runner. Without the async tag, much of the code
is assumed to be synchronous and runs without the safety net necessary
for async to occur. With the async property, that code is enabled.
This purely a performance issue. Things are slower when we need to
create handlers and logic to consider async communication.

The timeout itself can be ignored at this moment for your purposes.
This defines a timeout window for the whole test as I might have (n)
asynchronous steps that occur as part of one test. This serves as a
way for me to collectively indicate that they should take no more than
500ms. The timing on the individual handlers indicates how long each
sub-step might take. Truthfully, this is more confusing than it needs
to be and we need to find a way to simplify this a bit.

Mike

On Sep 30, 2:06 pm, Jesse Warden <jesse.war...@gmail.com> wrote:
> Ok, so I'll ask here then since technically fluint is backwards compatible
> with the RunWith.
> Why use [Test(async,timeout="500")] when you:
> - have to use the var response:Function = Async.asyncHandler thing anyway
> - and you have to set the timeout in that above function as well; I don't
> get what the tag is doing...
>
Reply all
Reply to author
Forward
0 new messages