[RESULT][VOTE] Moving JIRA issues and Confluence Wiki to GitHub

[Olaf Krueger]

Failed.

Result:
[0]: 4 votes 
[+1]: 2 votes

So we leave anything as it is.

Thanks for voting!

Olaf



--
Sent from: http://apache-flex-development.247.n4.nabble.com/

RE: [Vote] Moving JIRA issues and Confluence Wiki to GitHub

[Yishay Weiss]

+0.



As others, I don’t see a clear benefit. Both tracking systems are solid.



On the other hand, if a member of the community is keen to do something and 
sees some benefit in it I think we shouldn’t stand in his/her way.




From: Olaf Krueger 
Sent: Saturday, January 19, 2019 8:42:42 PM
To: dev@flex.apache.org
Subject: Re: [Vote] Moving JIRA issues and Confluence Wiki to GitHub

+1

In 2018, 15 issues were created by using JIRA.
So I don't expect too many duplicates for the future.
If it is possible, we would set JIRA and the Conflucence Wiki read-only.



--
Sent from: http://apache-flex-development.247.n4.nabble.com/

Re: [Vote] Moving JIRA issues and Confluence Wiki to GitHub

[Olaf Krueger]

+1 

In 2018, 15 issues were created by using JIRA.
So I don't expect too many duplicates for the future.
If it is possible, we would set JIRA and the Conflucence Wiki read-only.



--
Sent from: http://apache-flex-development.247.n4.nabble.com/

Re: [Vote] Moving JIRA issues and Confluence Wiki to GitHub

[Carlos Rovira]

+1

El El vie, 18 ene 2019 a las 22:12, Tom Chiverton 
escribió:

> Vote: -0
>
> I don't see there is much benefit, and it creates two streams of bugs. If
> it's
> true that Jira will be closed to new bugs but otherwise stay open, my vote
> changes to +0.
>
>
> --
> Tom
> Q. According to Intel the Pentium conforms to I.E.E.E. stnd. 754 and 854
> for
> floating point arithmetic. If you fly in an aircraft, drive over a bridge
> etc. etc. etc. designed on a Pentium, what is the correct pronunciation of
> 'I.E.E.E.'?
> A. Aaaaieee!
>
>
>
> --
Carlos Rovira
http://about.me/carlosrovira

Re: [Vote] Moving JIRA issues and Confluence Wiki to GitHub

[Tom Chiverton]

Vote: -0

I don't see there is much benefit, and it creates two streams of bugs. If it's 
true that Jira will be closed to new bugs but otherwise stay open, my vote 
changes to +0.


-- 
Tom
Q. According to Intel the Pentium conforms to I.E.E.E. stnd. 754 and 854 for
floating point arithmetic. If you fly in an aircraft, drive over a bridge
etc. etc. etc. designed on a Pentium, what is the correct pronunciation of
'I.E.E.E.'?
A. Aaaaieee!

Re: [Vote] Moving JIRA issues and Confluence Wiki to GitHub

[Piotr Zarzycki]

Hi,

+0, as long as we won't make JIRA and Confluence read only.

Thanks,
Piotr

pt., 18 sty 2019 o 00:19 Justin Mclean  napisał(a):

> Hi,
>
> +0 I see little reason to change and split where you need to look at
> things in two places.
>
> I assume that JIRA and Confluence will not be made read only?
>
> Thanks,
> Justin



-- 

Piotr Zarzycki

Patreon: *https://www.patreon.com/piotrzarzycki
*

Re: [Vote] Moving JIRA issues and Confluence Wiki to GitHub

[Justin Mclean]

Hi,

+0 I see little reason to change and split where you need to look at things in 
two places.

I assume that JIRA and Confluence will not be made read only?

Thanks,
Justin

[Vote] Moving JIRA issues and Confluence Wiki to GitHub

[Olaf Krueger]

Hi,
this is the official vote thread to approve the move from JIRA issues to
GitHub issues and to move the Confluence Wiki to the Github Wiki.

As discussed here [1], the idea is to don't migrate the JIRA issues to
GitHub. Instead, we want a clean restart at GitHub issues. The JIRA issues
will be treated as an archive.
We may will move some Wiki articles to GitHub.

Vote 
[ ] +1 Agree
[ ] -1 Veto (please provide specific comments) 

The vote passes if there is: 
- At least 3 +1 votes from the PMC 
- More positive votes than negative votes 

This vote will be open for at least 72 hours. 

[1] 
http://apache-flex-development.247.n4.nabble.com/Enabling-GitHub-quot-Issues-quot-and-quot-Wiki-quot-for-flex-repositories-td65601.html




--
Sent from: http://apache-flex-development.247.n4.nabble.com/

Re: Enabling GitHub "Issues" and "Wiki" for flex repositories

[Olaf Krueger]

@Justin

> i’m that sure that it is. JIRA has a number of significant benefits over
> GitHub issues including better searching and querying. (Yes I use both). 

Just to make sure that I got it right:
Does that mean that you basically support the idea of moving from JIRA to
GitHub issues or do you see some significant drawbacks?

@All
In order to get an "official" result, I propose to start a vote about it.
Will do this soon...

Thanks,
Olaf





--
Sent from: http://apache-flex-development.247.n4.nabble.com/

RE: [Non-DoD Source] Re: Enabling GitHub "Issues" and "Wiki" for flex repositories

[Kessler CTR Mark J]

+1

Especially those old Adobe issues that were imported.  To me they are mostly 
clutter because of the quantity, age, and missing attachments.  I would much 
rather have a clean start.  But brining in any new (from a given time range x 
number of months) wouldn't be a bonus, unless that's an entirely manual process.

-Mark K


-Original Message-
From: Carlos Rovira [mailto:carlosrov...@apache.org]
Sent: Wednesday, January 9, 2019 4:32 AM
To: dev@flex.apache.org
Subject: [Non-DoD Source] Re: Enabling GitHub "Issues" and "Wiki" for flex 
repositories

Hi Olaf,

I like your idea. I think going to GitHub is like when we migrate from SVN
to GIT, we must go with the times.
Is like FlexJS vs Royale wiki. We started a new one with newer info in
Royale, but still have useful content from FlexJS that we should not remove
and migrate as we can.

So for me is ok to star clean on Github and lock JIRA, (or just be able to
close or update JIRA content, but be unable to create new ones on JIRA.
People working in Flex should now that as part of the things they need to
know. If want to create a new ticket, first check JIRA, but create from
scratch or recreate in Github. This will be an improvement for Flex in the
end

just my 2...

Carlos




El mar., 8 ene. 2019 a las 23:18, Olaf Krueger ()
escribió:

> > Are we going to say that we won't fix any of the bugs currently open in
> JIRA?...
> > My concern is having two active bug bases...
>
> We maybe could set or treat or freeze JIRA as read-only and could link to
> the entire Flex-JIRA from GitHub in order to not losing the history.
> If somebody is going to fix one of those existing JIRA issues we could
> simply recreate it manually as a GitHub issue.
> If somebody will file a new GitHub issue we may want to check if there's a
> JIRA duplicate in order to link it from the GitHub issue.
> But on the other hand, I don't believe when somebody creates a new JIRA
> that
> somebody checks if there's already a JIRA duplicate.
>
> Thoughts?
>
> Olaf
>
>
>
> --
> Sent from: http://apache-flex-development.247.n4.nabble.com/
>


--
Carlos Rovira
http://about.me/carlosrovira

Re: Enabling GitHub "Issues" and "Wiki" for flex repositories

[Carlos Rovira]

Hi Olaf,

I like your idea. I think going to GitHub is like when we migrate from SVN
to GIT, we must go with the times.
Is like FlexJS vs Royale wiki. We started a new one with newer info in
Royale, but still have useful content from FlexJS that we should not remove
and migrate as we can.

So for me is ok to star clean on Github and lock JIRA, (or just be able to
close or update JIRA content, but be unable to create new ones on JIRA.
People working in Flex should now that as part of the things they need to
know. If want to create a new ticket, first check JIRA, but create from
scratch or recreate in Github. This will be an improvement for Flex in the
end

just my 2...

Carlos




El mar., 8 ene. 2019 a las 23:18, Olaf Krueger ()
escribió:

> > Are we going to say that we won't fix any of the bugs currently open in
> JIRA?...
> > My concern is having two active bug bases...
>
> We maybe could set or treat or freeze JIRA as read-only and could link to
> the entire Flex-JIRA from GitHub in order to not losing the history.
> If somebody is going to fix one of those existing JIRA issues we could
> simply recreate it manually as a GitHub issue.
> If somebody will file a new GitHub issue we may want to check if there's a
> JIRA duplicate in order to link it from the GitHub issue.
> But on the other hand, I don't believe when somebody creates a new JIRA
> that
> somebody checks if there's already a JIRA duplicate.
>
> Thoughts?
>
> Olaf
>
>
>
> --
> Sent from: http://apache-flex-development.247.n4.nabble.com/
>


-- 
Carlos Rovira
http://about.me/carlosrovira

Re: Enabling GitHub "Issues" and "Wiki" for flex repositories

[Olaf Krueger]

> Are we going to say that we won't fix any of the bugs currently open in
JIRA?...
> My concern is having two active bug bases...

We maybe could set or treat or freeze JIRA as read-only and could link to
the entire Flex-JIRA from GitHub in order to not losing the history.
If somebody is going to fix one of those existing JIRA issues we could
simply recreate it manually as a GitHub issue.
If somebody will file a new GitHub issue we may want to check if there's a
JIRA duplicate in order to link it from the GitHub issue.
But on the other hand, I don't believe when somebody creates a new JIRA that
somebody checks if there's already a JIRA duplicate.

Thoughts?

Olaf



--
Sent from: http://apache-flex-development.247.n4.nabble.com/

Re: Enabling GitHub "Issues" and "Wiki" for flex repositories

[Olaf Krueger]

> Do we know if it possible it import all historical JIRA issues ...

I don't know it this is possible.

> JIRA has a number of significant benefits over GitHub issues including
> better searching and querying

JIRA is probably more powerful but I guess the simplicity of GitHub is the
greater advantage here.

I think if we would go with GitHub issues it would be a good opportunity for
a "clean restart".
That doesn't mean that the existing JIRA issues are useless, these should
remain there.

Olaf






--
Sent from: http://apache-flex-development.247.n4.nabble.com/

Re: Enabling GitHub "Issues" and "Wiki" for flex repositories

[Justin Mclean]

Hi,

> I think it would be easier and more transparent for folks to handle future 
> development (if it will happen) by using GitHub.

i’m that sure that it is. JIRA has a number of significant benefits over GitHub 
issues including better searching and querying. (Yes I use both).

Do we know if it possible it import all historical JIRA issues (open and 
closed) into git issues? And if we do what if anything do we loose? Don’t 
forget the old Adobe issues are also in JIRA and having a history of which bugs 
have been resolved or not is very useful.

Thanks,
Justin

Re: Enabling GitHub "Issues" and "Wiki" for flex repositories

[Alex Harui]

Are we going to say that we won't fix any of the bugs currently open in JIRA?  
That might be ok.  It would help prioritize and make more clear what folks are 
still interested in.  My concern is having two active bug bases.  You may still 
end up referencing the JIRA bugs when marking new bugs as duplicates.

-Alex

On 1/8/19, 11:37 AM, "Olaf Krueger"  wrote:

I was thinking more like a cut of the old Wiki and JIRA than a migration of
the existing stuff.
I think it would be easier and more transparent for folks to handle future
development (if it will happen) by using GitHub.
We may migrate some wiki stuff someday, don't know.

Thanks,
Olaf





--
Sent from: 
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fapache-flex-development.247.n4.nabble.com%2Fdata=02%7C01%7Caharui%40adobe.com%7C134f454fbec74f08d97a08d675a0bada%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636825730552562081sdata=GF7m81Kmna%2BISLtw2z1wsCKXyoTO4ps151zPAu2E9D0%3Dreserved=0

Re: Enabling GitHub "Issues" and "Wiki" for flex repositories

[Olaf Krueger]

I was thinking more like a cut of the old Wiki and JIRA than a migration of
the existing stuff.
I think it would be easier and more transparent for folks to handle future
development (if it will happen) by using GitHub.
We may migrate some wiki stuff someday, don't know.

Thanks,
Olaf





--
Sent from: http://apache-flex-development.247.n4.nabble.com/

Re: Enabling GitHub "Issues" and "Wiki" for flex repositories

[Alex Harui]

I'm not opposed, but I'm not sure it is worth it.  I think there are a 800+ 
bugs filed since Flex came to Apache that are still open in JIRA so JIRA is 
known to our Flex users.  Not sure if it will be confusing to have two places 
to look and I don't think we should migrate those bugs either.  Same for the 
wiki.  There's good content there but that content may not be worth migrating.  
But if someone wants to take the time, update the website links, etc, that's 
fine with me.

My 2 cents,
-Alex

On 1/8/19, 8:41 AM, "Carlos Rovira"  wrote:

That's ok for me, thanks! :)

About the old flex wiki, I think we still have some good info about FlexJS
still not migrated to Royale. I assume all this migration is not about all
of that right?

thanks!



El mar., 8 ene. 2019 a las 10:10, Olaf Krueger ()
escribió:

> Hi,
> I would like to create an INFRA Jira in order to ask them to enable GitHub
> "Issues" and the "Wiki" for following repositories:
>
>  - flex-tlf.git
>  - flex-utilities.git
>  - flex-flexunit.git
>  - flex-site.git
>  - flex-external.git
>  - flex-sdk.git
>  - flex-blazeds.git
>  - flex-radii8.git
>
> We will maybe use these GitHub features mainly for the flex-sdk repo, but
> maybe it makes sense to just enable it for all flex repos?
>
> Any objections?
>
> Thanks,
> Olaf
>
>
>
>
> --
> Sent from: 
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fapache-flex-development.247.n4.nabble.com%2Fdata=02%7C01%7Caharui%40adobe.com%7C082657ca2436418529cb08d675882ff3%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636825625117173639sdata=FmviyLt3BGx3m7ysRUtjEMLfoUIGF35D%2F78xbhgsfz4%3Dreserved=0
>


-- 
Carlos Rovira

https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fabout.me%2Fcarlosroviradata=02%7C01%7Caharui%40adobe.com%7C082657ca2436418529cb08d675882ff3%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636825625117173639sdata=Y2RMBHeYo3aP52ueG26cd805XInmz5V4Q8kGfOVdQ1I%3Dreserved=0

Re: Enabling GitHub "Issues" and "Wiki" for flex repositories

[Carlos Rovira]

That's ok for me, thanks! :)

About the old flex wiki, I think we still have some good info about FlexJS
still not migrated to Royale. I assume all this migration is not about all
of that right?

thanks!



El mar., 8 ene. 2019 a las 10:10, Olaf Krueger ()
escribió:

> Hi,
> I would like to create an INFRA Jira in order to ask them to enable GitHub
> "Issues" and the "Wiki" for following repositories:
>
>  - flex-tlf.git
>  - flex-utilities.git
>  - flex-flexunit.git
>  - flex-site.git
>  - flex-external.git
>  - flex-sdk.git
>  - flex-blazeds.git
>  - flex-radii8.git
>
> We will maybe use these GitHub features mainly for the flex-sdk repo, but
> maybe it makes sense to just enable it for all flex repos?
>
> Any objections?
>
> Thanks,
> Olaf
>
>
>
>
> --
> Sent from: http://apache-flex-development.247.n4.nabble.com/
>


-- 
Carlos Rovira
http://about.me/carlosrovira

Enabling GitHub "Issues" and "Wiki" for flex repositories

[Olaf Krueger]

Hi,
I would like to create an INFRA Jira in order to ask them to enable GitHub
"Issues" and the "Wiki" for following repositories:

 - flex-tlf.git 
 - flex-utilities.git 
 - flex-flexunit.git 
 - flex-site.git 
 - flex-external.git 
 - flex-sdk.git 
 - flex-blazeds.git 
 - flex-radii8.git

We will maybe use these GitHub features mainly for the flex-sdk repo, but
maybe it makes sense to just enable it for all flex repos?

Any objections?

Thanks,
Olaf




--
Sent from: http://apache-flex-development.247.n4.nabble.com/

[FLEXJS][ROYALE] Github Issues replacing Apache JIRA

[Alex Harui]

Hi,

It looks like GitHub Issues has been enabled on the new Royale repos.  The
Royale folks will no longer be watching Apache JIRA for issues related to
FlexJS and Royale.  Existing FlexJS JIRA issues will not be automatically
migrated to GitHub Issues.  If you have an open FlexJS issue in Apache
JIRA please open a new issue in Github Issues.

GitHub Issues for the royale-asjs repo is here [1]

[1] https://github.com/apache/royale-asjs/issues

Thanks,
-Alex

CSS compiler issues

[Harbs]

I just created a couple of JIRAs on compiler issues that have been bugging me 
for a while.[1][2] Neither of these are super-important, but it would be nice 
if it worked better.

Harbs

[1]https://issues.apache.org/jira/browse/FLEX-35340 
<https://issues.apache.org/jira/browse/FLEX-35340>
[2]https://issues.apache.org/jira/browse/FLEX-35341 
<https://issues.apache.org/jira/browse/FLEX-35341>

Re: 4.16.0 Install issues

[Josh Tynjala]

Be sure to test on Windows 7. It has an older version of PowerShell, so
some things that work in newer versions may not work there. I had the unzip
working on Windows 10 after a long day of wrapping my head around running
PowerShell from AIR. The next morning, I discovered that I had to redo it
because I was using a new API that wasn't available in the older version of
PowerShell.

- Josh

On Fri, Jul 7, 2017 at 11:08 AM, Nicholas Kwiatkowski <
nicholaskwiatkow...@gmail.com> wrote:

> I think that was the direction I was going to go. Thanks!
>
> On Fri, Jul 7, 2017 at 11:30 AM, Josh Tynjala <joshtynj...@gmail.com>
> wrote:
>
> > In addition to MD5, the code the installer uses to unzip files can also
> > cause Error #1000.
> >
> > I came up with a solution for that using PowerShell that I posted on the
> > mailing list:
> >
> > http://mail-archives.apache.org/mod_mbox/flex-dev/201706.
> > mbox/%3CCADZSWG4j-DwOFUbFGrJaOZsXTgU1FfwmE9fJUP5
> > dnrgCGFkhLQ%40mail.gmail.com%3E
> >
> > I looked into MD5 with PowerShell, and from what I could tell from a
> quick
> > search, that should be possible too.
> >
> > - Josh
> >
> > On Wed, Jul 5, 2017 at 6:13 PM, Nicholas Kwiatkowski <que...@apache.org>
> > wrote:
> >
> > > Is anybody actually addressing the issues people have been reporting
> > about
> > > the installer and/or ANT script for Flex SDK 4.16.0?
> > >
> > > Right now I see two major issues that are preventing even people who
> are
> > > familiar with the SDK from doing installs :
> > >
> > >  - In the installer, selecting AIR 25.0 gives users a non-descript
> "error
> > > 1000".  This is due toe the md5 checking in the installer running out
> of
> > > memory
> > >  - Trying to install via ANT is also broken under Windows (any version
> of
> > > AIR SDK).  As packaged, it always errors because it tries to install
> the
> > > MacOS AIR SDK.  Additionally, the optional components that are
> currently
> > > hosted on sourceforge fail to download due to some SSL errors (I've
> > tested
> > > this with the latest java sdk and ANT build).
> > >
> > > The AIR installer issue will require us to rip-and-replace the md5
> > > calculation functions.  I've started looking at it, but I don't think
> it
> > > will be an easy feat.
> > > Fixing the ANT script for Windows trying to install the mac air dmg is
> an
> > > easy fix (but it will require us to do a dot release to push it out)
> > > The SSL errors are because sourceforge is using SANs on their SSL
> certs,
> > > and the current versions of ANT don't know how to read them to validate
> > > them.  This may be out of our control.  Anybody know if we can convince
> > > Adobe to either donate those chunks of code or at least to move them
> to a
> > > different host?  It looks like the code involved is OSMF, AFE, AGLJ,
> > > rideau and Flex-Fontkit.  The alternative to a different host is for us
> > to
> > > ignore SSL errors, but that could be potentially dangerous.
> > >
> > > If nobody is working on these, I can start to take a crack at them, but
> > > honestly, the installer is extremely fragile at this point and I'm not
> > > looking forward to even trying to figure out what is going on in there
> > > again.
> > >
> > > -Nick
> > >
> >
>

Re: 4.16.0 Install issues

[Nicholas Kwiatkowski]

I think that was the direction I was going to go. Thanks!

On Fri, Jul 7, 2017 at 11:30 AM, Josh Tynjala <joshtynj...@gmail.com> wrote:

> In addition to MD5, the code the installer uses to unzip files can also
> cause Error #1000.
>
> I came up with a solution for that using PowerShell that I posted on the
> mailing list:
>
> http://mail-archives.apache.org/mod_mbox/flex-dev/201706.
> mbox/%3CCADZSWG4j-DwOFUbFGrJaOZsXTgU1FfwmE9fJUP5
> dnrgCGFkhLQ%40mail.gmail.com%3E
>
> I looked into MD5 with PowerShell, and from what I could tell from a quick
> search, that should be possible too.
>
> - Josh
>
> On Wed, Jul 5, 2017 at 6:13 PM, Nicholas Kwiatkowski <que...@apache.org>
> wrote:
>
> > Is anybody actually addressing the issues people have been reporting
> about
> > the installer and/or ANT script for Flex SDK 4.16.0?
> >
> > Right now I see two major issues that are preventing even people who are
> > familiar with the SDK from doing installs :
> >
> >  - In the installer, selecting AIR 25.0 gives users a non-descript "error
> > 1000".  This is due toe the md5 checking in the installer running out of
> > memory
> >  - Trying to install via ANT is also broken under Windows (any version of
> > AIR SDK).  As packaged, it always errors because it tries to install the
> > MacOS AIR SDK.  Additionally, the optional components that are currently
> > hosted on sourceforge fail to download due to some SSL errors (I've
> tested
> > this with the latest java sdk and ANT build).
> >
> > The AIR installer issue will require us to rip-and-replace the md5
> > calculation functions.  I've started looking at it, but I don't think it
> > will be an easy feat.
> > Fixing the ANT script for Windows trying to install the mac air dmg is an
> > easy fix (but it will require us to do a dot release to push it out)
> > The SSL errors are because sourceforge is using SANs on their SSL certs,
> > and the current versions of ANT don't know how to read them to validate
> > them.  This may be out of our control.  Anybody know if we can convince
> > Adobe to either donate those chunks of code or at least to move them to a
> > different host?  It looks like the code involved is OSMF, AFE, AGLJ,
> > rideau and Flex-Fontkit.  The alternative to a different host is for us
> to
> > ignore SSL errors, but that could be potentially dangerous.
> >
> > If nobody is working on these, I can start to take a crack at them, but
> > honestly, the installer is extremely fragile at this point and I'm not
> > looking forward to even trying to figure out what is going on in there
> > again.
> >
> > -Nick
> >
>

Re: [FlexJS][VS Code] dual issues

[Alex Harui]

Don't know.  Right now, you can set up an FB project with the "JS" SWCs
and not see Flash APIs in the intelligence.  But I'm not clear that FB
allows for any switches/options in how it gets the data for code
intelligence.  I think Josh is saying he might use the first target.  I
will see if that works for FB if/when I get around to actually doing this.

-Alex

On 7/7/17, 1:04 AM, "piotrz" <piotrzarzyck...@gmail.com> wrote:

>Alex,
>
>Do you mean by teaching compiler that if I use -target=JSFlex - I will not
>see flash api hints in intellisence ?
>
>Thanks,
>Piotr
>
>
>
>-
>Apache Flex PMC
>piotrzarzyck...@gmail.com
>--
>View this message in context:
>https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fapache-fle
>x-development.247.n4.nabble.com%2FFlexJS-VS-Code-dual-issues-tp62894p6
>2929.html=02%7C01%7C%7C0e8aa1a17ecd4ca28add08d4c511556b%7Cfa7b1b5a7b3
>4438794aed2c178decee1%7C0%7C0%7C636350125614586979=z9APgkfyPFqzQagcK
>1%2FAWeV94WOsljZJcIl6qgC74k0%3D=0
>Sent from the Apache Flex Development mailing list archive at Nabble.com.

Re: 4.16.0 Install issues

[Josh Tynjala]

In addition to MD5, the code the installer uses to unzip files can also
cause Error #1000.

I came up with a solution for that using PowerShell that I posted on the
mailing list:

http://mail-archives.apache.org/mod_mbox/flex-dev/201706.mbox/%3CCADZSWG4j-DwOFUbFGrJaOZsXTgU1FfwmE9fJUP5dnrgCGFkhLQ%40mail.gmail.com%3E

I looked into MD5 with PowerShell, and from what I could tell from a quick
search, that should be possible too.

- Josh

On Wed, Jul 5, 2017 at 6:13 PM, Nicholas Kwiatkowski <que...@apache.org>
wrote:

> Is anybody actually addressing the issues people have been reporting about
> the installer and/or ANT script for Flex SDK 4.16.0?
>
> Right now I see two major issues that are preventing even people who are
> familiar with the SDK from doing installs :
>
>  - In the installer, selecting AIR 25.0 gives users a non-descript "error
> 1000".  This is due toe the md5 checking in the installer running out of
> memory
>  - Trying to install via ANT is also broken under Windows (any version of
> AIR SDK).  As packaged, it always errors because it tries to install the
> MacOS AIR SDK.  Additionally, the optional components that are currently
> hosted on sourceforge fail to download due to some SSL errors (I've tested
> this with the latest java sdk and ANT build).
>
> The AIR installer issue will require us to rip-and-replace the md5
> calculation functions.  I've started looking at it, but I don't think it
> will be an easy feat.
> Fixing the ANT script for Windows trying to install the mac air dmg is an
> easy fix (but it will require us to do a dot release to push it out)
> The SSL errors are because sourceforge is using SANs on their SSL certs,
> and the current versions of ANT don't know how to read them to validate
> them.  This may be out of our control.  Anybody know if we can convince
> Adobe to either donate those chunks of code or at least to move them to a
> different host?  It looks like the code involved is OSMF, AFE, AGLJ,
> rideau and Flex-Fontkit.  The alternative to a different host is for us to
> ignore SSL errors, but that could be potentially dangerous.
>
> If nobody is working on these, I can start to take a crack at them, but
> honestly, the installer is extremely fragile at this point and I'm not
> looking forward to even trying to figure out what is going on in there
> again.
>
> -Nick
>

Re: 4.16.0 Install issues

[Nicholas Kwiatkowski]

I'm looking at the 4.16.0 release package, not what is in the source tree.
That's why I was asking at first if anybody was working on it to make sure
it hadn't been tackled, or nobody was actively working on the fix.

-Nick

On Fri, Jul 7, 2017 at 11:12 AM, Alex Harui <aha...@adobe.com.invalid>
wrote:

> What version of installer.xml are you looking at?  I just synced up the
> develop branch in flex-sdk and took a look at there are merge conflict
> markers in the file!  Apparently, nobody has tried to install a nightly
> build of Flex SDK.  But I think the version from the release branch has
> the right fix in it, although I think it is still using "new Ant" syntax
> in the if/unless which the Installer does not support.  So whoever fixes
> this needs to also replace:
>
>   unless="${shouldUnzip}"
>
> With:
>
>   unless="shouldUnzip"
>
>
> And same for
>
>   if="shouldUnzip"
>
> -Alex
>
>
> On 7/7/17, 3:57 AM, "Nicholas Kwiatkowski" <nicholaskwiatkow...@gmail.com>
> wrote:
>
> >I installed the latest version of Java, but didn't touch the JCE (I have
> >to
> >keep the standard set on my machine and extend, per project because of
> >some
> >of the other projects I'm working on).  Just using the net.url package
> >within a quick java app seems grab the files from sourceforge with no
> >problems -- it's the way that ANT is calling it that is causing the issue.
> >It's quite possible they are overriding the standard security chain for
> >some reason.  Either way, asking our users to change out the cryptology
> >libraries and download them from another source isn't in good form.
> >
> >As far as the logic, there is clearly a bug.  It still bombs out on the
> >latest ANT version.  We didn't do a conditional in the ant script between
> >the call to unpack the dmg vs. zip files for the AIR install.  The mac one
> >probably works because ANT on mac understands the unzip directive, but the
> >other way complains that it doesn't know how to deal with a dmg file, so
> >the ant script errors out.  That one is an easy fix.
> >
> >-Nick
> >
> >
> >On Fri, Jul 7, 2017 at 1:54 AM, Alex Harui <aha...@adobe.com.invalid>
> >wrote:
> >
> >> Maybe I wasn't clear enough:
> >>
> >> I don't think the version of Ant will affect the SSL issues.  The SSL
> >> issues should be solvable via the JCE upgrade:
> >>
> >>
> >>https://na01.safelinks.protection.outlook.com/?url=
> http%3A%2F%2Fstackover
> >>flow.com%2Fquestions%2F38203971%2Fjavax-net-=
> 02%7C01%7C%7Cfe4cee1f24
> >>f849200a1108d4c526ee4e%7Cfa7b1b5a7b34438794aed2c178de
> cee1%7C0%7C0%7C63635
> >>0218379468309=3ngz8fD86odTrjJ5D8G1SKFu1j7H7q
> hOj%2F79UtWa3Cc%3D
> >>rved=0
> >> ssl-sslhandshakeexcep
> >> tion-received-fatal-alert-handshake-failure/38264878#38264878
> >>
> >>
> >> I'm not sure if there are issues about what Java version is used by Ant.
> >>
> >> The logic around Ant installing Mac artifacts on Windows should operate
> >> differently with recent Ant versions.  Did that problem go away with
> >>your
> >> Ant upgrade?
> >>
> >> -Alex
> >>
> >> On 7/6/17, 7:43 PM, "Nicholas Kwiatkowski"
> >><nicholaskwiatkow...@gmail.com>
> >> wrote:
> >>
> >> >I remember reading about needing the latest ANT, so I installed 1.10.1
> >>and
> >> >1.9.9 on my Windows 10 machine with the latest Java 8 install and still
> >> >had
> >> >the SSL errors.
> >> >
> >> >I just read through the threads where where there was a discussion
> >>about
> >> >it.  There were three things talked about :
> >> > - Having Adobe move the files to another provider (not going to
> >>happen)
> >> > - Grabbing the entire Flex 4.6 SDK and pulling the files out of there
> >> >(eww)
> >> > - Re-creating Fontkit.  Chris Dutz said he might be able to work on
> >>it,
> >> >but I havent seen anything else to that affect.
> >> >
> >> >-Nick
> >> >
> >> >
> >> >On Thu, Jul 6, 2017 at 11:44 AM, Alex Harui <aha...@adobe.com.invalid>
> >> >wrote:
> >> >
> >> >> I'm not spending any time on the installer other than occasionally
> >> >> pondering if there is some other workaround we could deploy.  I'm
> >> >>waiting
> >> >&

Re: 4.16.0 Install issues

[Alex Harui]

What version of installer.xml are you looking at?  I just synced up the
develop branch in flex-sdk and took a look at there are merge conflict
markers in the file!  Apparently, nobody has tried to install a nightly
build of Flex SDK.  But I think the version from the release branch has
the right fix in it, although I think it is still using "new Ant" syntax
in the if/unless which the Installer does not support.  So whoever fixes
this needs to also replace:

  unless="${shouldUnzip}"

With:

  unless="shouldUnzip"


And same for 

  if="shouldUnzip"

-Alex


On 7/7/17, 3:57 AM, "Nicholas Kwiatkowski" <nicholaskwiatkow...@gmail.com>
wrote:

>I installed the latest version of Java, but didn't touch the JCE (I have
>to
>keep the standard set on my machine and extend, per project because of
>some
>of the other projects I'm working on).  Just using the net.url package
>within a quick java app seems grab the files from sourceforge with no
>problems -- it's the way that ANT is calling it that is causing the issue.
>It's quite possible they are overriding the standard security chain for
>some reason.  Either way, asking our users to change out the cryptology
>libraries and download them from another source isn't in good form.
>
>As far as the logic, there is clearly a bug.  It still bombs out on the
>latest ANT version.  We didn't do a conditional in the ant script between
>the call to unpack the dmg vs. zip files for the AIR install.  The mac one
>probably works because ANT on mac understands the unzip directive, but the
>other way complains that it doesn't know how to deal with a dmg file, so
>the ant script errors out.  That one is an easy fix.
>
>-Nick
>
>
>On Fri, Jul 7, 2017 at 1:54 AM, Alex Harui <aha...@adobe.com.invalid>
>wrote:
>
>> Maybe I wasn't clear enough:
>>
>> I don't think the version of Ant will affect the SSL issues.  The SSL
>> issues should be solvable via the JCE upgrade:
>>
>> 
>>https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fstackover
>>flow.com%2Fquestions%2F38203971%2Fjavax-net-=02%7C01%7C%7Cfe4cee1f24
>>f849200a1108d4c526ee4e%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C63635
>>0218379468309=3ngz8fD86odTrjJ5D8G1SKFu1j7H7qhOj%2F79UtWa3Cc%3D
>>rved=0
>> ssl-sslhandshakeexcep
>> tion-received-fatal-alert-handshake-failure/38264878#38264878
>>
>>
>> I'm not sure if there are issues about what Java version is used by Ant.
>>
>> The logic around Ant installing Mac artifacts on Windows should operate
>> differently with recent Ant versions.  Did that problem go away with
>>your
>> Ant upgrade?
>>
>> -Alex
>>
>> On 7/6/17, 7:43 PM, "Nicholas Kwiatkowski"
>><nicholaskwiatkow...@gmail.com>
>> wrote:
>>
>> >I remember reading about needing the latest ANT, so I installed 1.10.1
>>and
>> >1.9.9 on my Windows 10 machine with the latest Java 8 install and still
>> >had
>> >the SSL errors.
>> >
>> >I just read through the threads where where there was a discussion
>>about
>> >it.  There were three things talked about :
>> > - Having Adobe move the files to another provider (not going to
>>happen)
>> > - Grabbing the entire Flex 4.6 SDK and pulling the files out of there
>> >(eww)
>> > - Re-creating Fontkit.  Chris Dutz said he might be able to work on
>>it,
>> >but I havent seen anything else to that affect.
>> >
>> >-Nick
>> >
>> >
>> >On Thu, Jul 6, 2017 at 11:44 AM, Alex Harui <aha...@adobe.com.invalid>
>> >wrote:
>> >
>> >> I'm not spending any time on the installer other than occasionally
>> >> pondering if there is some other workaround we could deploy.  I'm
>> >>waiting
>> >> on Adobe to put out a 64-bit native installer.  IMO, that's the least
>> >>work
>> >> on our part, but not sure when Adobe will push that out.
>> >>
>> >> The SSL errors from SourceForge have a workaround that's been posted
>>a
>> >>few
>> >> times.  The correct solution is to write our own font encoding
>>library.
>> >> Volunteers are needed.
>> >>
>> >> I thought Ant on Windows would work if you used a recent version of
>>Ant.
>> >>
>> >> Thanks,
>> >> -Alex
>> >>
>> >> On 7/5/17, 6:13 PM, "Nicholas Kwiatkowski" <que...@apache.org> wrote:
>> >>
>> >> >Is anybody actually addressing the issues people have been reporting
>

Re: 4.16.0 Install issues

[Nicholas Kwiatkowski]

I installed the latest version of Java, but didn't touch the JCE (I have to
keep the standard set on my machine and extend, per project because of some
of the other projects I'm working on).  Just using the net.url package
within a quick java app seems grab the files from sourceforge with no
problems -- it's the way that ANT is calling it that is causing the issue.
It's quite possible they are overriding the standard security chain for
some reason.  Either way, asking our users to change out the cryptology
libraries and download them from another source isn't in good form.

As far as the logic, there is clearly a bug.  It still bombs out on the
latest ANT version.  We didn't do a conditional in the ant script between
the call to unpack the dmg vs. zip files for the AIR install.  The mac one
probably works because ANT on mac understands the unzip directive, but the
other way complains that it doesn't know how to deal with a dmg file, so
the ant script errors out.  That one is an easy fix.

-Nick


On Fri, Jul 7, 2017 at 1:54 AM, Alex Harui <aha...@adobe.com.invalid> wrote:

> Maybe I wasn't clear enough:
>
> I don't think the version of Ant will affect the SSL issues.  The SSL
> issues should be solvable via the JCE upgrade:
>
> http://stackoverflow.com/questions/38203971/javax-net-
> ssl-sslhandshakeexcep
> tion-received-fatal-alert-handshake-failure/38264878#38264878
>
>
> I'm not sure if there are issues about what Java version is used by Ant.
>
> The logic around Ant installing Mac artifacts on Windows should operate
> differently with recent Ant versions.  Did that problem go away with your
> Ant upgrade?
>
> -Alex
>
> On 7/6/17, 7:43 PM, "Nicholas Kwiatkowski" <nicholaskwiatkow...@gmail.com>
> wrote:
>
> >I remember reading about needing the latest ANT, so I installed 1.10.1 and
> >1.9.9 on my Windows 10 machine with the latest Java 8 install and still
> >had
> >the SSL errors.
> >
> >I just read through the threads where where there was a discussion about
> >it.  There were three things talked about :
> > - Having Adobe move the files to another provider (not going to happen)
> > - Grabbing the entire Flex 4.6 SDK and pulling the files out of there
> >(eww)
> > - Re-creating Fontkit.  Chris Dutz said he might be able to work on it,
> >but I havent seen anything else to that affect.
> >
> >-Nick
> >
> >
> >On Thu, Jul 6, 2017 at 11:44 AM, Alex Harui <aha...@adobe.com.invalid>
> >wrote:
> >
> >> I'm not spending any time on the installer other than occasionally
> >> pondering if there is some other workaround we could deploy.  I'm
> >>waiting
> >> on Adobe to put out a 64-bit native installer.  IMO, that's the least
> >>work
> >> on our part, but not sure when Adobe will push that out.
> >>
> >> The SSL errors from SourceForge have a workaround that's been posted a
> >>few
> >> times.  The correct solution is to write our own font encoding library.
> >> Volunteers are needed.
> >>
> >> I thought Ant on Windows would work if you used a recent version of Ant.
> >>
> >> Thanks,
> >> -Alex
> >>
> >> On 7/5/17, 6:13 PM, "Nicholas Kwiatkowski" <que...@apache.org> wrote:
> >>
> >> >Is anybody actually addressing the issues people have been reporting
> >>about
> >> >the installer and/or ANT script for Flex SDK 4.16.0?
> >> >
> >> >Right now I see two major issues that are preventing even people who
> >>are
> >> >familiar with the SDK from doing installs :
> >> >
> >> > - In the installer, selecting AIR 25.0 gives users a non-descript
> >>"error
> >> >1000".  This is due toe the md5 checking in the installer running out
> >>of
> >> >memory
> >> > - Trying to install via ANT is also broken under Windows (any version
> >>of
> >> >AIR SDK).  As packaged, it always errors because it tries to install
> >>the
> >> >MacOS AIR SDK.  Additionally, the optional components that are
> >>currently
> >> >hosted on sourceforge fail to download due to some SSL errors (I've
> >>tested
> >> >this with the latest java sdk and ANT build).
> >> >
> >> >The AIR installer issue will require us to rip-and-replace the md5
> >> >calculation functions.  I've started looking at it, but I don't think
> >>it
> >> >will be an easy feat.
> >> >Fixing the ANT script for Windows trying to install the mac air dmg is
> >>an
>

Re: [FlexJS][VS Code] dual issues

[Harbs]

Sounds great! Looking forward!

Harbs

> On Jul 7, 2017, at 2:12 AM, Josh Tynjala <joshtynj...@gmail.com> wrote:
> 
> Okay, I figured out how to make VSCode switch to JS APIs instead of SWF
> APIs for completion and things. In the next version of the extension, when
> the targets compiler option is specified in asconfig.json, and the first
> value isn't "SWF", the extension will offer JS APIs. That includes giving
> you access to native browser APIs instead of Flash Player APIs. That's
> going to be a really nice change, I think. Thanks for pointing out this
> issue!
> 
> - Josh
> 
> On Thu, Jul 6, 2017 at 2:35 PM, Josh Tynjala <joshtynj...@gmail.com> wrote:
> 
>>> The js-output-type, library-path and external-library-path options are
>> probably not needed
>> 
>> I would get rid of -js-output-type. I assume that -targets takes
>> precedence, but just to be safe, you shouldn't use both. -js-output-type is
>> necessary for 0.7.0, but switch entirely to -targets with 0.8.0.
>> 
>>> It would be nice to have full support for the new compiler options
>> 
>> html-template, html-output-filename, js-compiler-option,
>> js-external-library-path, js-library-path, swf-external-library-path,
>> swf-library-path, and remove-circulars should all work in the
>> compilerOptions field if you're using the latest vscode-nextgenas 0.4.4 and
>> asconfigc 0.5.1. I just updated asconfigc to support these options last
>> week, and I haven't made an official announcement yet, so you might have
>> missed that one.
>> 
>> I think there may be one or two you'll still need to put into
>> additionalOptions, but you should be able to switch over most of them.
>> 
>>> 1. I’m getting “problems” even for files which are not used in my
>> project and are not open.
>> 
>> This is something I'm aware of in VSCode. The compiler seems to greedily
>> load everything that exists in a source path, even if you don't actually
>> use it. Personally I like this, but even if I were to make it an option
>> somehow, I'm not aware of a way to stop the compiler from working that way
>> either.
>> 
>>> 2. I’m getting lots of errors which are specific to SWF-only output:
>> 
>> I just glanced through the part of VSCode that passes options to the
>> compiler for code intelligence (different from the way that asconfigc runs
>> the compiler to build), and I don't think I'm actually using the value of
>> -targets. It's probably defaulting to SWF. I'll play around with that to
>> see if I can improve things when -targets is specifically JSFlex.
>> 
>>>  a. I’m not getting JS specific classes recognized.
>> 
>> Just to be clear, are you referring to DOM classes like HTMLButtonElement
>> and other browser APIs here? I just want to be sure so that I test the
>> right thing when I make changes.
>> 
>> - Josh
>> 
>> On Thu, Jul 6, 2017 at 2:16 PM, Harbs <harbs.li...@gmail.com> wrote:
>> 
>>> Since updating to “dual”has been reporting lots of errors. When I build,
>>> both the debug and release builds build correctly, but the PROBLEMS window
>>> and the live code hinting report lots of (non) issues. I’m not sure what
>>> the source of the errors are, but here are the details:
>>> 
>>> I’ve tried a few variations of the asconfig file, but here’s what I
>>> currently have:
>>> {
>>>"config": "flex",
>>>"compilerOptions": {
>>>"debug": true,
>>>"js-output-type": "flexjs",
>>>"targets": ["JSFlex"],
>>>"source-map": false,
>>>"library-path": [
>>>"lib"
>>>],
>>>"external-library-path": [
>>>"typedefs"
>>>]
>>>},
>>>"copySourcePathAssets": true,
>>>"additionalOptions": "-remove-circulars 
>>> -js-output-optimization=skipAsCoercions
>>> -html-template=src/resources/mdl-js-index-template.html
>>> -js-external-library-path+=typedefs -js-library-path+=lib",
>>>"files":
>>>[
>>>"src/PortedPrintUI.mxml"
>>>]
>>> }
>>> 
>>> The js-output-type, library-path and external-library-path options are
>>> probably not needed, but they don’t seem to adversely effect the builds.
>>> (It would be nice to have full support for the new compiler options.)
>>> 
>>> Problems:
>>> 1. I’m getting “problems” even for files which are not used in my project
>>> and are not open.
>>> 2. I’m getting lots of errors which are specific to SWF-only output:
>>>  a. I’m not getting JS specific classes recognized.
>>>  b. I’m getting lots of incompatible overrides (to Flash methods)
>>>  c. I’m getting overriding function not marked for override errors on
>>> functions that are implemented in flash (i.e. public function get
>>> scaleX():Number)
>>>  d. conflicts between variables and Flash getters.
>>>  e. etc...
>>> 
>>> Thanks,
>>> Harbs
>> 
>> 
>> 

Re: 4.16.0 Install issues

[piotrz]

It would be great if anyone who can will add to that jira any hint for
someone who will be working on that.

I have other things on my plate right now.

Thanks,
Piotr



-
Apache Flex PMC
piotrzarzyck...@gmail.com
--
View this message in context: 
http://apache-flex-development.247.n4.nabble.com/4-16-0-Install-issues-tp62855p62930.html
Sent from the Apache Flex Development mailing list archive at Nabble.com.

Re: [FlexJS][VS Code] dual issues

[piotrz]

Alex,

Do you mean by teaching compiler that if I use -target=JSFlex - I will not
see flash api hints in intellisence ?

Thanks,
Piotr



-
Apache Flex PMC
piotrzarzyck...@gmail.com
--
View this message in context: 
http://apache-flex-development.247.n4.nabble.com/FlexJS-VS-Code-dual-issues-tp62894p62929.html
Sent from the Apache Flex Development mailing list archive at Nabble.com.

Re: 4.16.0 Install issues

[Alex Harui]

Cool!  I missed seeing this JIRA.

I took a quick look and didn't see spline conversion handling between TTF
and CFF.  I think it may not handle specifying a TTF file for
embedAsCFF=true and vice versa.  But maybe we can live without it.

Any volunteers to take this on?

-Alex

On 7/6/17, 10:01 PM, "Olaf Krueger" <p...@olafkrueger.net> wrote:

>Nicholas Kwiatkowski-2 wrote
>>  - Re-creating Fontkit.  Chris Dutz said he might be able to work on it,
>> but I havent seen anything else to that affect.
>
>At ApacheCon somebody found a lib that maybe contains a working FontKit
>port.
>Piotr created a JIRA that contains the link [1]
>
>HTH,
>Olaf 
>
>[1]https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fissues
>.apache.org%2Fjira%2Fbrowse%2FFLEX-35315=02%7C01%7C%7C8b8ed67825f84e5
>d91a008d4c4f7c56b%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C63635001583
>1422210=ps7%2Fd3pmBt%2FY2pFquS9vvO5Dj2rJ8EohmcV7hfgU9F8%3D=
>0
>
>
>
>
>--
>View this message in context:
>https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fapache-fle
>x-development.247.n4.nabble.com%2F4-16-0-Install-issues-tp62855p62919.
>html=02%7C01%7C%7C8b8ed67825f84e5d91a008d4c4f7c56b%7Cfa7b1b5a7b344387
>94aed2c178decee1%7C0%7C0%7C636350015831422210=qtIXbmeRpuwC3ICJcyKsmG
>V%2FQ%2B%2FhG6jTxWBU1CO%2BJ4c%3D=0
>Sent from the Apache Flex Development mailing list archive at Nabble.com.

Re: 4.16.0 Install issues

[Alex Harui]

Maybe I wasn't clear enough:

I don't think the version of Ant will affect the SSL issues.  The SSL
issues should be solvable via the JCE upgrade:

http://stackoverflow.com/questions/38203971/javax-net-ssl-sslhandshakeexcep
tion-received-fatal-alert-handshake-failure/38264878#38264878


I'm not sure if there are issues about what Java version is used by Ant.

The logic around Ant installing Mac artifacts on Windows should operate
differently with recent Ant versions.  Did that problem go away with your
Ant upgrade?

-Alex

On 7/6/17, 7:43 PM, "Nicholas Kwiatkowski" <nicholaskwiatkow...@gmail.com>
wrote:

>I remember reading about needing the latest ANT, so I installed 1.10.1 and
>1.9.9 on my Windows 10 machine with the latest Java 8 install and still
>had
>the SSL errors.
>
>I just read through the threads where where there was a discussion about
>it.  There were three things talked about :
> - Having Adobe move the files to another provider (not going to happen)
> - Grabbing the entire Flex 4.6 SDK and pulling the files out of there
>(eww)
> - Re-creating Fontkit.  Chris Dutz said he might be able to work on it,
>but I havent seen anything else to that affect.
>
>-Nick
>
>
>On Thu, Jul 6, 2017 at 11:44 AM, Alex Harui <aha...@adobe.com.invalid>
>wrote:
>
>> I'm not spending any time on the installer other than occasionally
>> pondering if there is some other workaround we could deploy.  I'm
>>waiting
>> on Adobe to put out a 64-bit native installer.  IMO, that's the least
>>work
>> on our part, but not sure when Adobe will push that out.
>>
>> The SSL errors from SourceForge have a workaround that's been posted a
>>few
>> times.  The correct solution is to write our own font encoding library.
>> Volunteers are needed.
>>
>> I thought Ant on Windows would work if you used a recent version of Ant.
>>
>> Thanks,
>> -Alex
>>
>> On 7/5/17, 6:13 PM, "Nicholas Kwiatkowski" <que...@apache.org> wrote:
>>
>> >Is anybody actually addressing the issues people have been reporting
>>about
>> >the installer and/or ANT script for Flex SDK 4.16.0?
>> >
>> >Right now I see two major issues that are preventing even people who
>>are
>> >familiar with the SDK from doing installs :
>> >
>> > - In the installer, selecting AIR 25.0 gives users a non-descript
>>"error
>> >1000".  This is due toe the md5 checking in the installer running out
>>of
>> >memory
>> > - Trying to install via ANT is also broken under Windows (any version
>>of
>> >AIR SDK).  As packaged, it always errors because it tries to install
>>the
>> >MacOS AIR SDK.  Additionally, the optional components that are
>>currently
>> >hosted on sourceforge fail to download due to some SSL errors (I've
>>tested
>> >this with the latest java sdk and ANT build).
>> >
>> >The AIR installer issue will require us to rip-and-replace the md5
>> >calculation functions.  I've started looking at it, but I don't think
>>it
>> >will be an easy feat.
>> >Fixing the ANT script for Windows trying to install the mac air dmg is
>>an
>> >easy fix (but it will require us to do a dot release to push it out)
>> >The SSL errors are because sourceforge is using SANs on their SSL
>>certs,
>> >and the current versions of ANT don't know how to read them to validate
>> >them.  This may be out of our control.  Anybody know if we can convince
>> >Adobe to either donate those chunks of code or at least to move them
>>to a
>> >different host?  It looks like the code involved is OSMF, AFE, AGLJ,
>> >rideau and Flex-Fontkit.  The alternative to a different host is for
>>us to
>> >ignore SSL errors, but that could be potentially dangerous.
>> >
>> >If nobody is working on these, I can start to take a crack at them, but
>> >honestly, the installer is extremely fragile at this point and I'm not
>> >looking forward to even trying to figure out what is going on in there
>> >again.
>> >
>> >-Nick
>>
>>

Re: 4.16.0 Install issues

[Olaf Krueger]

Nicholas Kwiatkowski-2 wrote
>  - Re-creating Fontkit.  Chris Dutz said he might be able to work on it,
> but I havent seen anything else to that affect. 

At ApacheCon somebody found a lib that maybe contains a working FontKit
port.
Piotr created a JIRA that contains the link [1]

HTH,
Olaf 

[1]https://issues.apache.org/jira/browse/FLEX-35315




--
View this message in context: 
http://apache-flex-development.247.n4.nabble.com/4-16-0-Install-issues-tp62855p62919.html
Sent from the Apache Flex Development mailing list archive at Nabble.com.

Re: 4.16.0 Install issues

[Nicholas Kwiatkowski]

I remember reading about needing the latest ANT, so I installed 1.10.1 and
1.9.9 on my Windows 10 machine with the latest Java 8 install and still had
the SSL errors.

I just read through the threads where where there was a discussion about
it.  There were three things talked about :
 - Having Adobe move the files to another provider (not going to happen)
 - Grabbing the entire Flex 4.6 SDK and pulling the files out of there (eww)
 - Re-creating Fontkit.  Chris Dutz said he might be able to work on it,
but I havent seen anything else to that affect.

-Nick


On Thu, Jul 6, 2017 at 11:44 AM, Alex Harui <aha...@adobe.com.invalid>
wrote:

> I'm not spending any time on the installer other than occasionally
> pondering if there is some other workaround we could deploy.  I'm waiting
> on Adobe to put out a 64-bit native installer.  IMO, that's the least work
> on our part, but not sure when Adobe will push that out.
>
> The SSL errors from SourceForge have a workaround that's been posted a few
> times.  The correct solution is to write our own font encoding library.
> Volunteers are needed.
>
> I thought Ant on Windows would work if you used a recent version of Ant.
>
> Thanks,
> -Alex
>
> On 7/5/17, 6:13 PM, "Nicholas Kwiatkowski" <que...@apache.org> wrote:
>
> >Is anybody actually addressing the issues people have been reporting about
> >the installer and/or ANT script for Flex SDK 4.16.0?
> >
> >Right now I see two major issues that are preventing even people who are
> >familiar with the SDK from doing installs :
> >
> > - In the installer, selecting AIR 25.0 gives users a non-descript "error
> >1000".  This is due toe the md5 checking in the installer running out of
> >memory
> > - Trying to install via ANT is also broken under Windows (any version of
> >AIR SDK).  As packaged, it always errors because it tries to install the
> >MacOS AIR SDK.  Additionally, the optional components that are currently
> >hosted on sourceforge fail to download due to some SSL errors (I've tested
> >this with the latest java sdk and ANT build).
> >
> >The AIR installer issue will require us to rip-and-replace the md5
> >calculation functions.  I've started looking at it, but I don't think it
> >will be an easy feat.
> >Fixing the ANT script for Windows trying to install the mac air dmg is an
> >easy fix (but it will require us to do a dot release to push it out)
> >The SSL errors are because sourceforge is using SANs on their SSL certs,
> >and the current versions of ANT don't know how to read them to validate
> >them.  This may be out of our control.  Anybody know if we can convince
> >Adobe to either donate those chunks of code or at least to move them to a
> >different host?  It looks like the code involved is OSMF, AFE, AGLJ,
> >rideau and Flex-Fontkit.  The alternative to a different host is for us to
> >ignore SSL errors, but that could be potentially dangerous.
> >
> >If nobody is working on these, I can start to take a crack at them, but
> >honestly, the installer is extremely fragile at this point and I'm not
> >looking forward to even trying to figure out what is going on in there
> >again.
> >
> >-Nick
>
>

Re: [FlexJS][VS Code] dual issues

[Josh Tynjala]

The AS/MXML code intelligence in VSCode instantiates certain classes from
the compiler to get the data it needs. Previously, I was instantiating a
FlexProject no matter which targets were specified. Now, I instantiate a
FlexJSProject when I'm sure that JS should have a higher priority than SWF
(based on the value of -targets). This seems to be enough to get the
compiler to switch from Flash Player APIs to JS APIs in the data that it
provides, and i can see that it switches COMPILE::JS to true and COMPILE::
SWF to false.

With this change that I just made, I suspect that if we merged the pairs of
SWCs into one, VSCode might just work with no changes (except to tell it to
use the newer compiler).

- Josh


On Jul 6, 2017 4:23 PM, "Alex Harui" <aha...@adobe.com.invalid> wrote:

I agree, it will be great to get JS API code-intelligence.

Just out of curiosity, what did you have to change?  One work item for the
future is to merge the "JS" SWC into the "SWF" SWC so there is only one
SWC per library instead of the current pair like Core.SWC and CoreJS.swc.
I'm trying to assess the impact of doing so.  We can certainly teach the
compiler to pick different classes based on the current -target, but right
now for FB, all it knows is that one set of SWCs or the other is on the
library path.  If we merge the SWCS, there is no other way to tell the
compiler which APIs to use by default, and I'm not sure I can control that
from outside FB.

Maybe just the Maven artifacts will be merged and we'll keep the pairs of
SWCs for IDEs.

Thoughts?
-Alex

On 7/6/17, 4:12 PM, "Josh Tynjala" <joshtynj...@gmail.com> wrote:

>Okay, I figured out how to make VSCode switch to JS APIs instead of SWF
>APIs for completion and things. In the next version of the extension, when
>the targets compiler option is specified in asconfig.json, and the first
>value isn't "SWF", the extension will offer JS APIs. That includes giving
>you access to native browser APIs instead of Flash Player APIs. That's
>going to be a really nice change, I think. Thanks for pointing out this
>issue!
>
>- Josh
>
>On Thu, Jul 6, 2017 at 2:35 PM, Josh Tynjala <joshtynj...@gmail.com>
>wrote:
>
>> > The js-output-type, library-path and external-library-path options are
>> probably not needed
>>
>> I would get rid of -js-output-type. I assume that -targets takes
>> precedence, but just to be safe, you shouldn't use both.
>>-js-output-type is
>> necessary for 0.7.0, but switch entirely to -targets with 0.8.0.
>>
>> > It would be nice to have full support for the new compiler options
>>
>> html-template, html-output-filename, js-compiler-option,
>> js-external-library-path, js-library-path, swf-external-library-path,
>> swf-library-path, and remove-circulars should all work in the
>> compilerOptions field if you're using the latest vscode-nextgenas 0.4.4
>>and
>> asconfigc 0.5.1. I just updated asconfigc to support these options last
>> week, and I haven't made an official announcement yet, so you might have
>> missed that one.
>>
>> I think there may be one or two you'll still need to put into
>> additionalOptions, but you should be able to switch over most of them.
>>
>> > 1. I’m getting “problems” even for files which are not used in my
>> project and are not open.
>>
>> This is something I'm aware of in VSCode. The compiler seems to greedily
>> load everything that exists in a source path, even if you don't actually
>> use it. Personally I like this, but even if I were to make it an option
>> somehow, I'm not aware of a way to stop the compiler from working that
>>way
>> either.
>>
>> > 2. I’m getting lots of errors which are specific to SWF-only output:
>>
>> I just glanced through the part of VSCode that passes options to the
>> compiler for code intelligence (different from the way that asconfigc
>>runs
>> the compiler to build), and I don't think I'm actually using the value
>>of
>> -targets. It's probably defaulting to SWF. I'll play around with that to
>> see if I can improve things when -targets is specifically JSFlex.
>>
>> >   a. I’m not getting JS specific classes recognized.
>>
>> Just to be clear, are you referring to DOM classes like
>>HTMLButtonElement
>> and other browser APIs here? I just want to be sure so that I test the
>> right thing when I make changes.
>>
>> - Josh
>>
>> On Thu, Jul 6, 2017 at 2:16 PM, Harbs <harbs.li...@gmail.com> wrote:
>>
>>> Since updating to “dual”has been reporting lots of errors. When I
>>>build,
>>> both the debug and release builds build correctly, but th

Re: [FlexJS][VS Code] dual issues

[Alex Harui]

I agree, it will be great to get JS API code-intelligence.

Just out of curiosity, what did you have to change?  One work item for the
future is to merge the "JS" SWC into the "SWF" SWC so there is only one
SWC per library instead of the current pair like Core.SWC and CoreJS.swc.
I'm trying to assess the impact of doing so.  We can certainly teach the
compiler to pick different classes based on the current -target, but right
now for FB, all it knows is that one set of SWCs or the other is on the
library path.  If we merge the SWCS, there is no other way to tell the
compiler which APIs to use by default, and I'm not sure I can control that
from outside FB.

Maybe just the Maven artifacts will be merged and we'll keep the pairs of
SWCs for IDEs.

Thoughts?
-Alex

On 7/6/17, 4:12 PM, "Josh Tynjala" <joshtynj...@gmail.com> wrote:

>Okay, I figured out how to make VSCode switch to JS APIs instead of SWF
>APIs for completion and things. In the next version of the extension, when
>the targets compiler option is specified in asconfig.json, and the first
>value isn't "SWF", the extension will offer JS APIs. That includes giving
>you access to native browser APIs instead of Flash Player APIs. That's
>going to be a really nice change, I think. Thanks for pointing out this
>issue!
>
>- Josh
>
>On Thu, Jul 6, 2017 at 2:35 PM, Josh Tynjala <joshtynj...@gmail.com>
>wrote:
>
>> > The js-output-type, library-path and external-library-path options are
>> probably not needed
>>
>> I would get rid of -js-output-type. I assume that -targets takes
>> precedence, but just to be safe, you shouldn't use both.
>>-js-output-type is
>> necessary for 0.7.0, but switch entirely to -targets with 0.8.0.
>>
>> > It would be nice to have full support for the new compiler options
>>
>> html-template, html-output-filename, js-compiler-option,
>> js-external-library-path, js-library-path, swf-external-library-path,
>> swf-library-path, and remove-circulars should all work in the
>> compilerOptions field if you're using the latest vscode-nextgenas 0.4.4
>>and
>> asconfigc 0.5.1. I just updated asconfigc to support these options last
>> week, and I haven't made an official announcement yet, so you might have
>> missed that one.
>>
>> I think there may be one or two you'll still need to put into
>> additionalOptions, but you should be able to switch over most of them.
>>
>> > 1. I’m getting “problems” even for files which are not used in my
>> project and are not open.
>>
>> This is something I'm aware of in VSCode. The compiler seems to greedily
>> load everything that exists in a source path, even if you don't actually
>> use it. Personally I like this, but even if I were to make it an option
>> somehow, I'm not aware of a way to stop the compiler from working that
>>way
>> either.
>>
>> > 2. I’m getting lots of errors which are specific to SWF-only output:
>>
>> I just glanced through the part of VSCode that passes options to the
>> compiler for code intelligence (different from the way that asconfigc
>>runs
>> the compiler to build), and I don't think I'm actually using the value
>>of
>> -targets. It's probably defaulting to SWF. I'll play around with that to
>> see if I can improve things when -targets is specifically JSFlex.
>>
>> >   a. I’m not getting JS specific classes recognized.
>>
>> Just to be clear, are you referring to DOM classes like
>>HTMLButtonElement
>> and other browser APIs here? I just want to be sure so that I test the
>> right thing when I make changes.
>>
>> - Josh
>>
>> On Thu, Jul 6, 2017 at 2:16 PM, Harbs <harbs.li...@gmail.com> wrote:
>>
>>> Since updating to “dual”has been reporting lots of errors. When I
>>>build,
>>> both the debug and release builds build correctly, but the PROBLEMS
>>>window
>>> and the live code hinting report lots of (non) issues. I’m not sure
>>>what
>>> the source of the errors are, but here are the details:
>>>
>>> I’ve tried a few variations of the asconfig file, but here’s what I
>>> currently have:
>>> {
>>> "config": "flex",
>>> "compilerOptions": {
>>> "debug": true,
>>> "js-output-type": "flexjs",
>>> "targets": ["JSFlex"],
>>> "source-map": false,
>>> "library-path": [
>>> "lib"
>>> ],
>>>

Re: [FlexJS][VS Code] dual issues

[Josh Tynjala]

Okay, I figured out how to make VSCode switch to JS APIs instead of SWF
APIs for completion and things. In the next version of the extension, when
the targets compiler option is specified in asconfig.json, and the first
value isn't "SWF", the extension will offer JS APIs. That includes giving
you access to native browser APIs instead of Flash Player APIs. That's
going to be a really nice change, I think. Thanks for pointing out this
issue!

- Josh

On Thu, Jul 6, 2017 at 2:35 PM, Josh Tynjala <joshtynj...@gmail.com> wrote:

> > The js-output-type, library-path and external-library-path options are
> probably not needed
>
> I would get rid of -js-output-type. I assume that -targets takes
> precedence, but just to be safe, you shouldn't use both. -js-output-type is
> necessary for 0.7.0, but switch entirely to -targets with 0.8.0.
>
> > It would be nice to have full support for the new compiler options
>
> html-template, html-output-filename, js-compiler-option,
> js-external-library-path, js-library-path, swf-external-library-path,
> swf-library-path, and remove-circulars should all work in the
> compilerOptions field if you're using the latest vscode-nextgenas 0.4.4 and
> asconfigc 0.5.1. I just updated asconfigc to support these options last
> week, and I haven't made an official announcement yet, so you might have
> missed that one.
>
> I think there may be one or two you'll still need to put into
> additionalOptions, but you should be able to switch over most of them.
>
> > 1. I’m getting “problems” even for files which are not used in my
> project and are not open.
>
> This is something I'm aware of in VSCode. The compiler seems to greedily
> load everything that exists in a source path, even if you don't actually
> use it. Personally I like this, but even if I were to make it an option
> somehow, I'm not aware of a way to stop the compiler from working that way
> either.
>
> > 2. I’m getting lots of errors which are specific to SWF-only output:
>
> I just glanced through the part of VSCode that passes options to the
> compiler for code intelligence (different from the way that asconfigc runs
> the compiler to build), and I don't think I'm actually using the value of
> -targets. It's probably defaulting to SWF. I'll play around with that to
> see if I can improve things when -targets is specifically JSFlex.
>
> >   a. I’m not getting JS specific classes recognized.
>
> Just to be clear, are you referring to DOM classes like HTMLButtonElement
> and other browser APIs here? I just want to be sure so that I test the
> right thing when I make changes.
>
> - Josh
>
> On Thu, Jul 6, 2017 at 2:16 PM, Harbs <harbs.li...@gmail.com> wrote:
>
>> Since updating to “dual”has been reporting lots of errors. When I build,
>> both the debug and release builds build correctly, but the PROBLEMS window
>> and the live code hinting report lots of (non) issues. I’m not sure what
>> the source of the errors are, but here are the details:
>>
>> I’ve tried a few variations of the asconfig file, but here’s what I
>> currently have:
>> {
>> "config": "flex",
>> "compilerOptions": {
>> "debug": true,
>> "js-output-type": "flexjs",
>> "targets": ["JSFlex"],
>> "source-map": false,
>> "library-path": [
>> "lib"
>> ],
>> "external-library-path": [
>> "typedefs"
>> ]
>> },
>> "copySourcePathAssets": true,
>> "additionalOptions": "-remove-circulars 
>> -js-output-optimization=skipAsCoercions
>> -html-template=src/resources/mdl-js-index-template.html
>> -js-external-library-path+=typedefs -js-library-path+=lib",
>> "files":
>> [
>> "src/PortedPrintUI.mxml"
>> ]
>> }
>>
>> The js-output-type, library-path and external-library-path options are
>> probably not needed, but they don’t seem to adversely effect the builds.
>> (It would be nice to have full support for the new compiler options.)
>>
>> Problems:
>> 1. I’m getting “problems” even for files which are not used in my project
>> and are not open.
>> 2. I’m getting lots of errors which are specific to SWF-only output:
>>   a. I’m not getting JS specific classes recognized.
>>   b. I’m getting lots of incompatible overrides (to Flash methods)
>>   c. I’m getting overriding function not marked for override errors on
>> functions that are implemented in flash (i.e. public function get
>> scaleX():Number)
>>   d. conflicts between variables and Flash getters.
>>   e. etc...
>>
>> Thanks,
>> Harbs
>
>
>

Re: [FlexJS][VS Code] dual issues

[Josh Tynjala]

> The js-output-type, library-path and external-library-path options are
probably not needed

I would get rid of -js-output-type. I assume that -targets takes
precedence, but just to be safe, you shouldn't use both. -js-output-type is
necessary for 0.7.0, but switch entirely to -targets with 0.8.0.

> It would be nice to have full support for the new compiler options

html-template, html-output-filename, js-compiler-option,
js-external-library-path, js-library-path, swf-external-library-path,
swf-library-path, and remove-circulars should all work in the
compilerOptions field if you're using the latest vscode-nextgenas 0.4.4 and
asconfigc 0.5.1. I just updated asconfigc to support these options last
week, and I haven't made an official announcement yet, so you might have
missed that one.

I think there may be one or two you'll still need to put into
additionalOptions, but you should be able to switch over most of them.

> 1. I’m getting “problems” even for files which are not used in my project
and are not open.

This is something I'm aware of in VSCode. The compiler seems to greedily
load everything that exists in a source path, even if you don't actually
use it. Personally I like this, but even if I were to make it an option
somehow, I'm not aware of a way to stop the compiler from working that way
either.

> 2. I’m getting lots of errors which are specific to SWF-only output:

I just glanced through the part of VSCode that passes options to the
compiler for code intelligence (different from the way that asconfigc runs
the compiler to build), and I don't think I'm actually using the value of
-targets. It's probably defaulting to SWF. I'll play around with that to
see if I can improve things when -targets is specifically JSFlex.

>   a. I’m not getting JS specific classes recognized.

Just to be clear, are you referring to DOM classes like HTMLButtonElement
and other browser APIs here? I just want to be sure so that I test the
right thing when I make changes.

- Josh

On Thu, Jul 6, 2017 at 2:16 PM, Harbs <harbs.li...@gmail.com> wrote:

> Since updating to “dual”has been reporting lots of errors. When I build,
> both the debug and release builds build correctly, but the PROBLEMS window
> and the live code hinting report lots of (non) issues. I’m not sure what
> the source of the errors are, but here are the details:
>
> I’ve tried a few variations of the asconfig file, but here’s what I
> currently have:
> {
> "config": "flex",
> "compilerOptions": {
> "debug": true,
> "js-output-type": "flexjs",
> "targets": ["JSFlex"],
> "source-map": false,
> "library-path": [
> "lib"
> ],
> "external-library-path": [
> "typedefs"
> ]
> },
> "copySourcePathAssets": true,
> "additionalOptions": "-remove-circulars 
> -js-output-optimization=skipAsCoercions
> -html-template=src/resources/mdl-js-index-template.html
> -js-external-library-path+=typedefs -js-library-path+=lib",
> "files":
> [
> "src/PortedPrintUI.mxml"
> ]
> }
>
> The js-output-type, library-path and external-library-path options are
> probably not needed, but they don’t seem to adversely effect the builds.
> (It would be nice to have full support for the new compiler options.)
>
> Problems:
> 1. I’m getting “problems” even for files which are not used in my project
> and are not open.
> 2. I’m getting lots of errors which are specific to SWF-only output:
>   a. I’m not getting JS specific classes recognized.
>   b. I’m getting lots of incompatible overrides (to Flash methods)
>   c. I’m getting overriding function not marked for override errors on
> functions that are implemented in flash (i.e. public function get
> scaleX():Number)
>   d. conflicts between variables and Flash getters.
>   e. etc...
>
> Thanks,
> Harbs

[FlexJS][VS Code] dual issues

[Harbs]

Since updating to “dual”has been reporting lots of errors. When I build, both 
the debug and release builds build correctly, but the PROBLEMS window and the 
live code hinting report lots of (non) issues. I’m not sure what the source of 
the errors are, but here are the details:

I’ve tried a few variations of the asconfig file, but here’s what I currently 
have:
{
"config": "flex",
"compilerOptions": {
"debug": true,
"js-output-type": "flexjs",
"targets": ["JSFlex"],
"source-map": false,
"library-path": [
"lib"
],
"external-library-path": [
"typedefs"
]
},
"copySourcePathAssets": true,
"additionalOptions": "-remove-circulars 
-js-output-optimization=skipAsCoercions 
-html-template=src/resources/mdl-js-index-template.html 
-js-external-library-path+=typedefs -js-library-path+=lib",
"files":
[
"src/PortedPrintUI.mxml"
]
}

The js-output-type, library-path and external-library-path options are probably 
not needed, but they don’t seem to adversely effect the builds. (It would be 
nice to have full support for the new compiler options.)

Problems:
1. I’m getting “problems” even for files which are not used in my project and 
are not open.
2. I’m getting lots of errors which are specific to SWF-only output:
  a. I’m not getting JS specific classes recognized.
  b. I’m getting lots of incompatible overrides (to Flash methods)
  c. I’m getting overriding function not marked for override errors on 
functions that are implemented in flash (i.e. public function get 
scaleX():Number)
  d. conflicts between variables and Flash getters.
  e. etc...

Thanks,
Harbs

Re: 4.16.0 Install issues

[piotrz]

Hi Nick,

As for the font encoding we had discussion about that at ApacheCon [1] and I
raised jira [2]

[1]
http://apache-flex-development.247.n4.nabble.com/ApacheCon-FlexJS-Summit-FlexJS-1-0-Discussion-Round-Summary-td61700.html
[2] https://issues.apache.org/jira/browse/FLEX-35315

Piotr



-
Apache Flex PMC
piotrzarzyck...@gmail.com
--
View this message in context: 
http://apache-flex-development.247.n4.nabble.com/4-16-0-Install-issues-tp62855p62869.html
Sent from the Apache Flex Development mailing list archive at Nabble.com.

Re: 4.16.0 Install issues

[Alex Harui]

I'm not spending any time on the installer other than occasionally
pondering if there is some other workaround we could deploy.  I'm waiting
on Adobe to put out a 64-bit native installer.  IMO, that's the least work
on our part, but not sure when Adobe will push that out.

The SSL errors from SourceForge have a workaround that's been posted a few
times.  The correct solution is to write our own font encoding library.
Volunteers are needed.

I thought Ant on Windows would work if you used a recent version of Ant.

Thanks,
-Alex

On 7/5/17, 6:13 PM, "Nicholas Kwiatkowski" <que...@apache.org> wrote:

>Is anybody actually addressing the issues people have been reporting about
>the installer and/or ANT script for Flex SDK 4.16.0?
>
>Right now I see two major issues that are preventing even people who are
>familiar with the SDK from doing installs :
>
> - In the installer, selecting AIR 25.0 gives users a non-descript "error
>1000".  This is due toe the md5 checking in the installer running out of
>memory
> - Trying to install via ANT is also broken under Windows (any version of
>AIR SDK).  As packaged, it always errors because it tries to install the
>MacOS AIR SDK.  Additionally, the optional components that are currently
>hosted on sourceforge fail to download due to some SSL errors (I've tested
>this with the latest java sdk and ANT build).
>
>The AIR installer issue will require us to rip-and-replace the md5
>calculation functions.  I've started looking at it, but I don't think it
>will be an easy feat.
>Fixing the ANT script for Windows trying to install the mac air dmg is an
>easy fix (but it will require us to do a dot release to push it out)
>The SSL errors are because sourceforge is using SANs on their SSL certs,
>and the current versions of ANT don't know how to read them to validate
>them.  This may be out of our control.  Anybody know if we can convince
>Adobe to either donate those chunks of code or at least to move them to a
>different host?  It looks like the code involved is OSMF, AFE, AGLJ,
>rideau and Flex-Fontkit.  The alternative to a different host is for us to
>ignore SSL errors, but that could be potentially dangerous.
>
>If nobody is working on these, I can start to take a crack at them, but
>honestly, the installer is extremely fragile at this point and I'm not
>looking forward to even trying to figure out what is going on in there
>again.
>
>-Nick

4.16.0 Install issues

[Nicholas Kwiatkowski]

Is anybody actually addressing the issues people have been reporting about
the installer and/or ANT script for Flex SDK 4.16.0?

Right now I see two major issues that are preventing even people who are
familiar with the SDK from doing installs :

 - In the installer, selecting AIR 25.0 gives users a non-descript "error
1000".  This is due toe the md5 checking in the installer running out of
memory
 - Trying to install via ANT is also broken under Windows (any version of
AIR SDK).  As packaged, it always errors because it tries to install the
MacOS AIR SDK.  Additionally, the optional components that are currently
hosted on sourceforge fail to download due to some SSL errors (I've tested
this with the latest java sdk and ANT build).

The AIR installer issue will require us to rip-and-replace the md5
calculation functions.  I've started looking at it, but I don't think it
will be an easy feat.
Fixing the ANT script for Windows trying to install the mac air dmg is an
easy fix (but it will require us to do a dot release to push it out)
The SSL errors are because sourceforge is using SANs on their SSL certs,
and the current versions of ANT don't know how to read them to validate
them.  This may be out of our control.  Anybody know if we can convince
Adobe to either donate those chunks of code or at least to move them to a
different host?  It looks like the code involved is OSMF, AFE, AGLJ,
rideau and Flex-Fontkit.  The alternative to a different host is for us to
ignore SSL errors, but that could be potentially dangerous.

If nobody is working on these, I can start to take a crack at them, but
honestly, the installer is extremely fragile at this point and I'm not
looking forward to even trying to figure out what is going on in there
again.

-Nick

Re: [FLEXJS] ASDoc Issues

[Peter Ent]

Ah, I misunderstood which asdoc-config file.

On 6/30/17, 1:31 PM, "Alex Harui"  wrote:

>The asdoc target doesn't work because we've extended ActionScript for
>FlexJS.  We only care about the ASDoc Example.  I guess we should comment
>out the asdoc target for now.  Also, it looks like the ASDoc example uses
>its own asdic-config.xml file in its src/main/config folder.
>
>-Alex
>
>On 6/30/17, 10:25 AM, "Peter Ent"  wrote:
>
>>I'm running the asdoc target in flex-asjs and it has rejected a couple of
>>things in Core/main/flex/Promise.as such as:
>>
>>[asdoc] 
>>/Users/pent/dev/flex-asjs/frameworks/projects/Core/src/main/flex/Promise.
>>a
>>s(270): col: 19 Error: Syntax error: expecting identifier before catch.
>>[asdoc]
>>[asdoc] public function catch(onRejected:Function = null):IThenable
>>
>>Since "catch" is a keyword I'm surprised this compiles at all. Should
>>Promise.as be excluded from ASDoc for now?
>>
>>‹peter
>

Re: [FLEXJS] ASDoc Issues

[Alex Harui]

The asdoc target doesn't work because we've extended ActionScript for
FlexJS.  We only care about the ASDoc Example.  I guess we should comment
out the asdoc target for now.  Also, it looks like the ASDoc example uses
its own asdic-config.xml file in its src/main/config folder.

-Alex

On 6/30/17, 10:25 AM, "Peter Ent"  wrote:

>I'm running the asdoc target in flex-asjs and it has rejected a couple of
>things in Core/main/flex/Promise.as such as:
>
>[asdoc] 
>/Users/pent/dev/flex-asjs/frameworks/projects/Core/src/main/flex/Promise.a
>s(270): col: 19 Error: Syntax error: expecting identifier before catch.
>[asdoc]
>[asdoc] public function catch(onRejected:Function = null):IThenable
>
>Since "catch" is a keyword I'm surprised this compiles at all. Should
>Promise.as be excluded from ASDoc for now?
>
>—peter

[FLEXJS] ASDoc Issues

[Peter Ent]

I'm running the asdoc target in flex-asjs and it has rejected a couple of 
things in Core/main/flex/Promise.as such as:

[asdoc] 
/Users/pent/dev/flex-asjs/frameworks/projects/Core/src/main/flex/Promise.as(270):
 col: 19 Error: Syntax error: expecting identifier before catch.
[asdoc]
[asdoc] public function catch(onRejected:Function = null):IThenable

Since "catch" is a keyword I'm surprised this compiles at all. Should 
Promise.as be excluded from ASDoc for now?

—peter

Re: [FlexJS] [Typedefs] [Maven] build issues

[Christofer Dutz]

The only 100% safe option in my opinion would be to create a tool to do the 
patching. This tool might simply be a wrapper around some open-source patch 
tool with a suitable license and provide a Maven goal to do the patching. For 
Ant a matching wrapper should make sure the patches are applied equally in Ant 
and Maven.

Chris

Am 07.06.17, 02:18 schrieb "Justin Mclean" :

Hi,

> I think others may have experienced this also, at least I have seen 
similar things mentioned previously.

I’ve run into this issue before on OSX and it been discussed on list - was 
not able to find the root cause however.

Thanks,
Justin

AW: [FlexJS] [Typedefs] [Maven] build issues

[Christofer Dutz]

Nope … don’t have any issues on a Mac.

I wouldn’t say that I am really happy with the patching and all, the way it’s 
done today. We are relying on an external executable. As you might have 
noticed, I don’t like executing Things outside the VM and like to have Things 
inside the Java world and hereby under full control. I wouldn’t bet my life on 
using the git.exe on Windows to do the patching to work identically on Linux 
and Mac for example.

Chris

Von: Greg Dove<mailto:greg.d...@gmail.com>
Gesendet: Dienstag, 6. Juni 2017 22:45
An: dev@flex.apache.org<mailto:dev@flex.apache.org>
Betreff: [FlexJS] [Typedefs] [Maven] build issues

I have had a few issues with maven build of typedefs (in develop) over recent 
times.

I think others may have experienced this also, at least I have seen similar 
things mentioned previously.

For me this seems to be related to patches being applied when the content that 
is being downloaded is already patched (I *think*, based on eyeballing the 
patch and the download js content).

In the past I had this with svg.js and today I had it happen with google maps.
I hand edited  src/main/patch/js.patch to remove the already patched parts for 
svg.js

and today I simply removed the patch part of the pom.xml inside googlemaps 
subdir to get it to build.

I have not committed any of this because I have no knowledge or confidence in 
terms of what the correct approach should be here.

ant build seems to not have an issue with applying the patch content that 
(apparently) is no longer needed.

I am on windows if that possibly makes any difference to the way git apples 
patches.

Chris, or others, are you ever seeing the above issue in the dev maven build 
for typedefs?

[FlexJS] [Typedefs] [Maven] build issues

[Greg Dove]

I have had a few issues with maven build of typedefs (in develop) over recent 
times.

I think others may have experienced this also, at least I have seen similar 
things mentioned previously.

For me this seems to be related to patches being applied when the content that 
is being downloaded is already patched (I *think*, based on eyeballing the 
patch and the download js content). 

In the past I had this with svg.js and today I had it happen with google maps. 
I hand edited  src/main/patch/js.patch to remove the already patched parts for 
svg.js 

and today I simply removed the patch part of the pom.xml inside googlemaps 
subdir to get it to build.

I have not committed any of this because I have no knowledge or confidence in 
terms of what the correct approach should be here.

ant build seems to not have an issue with applying the patch content that 
(apparently) is no longer needed.

I am on windows if that possibly makes any difference to the way git apples 
patches.

Chris, or others, are you ever seeing the above issue in the dev maven build 
for typedefs?

Re: dual issues

[Harbs]

In compilerOptions.

Here’s my tasks.json file:
{
// See https://go.microsoft.com/fwlink/?LinkId=733558
// for the documentation about the tasks.json format
"version": "0.1.0",
"command": "asconfigc",
"isShellCommand": true,
"args": [
"--flexHome=/FlexSDK/FlexJSNightly"
],
"windows": {
"args": [
"--flexHome=C:\\dev\\flexjs_builds\\nightly_07"
]
},
"showOutput": "always"
}

asconfig.json:
{
"config": "flex",
"compilerOptions": {
"debug": true,
"targets": ["JSFlex"],
"source-map": false,
"library-path": [
"lib"
],
"external-library-path": [
"typedefs"
]
},
"copySourcePathAssets": true,
"additionalOptions": "-remove-circulars 
-js-output-optimization=skipAsCoercions 
-html-template=src/resources/mdl-js-index-template.html 
-js-external-library-path+=typedefs -js-library-path+=lib",
"files":
[
"src/PortedPrintUI.mxml"
]
}

I’m having other issues after switching to dual as well. Maybe I should show 
you…

Harbs

> On Jun 5, 2017, at 4:31 PM, Josh Tynjala <joshtynj...@gmail.com> wrote:
> 
> Are you setting debug in an argument passed to asconfigc, like this?
> 
> asconfigc --debug=false
> 
> Or is it specified in asconfig.json in the compilerOptions section?
> 
> Both will work, but the asconfigc command line option will always override
> whatever is in asconfig.json. In VSCode, the generated tasks.json now
> defaults to --debug=true, so that could take precedence if you're trying to
> set debug to false in asconfig.json. You might want to double-check both
> locations (asconfig.json and tasks.json).
> 
> - Josh
> 
> On Sun, Jun 4, 2017 at 2:08 AM, Harbs <harbs.li...@gmail.com> wrote:
> 
>> I’m seeing similar times using both ant and asconfig for debug. (Both are
>> good.)
>> 
>> asconfig is not compiling a release build at all — even if I specify
>> debug=false.
>> 
>> I have not tried with Maven.
>> 
>>> On Jun 4, 2017, at 12:06 PM, Justin Mclean <jus...@classsoftware.com>
>> wrote:
>>> 
>>> Hi,
>>> 
>>>> Justin you are compiling by Maven ?
>>> 
>>> Yep and Debug builds are noticeably slower for me.
>>> 
>>> Justin
>> 
>> 

Re: dual issues

[Josh Tynjala]

Are you setting debug in an argument passed to asconfigc, like this?

asconfigc --debug=false

Or is it specified in asconfig.json in the compilerOptions section?

Both will work, but the asconfigc command line option will always override
whatever is in asconfig.json. In VSCode, the generated tasks.json now
defaults to --debug=true, so that could take precedence if you're trying to
set debug to false in asconfig.json. You might want to double-check both
locations (asconfig.json and tasks.json).

- Josh

On Sun, Jun 4, 2017 at 2:08 AM, Harbs  wrote:

> I’m seeing similar times using both ant and asconfig for debug. (Both are
> good.)
>
> asconfig is not compiling a release build at all — even if I specify
> debug=false.
>
> I have not tried with Maven.
>
> > On Jun 4, 2017, at 12:06 PM, Justin Mclean 
> wrote:
> >
> > Hi,
> >
> >> Justin you are compiling by Maven ?
> >
> > Yep and Debug builds are noticeably slower for me.
> >
> > Justin
>
>

Re: dual issues

[Alex Harui]

For me, the Ant build calls the "compile" target in
examples/build_example.xml and sets the config_arg to use
frameworks/js-config.xml which has targets set to just JS.

A better example might be examples/flexjs/MDLExample.  For Ant, it has a
src/main/config/compile-app-config.xml which specifies:
JSFlex

And for me, no SWF is produced.


HTH,
-Alex

On 6/4/17, 10:31 AM, "Harbs"  wrote:

>That is using -js-output-type=FLEXJS
>
>My understanding is that we are supposed to use -targets=JSFlex now.
>
>Is that not right?
>
>> On Jun 4, 2017, at 5:43 PM, Alex Harui  wrote:
>> 
>> The examples/native/ButtonExample does not produce a SWF for me, so
>>maybe
>> compare that setup against yours.
>> 
>> HTH,
>> -Alex
>> 
>> On 6/4/17, 12:56 AM, "Harbs"  wrote:
>> 
>>> Good news: Prior to the switch to dual, I was not getting a minified
>>> release build working. That’s now working.
>>> Bad news: It takes much longer than it used to, and I’m getting a swf
>>> file even though I’m only using -targets=JSFlex
>>> 
>>> Harbs
>> 
>

Re: dual issues

[Harbs]

That is using -js-output-type=FLEXJS

My understanding is that we are supposed to use -targets=JSFlex now.

Is that not right?

> On Jun 4, 2017, at 5:43 PM, Alex Harui  wrote:
> 
> The examples/native/ButtonExample does not produce a SWF for me, so maybe
> compare that setup against yours.
> 
> HTH,
> -Alex
> 
> On 6/4/17, 12:56 AM, "Harbs"  wrote:
> 
>> Good news: Prior to the switch to dual, I was not getting a minified
>> release build working. That’s now working.
>> Bad news: It takes much longer than it used to, and I’m getting a swf
>> file even though I’m only using -targets=JSFlex
>> 
>> Harbs
> 

Re: dual issues

[Alex Harui]

The examples/native/ButtonExample does not produce a SWF for me, so maybe
compare that setup against yours.

HTH,
-Alex

On 6/4/17, 12:56 AM, "Harbs"  wrote:

>Good news: Prior to the switch to dual, I was not getting a minified
>release build working. That’s now working.
>Bad news: It takes much longer than it used to, and I’m getting a swf
>file even though I’m only using -targets=JSFlex
>
>Harbs

Re: dual issues

[Harbs]

I’m seeing similar times using both ant and asconfig for debug. (Both are good.)

asconfig is not compiling a release build at all — even if I specify 
debug=false.

I have not tried with Maven.

> On Jun 4, 2017, at 12:06 PM, Justin Mclean  wrote:
> 
> Hi,
> 
>> Justin you are compiling by Maven ?
> 
> Yep and Debug builds are noticeably slower for me.
> 
> Justin

Re: dual issues

[Justin Mclean]

Hi,

> Justin you are compiling by Maven ?

Yep and Debug builds are noticeably slower for me.

Justin

Re: dual issues

[Harbs]

To be clear:
Debug builds are as fast if not faster than they used to be.
Release build are much slower.

I’m getting a debug build (with remove circulars) of my (complex) app taking 
about 10 seconds.
The release build takes about 3 minutes.


> On Jun 4, 2017, at 11:57 AM, Justin Mclean  wrote:
> 
> Hi,
> 
>> Bad news: It takes much longer than it used to, and I’m getting a swf file 
>> even though I’m only using -targets=JSFlex
> 
> Compiles are also taking a lot longer for me and I’m also only using the 
> JSFlex target.
> 
> Thanks,
> Justin

Re: dual issues

[piotrz]

As I remember Harbs is compiling it by ANT ? Am I right ? Justin you are
compiling by Maven ?

Piotr



-
Apache Flex PMC
piotrzarzyck...@gmail.com
--
View this message in context: 
http://apache-flex-development.247.n4.nabble.com/dual-issues-tp61955p62070.html
Sent from the Apache Flex Development mailing list archive at Nabble.com.

Re: dual issues

[Justin Mclean]

Hi,

> Bad news: It takes much longer than it used to, and I’m getting a swf file 
> even though I’m only using -targets=JSFlex

Compiles are also taking a lot longer for me and I’m also only using the JSFlex 
target.

Thanks,
Justin

Re: dual issues

[piotrz]

Ahh..That's not good. Another thing for resolve in release definitly.

Piotr



-
Apache Flex PMC
piotrzarzyck...@gmail.com
--
View this message in context: 
http://apache-flex-development.247.n4.nabble.com/dual-issues-tp61955p62066.html
Sent from the Apache Flex Development mailing list archive at Nabble.com.

Re: dual issues

[Harbs]

Good news: Prior to the switch to dual, I was not getting a minified release 
build working. That’s now working.
Bad news: It takes much longer than it used to, and I’m getting a swf file even 
though I’m only using -targets=JSFlex

Harbs

Re: dual issues

[Harbs]

Objects like Promise, Map, Set etc. are available by default in pretty much all 
modern browsers.

Things like await, arrow functions, let, etc. are being added by browsers all 
the time. I’m not sure exactly which ones have what right now, but they are 
pretty far along already, so that article is already out of date. ;-)

Either way, we’re not outputting ES6-style code (currently), so those features 
are a non-issue.

> On Jun 4, 2017, at 10:34 AM, Alex Harui  wrote:
> 
> I thought es6 wasn't always on by default?
> 
> https://thenextweb.com/dd/2016/03/09/6-reasons-need-learn-javascript-es6-no
> w-not-later/#.tnw_5KKUOOPX
> 
> I haven't been paying attention so maybe all browsers are fully compliant
> now?
> 
> -Alex
> 
> On 6/4/17, 12:01 AM, "Harbs"  wrote:
> 
>> Not sure what you mean here.
>> 
>>> On Jun 4, 2017, at 8:52 AM, Alex Harui  wrote:
>>> 
>>> Does FlexJS emit the right code to enable es6 APIs?
>> 
> 

Re: dual issues

[Alex Harui]

I thought es6 wasn't always on by default?

https://thenextweb.com/dd/2016/03/09/6-reasons-need-learn-javascript-es6-no
w-not-later/#.tnw_5KKUOOPX

I haven't been paying attention so maybe all browsers are fully compliant
now?

-Alex

On 6/4/17, 12:01 AM, "Harbs"  wrote:

>Not sure what you mean here.
>
>> On Jun 4, 2017, at 8:52 AM, Alex Harui  wrote:
>> 
>> Does FlexJS emit the right code to enable es6 APIs?
>

Re: dual issues

[Harbs]

Not sure what you mean here.

> On Jun 4, 2017, at 8:52 AM, Alex Harui  wrote:
> 
> Does FlexJS emit the right code to enable es6 APIs?

Re: dual issues

[Alex Harui]

On 6/3/17, 2:02 PM, "Harbs"  wrote:

>I just ran into another issue.
>
>I’m not sure why dual would have changed this, but I’m now getting an
>(erroneous) error when trying to use Promises:
>
>Incorrect number of arguments.  Expected no more than 0
>
>   return new Promise(function(resolve:*,reject:*):void{
>  ^

Looking at the typedefs, Promise is in missing.js and defined to have no
args.  It is also in es6.js.  Maybe there is some randomness about which
definition wins.  Seems like you could remove the one from missing.js.
Does FlexJS emit the right code to enable es6 APIs?

-Alex

Re: dual issues

[Harbs]

I just ran into another issue.

I’m not sure why dual would have changed this, but I’m now getting an 
(erroneous) error when trying to use Promises:

Incorrect number of arguments.  Expected no more than 0

return new Promise(function(resolve:*,reject:*):void{
   ^

> On Jun 1, 2017, at 10:39 AM, Harbs <harbs.li...@gmail.com> wrote:
> 
> My bad. I accidentally overwrote my copy of BasicJS.swc.
> 
> After fixing some issues, I got my app to run using the latest dual changes.
> 
> All in all, switching to dual went a lot better than I expected.
> 
> Great work, Alex!
> 
> I think the switch to the dual approach was the right one. There’s still 
> probably some rough edges to improve, but it’s definitely usable.
> 
> Thanks,
> Harbs
> 
>> On Jun 1, 2017, at 5:57 AM, Harbs <harbs.li...@gmail.com> wrote:
>> 
>> Another weird issue:
>> 
>>> /Users/harbs/Documents/git/PrintUI/printui-flexjs/PortedPrintUI/src/com/printui/dummy/view/Image.as(12):
>>>  col: 15 interface method imageElement in interface IImage not implemented 
>>> by class Image
>>> 
>>> public class Image extends BinaryImage
>>>  ^
>>> 
>>> /Users/harbs/Documents/git/PrintUI/printui-flexjs/PortedPrintUI/src/com/printui/dummy/view/Image.as(12):
>>>  col: 15 interface method applyImageData in interface IImage not 
>>> implemented by class Image
>>> 
>>> public class Image extends BinaryImage
>>> 
>> 
>> org.apache.flex.html.BinaryImage which in tern subclasses 
>> org.apache.flex.html.Image. org.apache.flex.html.Image implements these two 
>> methods. I figured the compiler might be getting confused by the fact that 
>> Image is named the same as an ancestor class, so I changed the name and I'm 
>> getting the same error in additon to a null pointer error:
>> 
>>> 
>>> /Users/harbs/Documents/git/PrintUI/printui-flexjs/PortedPrintUI/src/com/printui/dummy/view/PUIImage.as(12):
>>>  col: 15 interface method imageElement in interface IImage not implemented 
>>> by class PUIImage
>>> 
>>> public class PUIImage extends BinaryImage
>>>  ^
>>> 
>>> /Users/harbs/Documents/git/PrintUI/printui-flexjs/PortedPrintUI/src/com/printui/dummy/view/PUIImage.as(12):
>>>  col: 15 interface method applyImageData in interface IImage not 
>>> implemented by class PUIImage
>>> 
>>> public class PUIImage extends BinaryImage
>>>  ^
>>> 
>>> /Users/harbs/Documents/git/PrintUI/printui-flexjs/PortedPrintUI/src/com/printui/view/components/DesignAreaComponents/drawnObjects/ImageObject.as(59):
>>>  col: 3 java.lang.NullPointerException
>>> at 
>>> org.apache.flex.compiler.internal.projects.FlexProject.isValidTypeConversion(FlexProject.java:2304)
>>> at 
>>> org.apache.flex.compiler.internal.semantics.MethodBodySemanticChecker.checkImplicitConversion(MethodBodySemanticChecker.java:564)
>>> at 
>>> org.apache.flex.compiler.internal.semantics.MethodBodySemanticChecker.checkImplicitConversion(MethodBodySemanticChecker.java:374)
>>> at 
>>> org.apache.flex.compiler.internal.semantics.MethodBodySemanticChecker.checkAssignment(MethodBodySemanticChecker.java:246)
>>> at 
>>> org.apache.flex.compiler.internal.as.codegen.ABCGeneratingReducer.reduce_assignToNameExpr_to_void_expression(ABCGeneratingReducer.java:2947)
>>> at 
>>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.action_237(CmcEmitter.java:4451)
>>> at 
>>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.dispatchAction(CmcEmitter.java:8067)
>>> at 
>>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduceAntecedent(CmcEmitter.java:39127)
>>> at 
>>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.dispatchAction(CmcEmitter.java:9362)
>>> at 
>>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduceAntecedent(CmcEmitter.java:39127)
>>> at 
>>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduce(CmcEmitter.java:39104)
>>> at 
>>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduceSubgoals(CmcEmitter.java:39155)
>>> at 
>>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduceAntecedent(CmcEmitter.java:39126)
>>> at 
>>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduce(CmcEmitter.java:39104)
>>> at

Re: dual issues

[Josh Tynjala]

When they're defined in the *-config.xml files, they'll be picked up
automatically without any special code in VSCode. However, now that you
mention it, I need to allow app developers to customize these in
asconfig.json too.

- Josh

On Jun 1, 2017 12:04 AM, "Alex Harui"  wrote:

Josh, there are new compiler options like js-external-library-path,
js-library-path, swf-library-path, swf-external-library-path.  Does your
VSCode integration handle these?

-Alex

On 5/31/17, 8:42 PM, "Josh Tynjala"  wrote:

>As I understand it, the compiler should just take care of differences
>between JS and SWF, as long as your targets option is correct. VSCode
>shouldn't need to do anything​ special except to pass in the correct
>options.
>
>There's nothing related to dual changes on my to-do list at this time, but
>maybe I missed something important...
>
>- Josh
>
>On May 31, 2017 7:52 PM, "Harbs"  wrote:
>
>Josh,
>
>VS Code is now reporting lots of errors in the Problems window. I’m
>guessing it does not know how to differentiate between JS and SWF builds
>using dual?
>
>Is that something on your to-do list?
>
>Thanks,
>Harbs
>
>> On Jun 1, 2017, at 5:10 AM, Harbs  wrote:
>>
>> Bingo. Thanks!
>>
>> SWF output still gets all kinds of errors, but the JS output is pretty
>error free. I still have some fixing up to do before I can see how well it
>actually works, but at least the compiler is not complaining now.
>>
>>> On Jun 1, 2017, at 4:45 AM, Josh Tynjala  wrote:
>>>
>>> Try replacing the js-output-type compiler option with:
>>>
>>> "targets": ["JSFlex"]
>>>
>>> - Josh
>>>
>>> On May 31, 2017 6:26 PM, "Harbs"  wrote:
>>>
 I’m using asconfig in VSCode to compile with these settings:
   "config": "flex",
   "compilerOptions": {
   "debug": false,
   "js-output-type": "flexjs",
   "source-map": false,
   "library-path": [
   "lib"
   ],

> On Jun 1, 2017, at 1:04 AM, Justin Mclean 
 wrote:
>
> HI,
>
>> Am I missing something?
>
> Perhaps try JSFlex in your pom.xml? I believe the
 name changed from pre to post dual branch.
>
> Justin
>


>>

Re: dual issues

[piotrz]

Hi Harbs,

Great news! This approach us to release!

Piotr



-
Apache Flex PMC
piotrzarzyck...@gmail.com
--
View this message in context: 
http://apache-flex-development.247.n4.nabble.com/dual-issues-tp61955p61988.html
Sent from the Apache Flex Development mailing list archive at Nabble.com.

Re: dual issues

[Harbs]

My bad. I accidentally overwrote my copy of BasicJS.swc.

After fixing some issues, I got my app to run using the latest dual changes.

All in all, switching to dual went a lot better than I expected.

Great work, Alex!

I think the switch to the dual approach was the right one. There’s still 
probably some rough edges to improve, but it’s definitely usable.

Thanks,
Harbs

> On Jun 1, 2017, at 5:57 AM, Harbs <harbs.li...@gmail.com> wrote:
> 
> Another weird issue:
> 
>> /Users/harbs/Documents/git/PrintUI/printui-flexjs/PortedPrintUI/src/com/printui/dummy/view/Image.as(12):
>>  col: 15 interface method imageElement in interface IImage not implemented 
>> by class Image
>> 
>>  public class Image extends BinaryImage
>>   ^
>> 
>> /Users/harbs/Documents/git/PrintUI/printui-flexjs/PortedPrintUI/src/com/printui/dummy/view/Image.as(12):
>>  col: 15 interface method applyImageData in interface IImage not implemented 
>> by class Image
>> 
>>  public class Image extends BinaryImage
>> 
> 
> org.apache.flex.html.BinaryImage which in tern subclasses 
> org.apache.flex.html.Image. org.apache.flex.html.Image implements these two 
> methods. I figured the compiler might be getting confused by the fact that 
> Image is named the same as an ancestor class, so I changed the name and I'm 
> getting the same error in additon to a null pointer error:
> 
>>  
>> /Users/harbs/Documents/git/PrintUI/printui-flexjs/PortedPrintUI/src/com/printui/dummy/view/PUIImage.as(12):
>>  col: 15 interface method imageElement in interface IImage not implemented 
>> by class PUIImage
>> 
>>  public class PUIImage extends BinaryImage
>>   ^
>> 
>> /Users/harbs/Documents/git/PrintUI/printui-flexjs/PortedPrintUI/src/com/printui/dummy/view/PUIImage.as(12):
>>  col: 15 interface method applyImageData in interface IImage not implemented 
>> by class PUIImage
>> 
>>  public class PUIImage extends BinaryImage
>>   ^
>> 
>> /Users/harbs/Documents/git/PrintUI/printui-flexjs/PortedPrintUI/src/com/printui/view/components/DesignAreaComponents/drawnObjects/ImageObject.as(59):
>>  col: 3 java.lang.NullPointerException
>>  at 
>> org.apache.flex.compiler.internal.projects.FlexProject.isValidTypeConversion(FlexProject.java:2304)
>>  at 
>> org.apache.flex.compiler.internal.semantics.MethodBodySemanticChecker.checkImplicitConversion(MethodBodySemanticChecker.java:564)
>>  at 
>> org.apache.flex.compiler.internal.semantics.MethodBodySemanticChecker.checkImplicitConversion(MethodBodySemanticChecker.java:374)
>>  at 
>> org.apache.flex.compiler.internal.semantics.MethodBodySemanticChecker.checkAssignment(MethodBodySemanticChecker.java:246)
>>  at 
>> org.apache.flex.compiler.internal.as.codegen.ABCGeneratingReducer.reduce_assignToNameExpr_to_void_expression(ABCGeneratingReducer.java:2947)
>>  at 
>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.action_237(CmcEmitter.java:4451)
>>  at 
>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.dispatchAction(CmcEmitter.java:8067)
>>  at 
>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduceAntecedent(CmcEmitter.java:39127)
>>  at 
>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.dispatchAction(CmcEmitter.java:9362)
>>  at 
>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduceAntecedent(CmcEmitter.java:39127)
>>  at 
>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduce(CmcEmitter.java:39104)
>>  at 
>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduceSubgoals(CmcEmitter.java:39155)
>>  at 
>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduceAntecedent(CmcEmitter.java:39126)
>>  at 
>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduce(CmcEmitter.java:39104)
>>  at 
>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduceSubgoals(CmcEmitter.java:39146)
>>  at 
>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduceAntecedent(CmcEmitter.java:39126)
>>  at 
>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduce(CmcEmitter.java:39104)
>>  at 
>> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.burm(CmcEmitter.java:39406)
>>  at 
>> org.apache.flex.compiler.internal.as.codegen.ABCGenerator.generateInstructions(ABCGenerator.java:228)
>>  at 
>> org.apache.flex.compiler.internal.as.codegen.ABCGenerator.generateMethodBodyForFunction(ABCGenerator.java:393)
>>  at 

Re: dual issues

[Alex Harui]

Josh, there are new compiler options like js-external-library-path,
js-library-path, swf-library-path, swf-external-library-path.  Does your
VSCode integration handle these?

-Alex

On 5/31/17, 8:42 PM, "Josh Tynjala"  wrote:

>As I understand it, the compiler should just take care of differences
>between JS and SWF, as long as your targets option is correct. VSCode
>shouldn't need to do anything​ special except to pass in the correct
>options.
>
>There's nothing related to dual changes on my to-do list at this time, but
>maybe I missed something important...
>
>- Josh
>
>On May 31, 2017 7:52 PM, "Harbs"  wrote:
>
>Josh,
>
>VS Code is now reporting lots of errors in the Problems window. I’m
>guessing it does not know how to differentiate between JS and SWF builds
>using dual?
>
>Is that something on your to-do list?
>
>Thanks,
>Harbs
>
>> On Jun 1, 2017, at 5:10 AM, Harbs  wrote:
>>
>> Bingo. Thanks!
>>
>> SWF output still gets all kinds of errors, but the JS output is pretty
>error free. I still have some fixing up to do before I can see how well it
>actually works, but at least the compiler is not complaining now.
>>
>>> On Jun 1, 2017, at 4:45 AM, Josh Tynjala  wrote:
>>>
>>> Try replacing the js-output-type compiler option with:
>>>
>>> "targets": ["JSFlex"]
>>>
>>> - Josh
>>>
>>> On May 31, 2017 6:26 PM, "Harbs"  wrote:
>>>
 I’m using asconfig in VSCode to compile with these settings:
   "config": "flex",
   "compilerOptions": {
   "debug": false,
   "js-output-type": "flexjs",
   "source-map": false,
   "library-path": [
   "lib"
   ],

> On Jun 1, 2017, at 1:04 AM, Justin Mclean 
 wrote:
>
> HI,
>
>> Am I missing something?
>
> Perhaps try JSFlex in your pom.xml? I believe the
 name changed from pre to post dual branch.
>
> Justin
>


>>

Re: dual issues

[Alex Harui]

If targets only has JSFlex, there shouldn't be any SWF compile at all.
The console should spit out what it thinks it needs to do.

Keep in mind that there are now SWCs for SWFs and SWCs for JS.  Prior to
dual, there was a main SWC with SWF APIs that the compiler saw, and JS
that it substituted in the output.  The JS SWC was only used to build
downstream JS SWCs.

If you have a custom SWC before dual and it has platform dependent APIs,
and you are only interested in JS output, you need to make sure your SWC's
upstream SWCs are the JS SWCs and not the SWF SWCs.

That in turn requires being more careful about which SWCs are on the
-library-path vs -swf-library-path vs -js-library-path.  In theory, if you
only want JS output, it should be as simple as making sure everything on
the -library-path is a SWC for JS.

HTH,
-Alex

On 5/31/17, 8:42 PM, "Josh Tynjala"  wrote:

>As I understand it, the compiler should just take care of differences
>between JS and SWF, as long as your targets option is correct. VSCode
>shouldn't need to do anything​ special except to pass in the correct
>options.
>
>There's nothing related to dual changes on my to-do list at this time, but
>maybe I missed something important...
>
>- Josh
>
>On May 31, 2017 7:52 PM, "Harbs"  wrote:
>
>Josh,
>
>VS Code is now reporting lots of errors in the Problems window. I’m
>guessing it does not know how to differentiate between JS and SWF builds
>using dual?
>
>Is that something on your to-do list?
>
>Thanks,
>Harbs
>
>> On Jun 1, 2017, at 5:10 AM, Harbs  wrote:
>>
>> Bingo. Thanks!
>>
>> SWF output still gets all kinds of errors, but the JS output is pretty
>error free. I still have some fixing up to do before I can see how well it
>actually works, but at least the compiler is not complaining now.
>>
>>> On Jun 1, 2017, at 4:45 AM, Josh Tynjala  wrote:
>>>
>>> Try replacing the js-output-type compiler option with:
>>>
>>> "targets": ["JSFlex"]
>>>
>>> - Josh
>>>
>>> On May 31, 2017 6:26 PM, "Harbs"  wrote:
>>>
 I’m using asconfig in VSCode to compile with these settings:
   "config": "flex",
   "compilerOptions": {
   "debug": false,
   "js-output-type": "flexjs",
   "source-map": false,
   "library-path": [
   "lib"
   ],

> On Jun 1, 2017, at 1:04 AM, Justin Mclean 
 wrote:
>
> HI,
>
>> Am I missing something?
>
> Perhaps try JSFlex in your pom.xml? I believe the
 name changed from pre to post dual branch.
>
> Justin
>


>>

Re: dual issues

[Josh Tynjala]

As I understand it, the compiler should just take care of differences
between JS and SWF, as long as your targets option is correct. VSCode
shouldn't need to do anything​ special except to pass in the correct
options.

There's nothing related to dual changes on my to-do list at this time, but
maybe I missed something important...

- Josh

On May 31, 2017 7:52 PM, "Harbs"  wrote:

Josh,

VS Code is now reporting lots of errors in the Problems window. I’m
guessing it does not know how to differentiate between JS and SWF builds
using dual?

Is that something on your to-do list?

Thanks,
Harbs

> On Jun 1, 2017, at 5:10 AM, Harbs  wrote:
>
> Bingo. Thanks!
>
> SWF output still gets all kinds of errors, but the JS output is pretty
error free. I still have some fixing up to do before I can see how well it
actually works, but at least the compiler is not complaining now.
>
>> On Jun 1, 2017, at 4:45 AM, Josh Tynjala  wrote:
>>
>> Try replacing the js-output-type compiler option with:
>>
>> "targets": ["JSFlex"]
>>
>> - Josh
>>
>> On May 31, 2017 6:26 PM, "Harbs"  wrote:
>>
>>> I’m using asconfig in VSCode to compile with these settings:
>>>   "config": "flex",
>>>   "compilerOptions": {
>>>   "debug": false,
>>>   "js-output-type": "flexjs",
>>>   "source-map": false,
>>>   "library-path": [
>>>   "lib"
>>>   ],
>>>
 On Jun 1, 2017, at 1:04 AM, Justin Mclean 
>>> wrote:

 HI,

> Am I missing something?

 Perhaps try JSFlex in your pom.xml? I believe the
>>> name changed from pre to post dual branch.

 Justin

>>>
>>>
>

Re: dual issues

[Harbs]

Another weird issue:

> /Users/harbs/Documents/git/PrintUI/printui-flexjs/PortedPrintUI/src/com/printui/dummy/view/Image.as(12):
>  col: 15 interface method imageElement in interface IImage not implemented by 
> class Image
> 
>   public class Image extends BinaryImage
>^
> 
> /Users/harbs/Documents/git/PrintUI/printui-flexjs/PortedPrintUI/src/com/printui/dummy/view/Image.as(12):
>  col: 15 interface method applyImageData in interface IImage not implemented 
> by class Image
> 
>   public class Image extends BinaryImage
> 

org.apache.flex.html.BinaryImage which in tern subclasses 
org.apache.flex.html.Image. org.apache.flex.html.Image implements these two 
methods. I figured the compiler might be getting confused by the fact that 
Image is named the same as an ancestor class, so I changed the name and I'm 
getting the same error in additon to a null pointer error:

>   
> /Users/harbs/Documents/git/PrintUI/printui-flexjs/PortedPrintUI/src/com/printui/dummy/view/PUIImage.as(12):
>  col: 15 interface method imageElement in interface IImage not implemented by 
> class PUIImage
> 
>   public class PUIImage extends BinaryImage
>^
> 
> /Users/harbs/Documents/git/PrintUI/printui-flexjs/PortedPrintUI/src/com/printui/dummy/view/PUIImage.as(12):
>  col: 15 interface method applyImageData in interface IImage not implemented 
> by class PUIImage
> 
>   public class PUIImage extends BinaryImage
>^
> 
> /Users/harbs/Documents/git/PrintUI/printui-flexjs/PortedPrintUI/src/com/printui/view/components/DesignAreaComponents/drawnObjects/ImageObject.as(59):
>  col: 3 java.lang.NullPointerException
>   at 
> org.apache.flex.compiler.internal.projects.FlexProject.isValidTypeConversion(FlexProject.java:2304)
>   at 
> org.apache.flex.compiler.internal.semantics.MethodBodySemanticChecker.checkImplicitConversion(MethodBodySemanticChecker.java:564)
>   at 
> org.apache.flex.compiler.internal.semantics.MethodBodySemanticChecker.checkImplicitConversion(MethodBodySemanticChecker.java:374)
>   at 
> org.apache.flex.compiler.internal.semantics.MethodBodySemanticChecker.checkAssignment(MethodBodySemanticChecker.java:246)
>   at 
> org.apache.flex.compiler.internal.as.codegen.ABCGeneratingReducer.reduce_assignToNameExpr_to_void_expression(ABCGeneratingReducer.java:2947)
>   at 
> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.action_237(CmcEmitter.java:4451)
>   at 
> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.dispatchAction(CmcEmitter.java:8067)
>   at 
> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduceAntecedent(CmcEmitter.java:39127)
>   at 
> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.dispatchAction(CmcEmitter.java:9362)
>   at 
> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduceAntecedent(CmcEmitter.java:39127)
>   at 
> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduce(CmcEmitter.java:39104)
>   at 
> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduceSubgoals(CmcEmitter.java:39155)
>   at 
> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduceAntecedent(CmcEmitter.java:39126)
>   at 
> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduce(CmcEmitter.java:39104)
>   at 
> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduceSubgoals(CmcEmitter.java:39146)
>   at 
> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduceAntecedent(CmcEmitter.java:39126)
>   at 
> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.reduce(CmcEmitter.java:39104)
>   at 
> org.apache.flex.compiler.internal.as.codegen.CmcEmitter.burm(CmcEmitter.java:39406)
>   at 
> org.apache.flex.compiler.internal.as.codegen.ABCGenerator.generateInstructions(ABCGenerator.java:228)
>   at 
> org.apache.flex.compiler.internal.as.codegen.ABCGenerator.generateMethodBodyForFunction(ABCGenerator.java:393)
>   at 
> org.apache.flex.compiler.internal.as.codegen.ABCGenerator.generateMethodBodyForFunction(ABCGenerator.java:347)
>   at 
> org.apache.flex.compiler.internal.as.codegen.ABCGenerator.generateFunction(ABCGenerator.java:262)
>   at 
> org.apache.flex.compiler.internal.as.codegen.ClassDirectiveProcessor.finishClassDefinition(ClassDirectiveProcessor.java:639)
>   at 
> org.apache.flex.compiler.internal.as.codegen.GlobalDirectiveProcessor.declareClass(GlobalDirectiveProcessor.java:430)
>   at 
> org.apache.flex.compiler.internal.as.codegen.DirectiveProcessor.processNode(DirectiveProcessor.java:206)
>   at 
> org.apache.flex.compiler.internal.as.codegen.DirectiveProcessor.traverse(DirectiveProcessor.java:188)
>   at 
> org.apache.flex.compiler.internal.as.codegen.GlobalDirectiveProcessor.declarePackage(GlobalDirectiveProcessor.java:455)
>   at 
> org.apache.flex.compiler.internal.as.codegen.DirectiveProcessor.processNode(DirectiveProcessor.java:223)
>   at 

Re: dual issues

[Harbs]

Josh,

VS Code is now reporting lots of errors in the Problems window. I’m guessing it 
does not know how to differentiate between JS and SWF builds using dual?

Is that something on your to-do list?

Thanks,
Harbs

> On Jun 1, 2017, at 5:10 AM, Harbs  wrote:
> 
> Bingo. Thanks!
> 
> SWF output still gets all kinds of errors, but the JS output is pretty error 
> free. I still have some fixing up to do before I can see how well it actually 
> works, but at least the compiler is not complaining now.
> 
>> On Jun 1, 2017, at 4:45 AM, Josh Tynjala  wrote:
>> 
>> Try replacing the js-output-type compiler option with:
>> 
>> "targets": ["JSFlex"]
>> 
>> - Josh
>> 
>> On May 31, 2017 6:26 PM, "Harbs"  wrote:
>> 
>>> I’m using asconfig in VSCode to compile with these settings:
>>>   "config": "flex",
>>>   "compilerOptions": {
>>>   "debug": false,
>>>   "js-output-type": "flexjs",
>>>   "source-map": false,
>>>   "library-path": [
>>>   "lib"
>>>   ],
>>> 
 On Jun 1, 2017, at 1:04 AM, Justin Mclean 
>>> wrote:
 
 HI,
 
> Am I missing something?
 
 Perhaps try JSFlex in your pom.xml? I believe the
>>> name changed from pre to post dual branch.
 
 Justin
 
>>> 
>>> 
> 

Re: dual issues

[Harbs]

Bingo. Thanks!

SWF output still gets all kinds of errors, but the JS output is pretty error 
free. I still have some fixing up to do before I can see how well it actually 
works, but at least the compiler is not complaining now.

> On Jun 1, 2017, at 4:45 AM, Josh Tynjala  wrote:
> 
> Try replacing the js-output-type compiler option with:
> 
> "targets": ["JSFlex"]
> 
> - Josh
> 
> On May 31, 2017 6:26 PM, "Harbs"  wrote:
> 
>> I’m using asconfig in VSCode to compile with these settings:
>>"config": "flex",
>>"compilerOptions": {
>>"debug": false,
>>"js-output-type": "flexjs",
>>"source-map": false,
>>"library-path": [
>>"lib"
>>],
>> 
>>> On Jun 1, 2017, at 1:04 AM, Justin Mclean 
>> wrote:
>>> 
>>> HI,
>>> 
 Am I missing something?
>>> 
>>> Perhaps try JSFlex in your pom.xml? I believe the
>> name changed from pre to post dual branch.
>>> 
>>> Justin
>>> 
>> 
>> 

Re: dual issues

[Harbs]

The examples I see seem to use COMPILE blocks.

I’m not sure how this translates to the app level. Is there a way to define a 
single method which overrides a SWF one and implements a new JS one (without 
using COMPILE blocks)? That seems to me like a pretty common use case for 
migrated apps.

> On Jun 1, 2017, at 2:18 AM, Alex Harui  wrote:
> 
> 
> 
> On 5/31/17, 2:38 PM, "Harbs"  > wrote:
> 
>> I just switched to dual and I’m getting lots of errors. These errors seem
>> to be exactly what I was afraid of, and I’m not sure how to resolve
>> without being restricted by naming of properties and methods.
>> 
>> 1. I’m using a mask setter and getter for images it’s now giving override
>> errors.
>> 2. contains() is an incompatible override.
>> 3. scaleX and scaleY
>> 4. blendMode
>> 5. rotation
>> 6. filters
>> 7. mouseEnabled
>> etc….
>> 
>> I thought there was supposed to be something built into the compiler to
>> hide the flash-specific APIs and allow reuse of them.
> 
> Not just in the compiler, but in metadata.  Search the source for
> [SWFOverride].  I guess I don't understand why you want/need to change the
> type of scaleX/Y, blendMode, etc.  Wouldn't mouseEnabled always be a
> boolean?
> 
> SWFOverride is designed for properties like mask and methods like
> contains() that are defined as a flash.display.DisplayObject by Flash and
> we probably want to use UIBase instead.  SWFOverride will work because
> UIBase is a DisplayObject on Flash.  I did not anticipate a requirement
> that mouseEnabled would be something other than a boolean.  What types are
> you using for these other properties?  Or maybe you just need COMPILE::JS
> on mouseEnabled?
> 
> HTH,
> -Alex

Re: dual issues

[Justin Mclean]

Hi,

> I’m using asconfig in VSCode to compile with these settings:
>   "config": "flex",
>   "compilerOptions": {
>   "debug": false,
>   "js-output-type": "flexjs",
>   "source-map": false,
>   "library-path": [
>   "lib"
>   ],


I’ve not used VSCode but perhaps try “JSFlex" rather than “flexjs" in there or 
perhaps “targets": “JSFlex"?

Justin

Re: dual issues

[Josh Tynjala]

Try replacing the js-output-type compiler option with:

"targets": ["JSFlex"]

- Josh

On May 31, 2017 6:26 PM, "Harbs"  wrote:

> I’m using asconfig in VSCode to compile with these settings:
> "config": "flex",
> "compilerOptions": {
> "debug": false,
> "js-output-type": "flexjs",
> "source-map": false,
> "library-path": [
> "lib"
> ],
>
> > On Jun 1, 2017, at 1:04 AM, Justin Mclean 
> wrote:
> >
> > HI,
> >
> >> Am I missing something?
> >
> > Perhaps try JSFlex in your pom.xml? I believe the
> name changed from pre to post dual branch.
> >
> > Justin
> >
>
>

Re: dual issues

[Harbs]

In case I wasn’t clear, I’m not outputting SWF — only JS output.

> On Jun 1, 2017, at 4:26 AM, Harbs  wrote:
> 
> I’m using asconfig in VSCode to compile with these settings:
> "config": "flex",
> "compilerOptions": {
> "debug": false,
> "js-output-type": "flexjs",
> "source-map": false,
> "library-path": [
> "lib"
> ],
> 
>> On Jun 1, 2017, at 1:04 AM, Justin Mclean > > wrote:
>> 
>> HI,
>> 
>>> Am I missing something?
>> 
>> Perhaps try JSFlex in your pom.xml? I believe the name 
>> changed from pre to post dual branch.
>> 
>> Justin
>> 
> 

Re: dual issues

[Harbs]

I’m using asconfig in VSCode to compile with these settings:
"config": "flex",
"compilerOptions": {
"debug": false,
"js-output-type": "flexjs",
"source-map": false,
"library-path": [
"lib"
],

> On Jun 1, 2017, at 1:04 AM, Justin Mclean  wrote:
> 
> HI,
> 
>> Am I missing something?
> 
> Perhaps try JSFlex in your pom.xml? I believe the name 
> changed from pre to post dual branch.
> 
> Justin
> 

Re: dual issues

[Alex Harui]

On 5/31/17, 2:38 PM, "Harbs"  wrote:

>I just switched to dual and I’m getting lots of errors. These errors seem
>to be exactly what I was afraid of, and I’m not sure how to resolve
>without being restricted by naming of properties and methods.
>
>1. I’m using a mask setter and getter for images it’s now giving override
>errors.
>2. contains() is an incompatible override.
>3. scaleX and scaleY
>4. blendMode
>5. rotation
>6. filters
>7. mouseEnabled
>etc….
>
>I thought there was supposed to be something built into the compiler to
>hide the flash-specific APIs and allow reuse of them.

Not just in the compiler, but in metadata.  Search the source for
[SWFOverride].  I guess I don't understand why you want/need to change the
type of scaleX/Y, blendMode, etc.  Wouldn't mouseEnabled always be a
boolean?

SWFOverride is designed for properties like mask and methods like
contains() that are defined as a flash.display.DisplayObject by Flash and
we probably want to use UIBase instead.  SWFOverride will work because
UIBase is a DisplayObject on Flash.  I did not anticipate a requirement
that mouseEnabled would be something other than a boolean.  What types are
you using for these other properties?  Or maybe you just need COMPILE::JS
on mouseEnabled?

HTH,
-Alex

dual issues

[Harbs]

I just switched to dual and I’m getting lots of errors. These errors seem to be 
exactly what I was afraid of, and I’m not sure how to resolve without being 
restricted by naming of properties and methods.

1. I’m using a mask setter and getter for images it’s now giving override 
errors.
2. contains() is an incompatible override.
3. scaleX and scaleY
4. blendMode
5. rotation
6. filters
7. mouseEnabled
etc….

I thought there was supposed to be something built into the compiler to hide 
the flash-specific APIs and allow reuse of them.

Am I missing something?

Thanks,
Harbs

Re: 4.16.0 Windows Install issues (was Re: [ANNOUNCE] Apache Flex 4.16.0 released)

[Vincent Sotto]

just add option to select mirror, rise and eu mirror did not work for me,
but us did

http://www-eu.apache.org/dist/flex/4.16.0/binaries/apache-flex-sdk-4.16.0-bin.zip
http://mirror.rise.ph/apache/flex/4.16.0/binaries/apache-flex-sdk-4.16.0-bin.zip

while this worked

http://www-us.apache.org/dist/flex/4.16.0/binaries/apache-flex-sdk-4.16.0-bin.zip

On Sun, Mar 26, 2017 at 12:50 AM, Alex Harui  wrote:

>
>
> On 3/25/17, 9:19 AM, "piotrz"  wrote:
>
> >I was referring to the idea where we stay with current installer as is and
> >popup with the message to the user if problems during download occurs.
> >
> >Message could contains link to the zip.
>
> I don't think the current installer knows how to pop up a message.  I
> think we could halt the install until the user hits a key.  We would have
> to change the installer.xml and repackage that in the 4.16.0 binary to
> make that happen though.  Or are you thinking of some other way to do
> generate the popup?
>
>
> -Alex
>
>

Re: 4.16.0 Windows Install issues (was Re: [ANNOUNCE] Apache Flex 4.16.0 released)

[Alex Harui]

On 3/25/17, 9:19 AM, "piotrz"  wrote:

>I was referring to the idea where we stay with current installer as is and
>popup with the message to the user if problems during download occurs.
>
>Message could contains link to the zip.

I don't think the current installer knows how to pop up a message.  I
think we could halt the install until the user hits a key.  We would have
to change the installer.xml and repackage that in the 4.16.0 binary to
make that happen though.  Or are you thinking of some other way to do
generate the popup?


-Alex

Re: 4.16.0 Windows Install issues (was Re: [ANNOUNCE] Apache Flex 4.16.0 released)

[piotrz]

I was referring to the idea where we stay with current installer as is and
popup with the message to the user if problems during download occurs.

Message could contains link to the zip.

Piotr



-
Apache Flex PMC
piotrzarzyck...@gmail.com
--
View this message in context: 
http://apache-flex-development.247.n4.nabble.com/4-16-0-Windows-Install-issues-was-Re-ANNOUNCE-Apache-Flex-4-16-0-released-tp60626p60768.html
Sent from the Apache Flex Development mailing list archive at Nabble.com.

Re: 4.16.0 Windows Install issues (was Re: [ANNOUNCE] Apache Flex 4.16.0 released)

[Alex Harui]

On 3/25/17, 9:05 AM, "piotrz"  wrote:

>Alex,
>
>Can we exact specify if user face this problem ? Maybe we could do this
>zip
>file, and notify user by popup - Hey you have to download 64 bit
>installer.
>
>In the other words zip file is best option for me.

I'm not sure what kind of popup you are thinking of.  What code/script
would we change to generate the popup?

Also, IMO, the PMC has to reach consensus that we can distribute a binary
that doesn't have a source package that automatically generates it.  I'm
personally ok with that.  I don't think automatic generation is a
requirement.  Or you, personally, can post the zip and blog about it.  But
for us to promote it off of our website probably means we need PMC
approval first.

Thoughts?
-Alex

Re: 4.16.0 Windows Install issues (was Re: [ANNOUNCE] Apache Flex 4.16.0 released)

[piotrz]

Alex,

Can we exact specify if user face this problem ? Maybe we could do this zip
file, and notify user by popup - Hey you have to download 64 bit installer.

In the other words zip file is best option for me.

Piotr



-
Apache Flex PMC
piotrzarzyck...@gmail.com
--
View this message in context: 
http://apache-flex-development.247.n4.nabble.com/4-16-0-Windows-Install-issues-was-Re-ANNOUNCE-Apache-Flex-4-16-0-released-tp60626p60766.html
Sent from the Apache Flex Development mailing list archive at Nabble.com.

Re: 4.16.0 Windows Install issues (was Re: [ANNOUNCE] Apache Flex 4.16.0 released)

[Alex Harui]

Can you explain in more detail what steps you would take?

Thanks,
-Alex

On 3/20/17, 1:30 PM, "omup...@gmail.com on behalf of OmPrakash Muppirala"
 wrote:

>We could do a patch release that does not add any new feature for the 64
>bit installer binary.
>
>We would need to bump up the version so that the auto update functionality
>would be triggered.
>
>In this we probably don't even need a source release as we will only be
>modifying the convenience binaries.
>
>Thanks,
>Om
>
>
>On Mar 20, 2017 11:00 AM, "Alex Harui"  wrote:
>
>Renaming thread.  On next post I will drop the [ANNOUNCE] tag so we don't
>fill the inboxes of those using an Announce filter.
>
>So, I have heard back from my internal contacts.  There is no plan to
>allow the packaging of a 64-bit app the way we package the 32-bit
>Installer.  Essentially, that means that we have to write our own
>installer for the 64-bit Apache Flex SDK Installer.
>
>Also, the much larger size of AIR 25 is intentional.  There is a feature
>called "bitcode" that is responsible for the increase in size.
>
>Here are some possible solutions, of which we don't have to choose just
>one:
>1) New Installer release.  Try to add the call to clear() and maybe
>rewrite the way we do MD5 checking to not suck in the entire file at once.
> We are overdue for an Installer release.  We really should replace the
>way we get the mirror data with a JSON call.  But on the other hand, that
>will take us several days or weeks to do this, and I'd rather do other
>things.  But if volunteers can step up to do this, that would be great.
>It just won't solve the problem in the short term.
>
>2) Remove AIR 25 from the list of choices until we do #1
>
>3) Make the 64-bit version of the installer available as a zip file.
>Folks would have to download it, unpack it and run it instead of the usual
>installer.
>
>4) Add the installation of the 64-bit installer to the 32-bit Installer!
>I don't know for sure if we can do that, but we should be able to script
>the downloading and unpacking described in #3.
>
>Note that both #3 and #4 would probably not be official Apache releases.
>The 3.2 build script doesn't create a bundle, it only creates an installer
>EXE which will be 32-bit.  Usually, the binary artifact is the result of
>running the build script.  In order to create the 64-bit bundle, I had to
>tweak the build script which effectively changes the source package.  In
>order to do #4, we'd have to create the installer.xml for it.  So, we
>would not replace the existing 3.2 release or deploy the zip on the
>mirrors.  We'd have to host the zip somewhere else, more like making
>nightly builds available to those having problems.
>
>Thoughts?
>-Alex
>
>On 3/19/17, 10:02 PM, "Alex Harui"  wrote:
>
>>An update:  I was able to generate a 64-bit bundle, but not an installer
>>EXE that installs a 64-bit app.  It seems to generate a 32-bit app.
>>
>>Has anyone seen any info on creating an installer EXE for a 64-bit app?
>>I'll ask internally at Adobe as well.
>>
>>Thanks,
>>-Alex
>>
>>On 3/18/17, 9:33 PM, "Alex Harui"  wrote:
>>
>>>I haven't tried clear().  I'm trying to see if 64-bit will work because
>>>then I think we can just deploy a new binary instead of having to go
>>>through an entire release process.
>>>
>>>I spent some time this afternoon and could not get the AIR 25 SDK Beta
>>>with 64-bit support to actually generate a 64-bit exe.  Has anyone seen
>>>any documentation on the process?
>>>
>>>Thanks,
>>>-Alex
>>>
>>>On 3/18/17, 8:17 AM, "Josh Tynjala"  wrote:
>>>
I wonder if it would help to call clear() on the URLLoader's ByteArray
in
the Get AntOnAIR task after we save it to a file. While it's supposed
to
be
GCed eventually, clearing the memory immediately is probably a good
idea
with these large files. I've had succuss doing something similar in
cases
where BitmapData used too much cumulative memory before it could be
GCed.

- Josh

On Mar 18, 2017 7:24 AM, "Alex Harui"  wrote:

> Do any of the Windows platforms we support not handle 64-bit apps?
> Windows Server 2008 for example?
>
> Otherwise, it seems like we should rebuild the Windows Installer on a
> 64-bit AIR version and replace the 32-bit one that we've deployed?
>
> Thoughts?
> -Alex
>
> On 3/17/17, 3:46 PM, "piotrz"  wrote:
>
> >Hi :),
> >
> >Can someone post also this information to our Facebook fan page ?
> >
> >Thanks,
> >Piotr
> >
> >
> >
> >-
> >Apache Flex PMC
> >piotrzarzyck...@gmail.com
> >--
> >View this message in context:
> >https://na01.safelinks.protection.outlook.com/?url=
> http%3A%2F%2Fapache-fle
> >x-development.247.n4.nabble.com%2FANNOUNCE-Apache-
> 

Re: 4.16.0 Windows Install issues (was Re: [ANNOUNCE] Apache Flex 4.16.0 released)

[OmPrakash Muppirala]

We could do a patch release that does not add any new feature for the 64
bit installer binary.

We would need to bump up the version so that the auto update functionality
would be triggered.

In this we probably don't even need a source release as we will only be
modifying the convenience binaries.

Thanks,
Om


On Mar 20, 2017 11:00 AM, "Alex Harui"  wrote:

Renaming thread.  On next post I will drop the [ANNOUNCE] tag so we don't
fill the inboxes of those using an Announce filter.

So, I have heard back from my internal contacts.  There is no plan to
allow the packaging of a 64-bit app the way we package the 32-bit
Installer.  Essentially, that means that we have to write our own
installer for the 64-bit Apache Flex SDK Installer.

Also, the much larger size of AIR 25 is intentional.  There is a feature
called "bitcode" that is responsible for the increase in size.

Here are some possible solutions, of which we don't have to choose just
one:
1) New Installer release.  Try to add the call to clear() and maybe
rewrite the way we do MD5 checking to not suck in the entire file at once.
 We are overdue for an Installer release.  We really should replace the
way we get the mirror data with a JSON call.  But on the other hand, that
will take us several days or weeks to do this, and I'd rather do other
things.  But if volunteers can step up to do this, that would be great.
It just won't solve the problem in the short term.

2) Remove AIR 25 from the list of choices until we do #1

3) Make the 64-bit version of the installer available as a zip file.
Folks would have to download it, unpack it and run it instead of the usual
installer.

4) Add the installation of the 64-bit installer to the 32-bit Installer!
I don't know for sure if we can do that, but we should be able to script
the downloading and unpacking described in #3.

Note that both #3 and #4 would probably not be official Apache releases.
The 3.2 build script doesn't create a bundle, it only creates an installer
EXE which will be 32-bit.  Usually, the binary artifact is the result of
running the build script.  In order to create the 64-bit bundle, I had to
tweak the build script which effectively changes the source package.  In
order to do #4, we'd have to create the installer.xml for it.  So, we
would not replace the existing 3.2 release or deploy the zip on the
mirrors.  We'd have to host the zip somewhere else, more like making
nightly builds available to those having problems.

Thoughts?
-Alex

On 3/19/17, 10:02 PM, "Alex Harui"  wrote:

>An update:  I was able to generate a 64-bit bundle, but not an installer
>EXE that installs a 64-bit app.  It seems to generate a 32-bit app.
>
>Has anyone seen any info on creating an installer EXE for a 64-bit app?
>I'll ask internally at Adobe as well.
>
>Thanks,
>-Alex
>
>On 3/18/17, 9:33 PM, "Alex Harui"  wrote:
>
>>I haven't tried clear().  I'm trying to see if 64-bit will work because
>>then I think we can just deploy a new binary instead of having to go
>>through an entire release process.
>>
>>I spent some time this afternoon and could not get the AIR 25 SDK Beta
>>with 64-bit support to actually generate a 64-bit exe.  Has anyone seen
>>any documentation on the process?
>>
>>Thanks,
>>-Alex
>>
>>On 3/18/17, 8:17 AM, "Josh Tynjala"  wrote:
>>
>>>I wonder if it would help to call clear() on the URLLoader's ByteArray
>>>in
>>>the Get AntOnAIR task after we save it to a file. While it's supposed to
>>>be
>>>GCed eventually, clearing the memory immediately is probably a good idea
>>>with these large files. I've had succuss doing something similar in
>>>cases
>>>where BitmapData used too much cumulative memory before it could be
>>>GCed.
>>>
>>>- Josh
>>>
>>>On Mar 18, 2017 7:24 AM, "Alex Harui"  wrote:
>>>
 Do any of the Windows platforms we support not handle 64-bit apps?
 Windows Server 2008 for example?

 Otherwise, it seems like we should rebuild the Windows Installer on a
 64-bit AIR version and replace the 32-bit one that we've deployed?

 Thoughts?
 -Alex

 On 3/17/17, 3:46 PM, "piotrz"  wrote:

 >Hi :),
 >
 >Can someone post also this information to our Facebook fan page ?
 >
 >Thanks,
 >Piotr
 >
 >
 >
 >-
 >Apache Flex PMC
 >piotrzarzyck...@gmail.com
 >--
 >View this message in context:
 >https://na01.safelinks.protection.outlook.com/?url=
 http%3A%2F%2Fapache-fle
 >x-development.247.n4.nabble.com%2FANNOUNCE-Apache-
 Flex-4-16-0-released
 >-tp60427p60557.html=02%7C01%7C%7C526823ea8c5e49aa413108d46d88
 85d2%7Cf
 >a7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636253880511588257=
 VGA%2B5
 >OrpJvyyOLnKX9G4wEv587gNa%2Fa4NInd9G0Auuk%3D=0
 >Sent from the Apache Flex Development mailing list archive at
Nabble.com.


>>
>

4.16.0 Windows Install issues (was Re: [ANNOUNCE] Apache Flex 4.16.0 released)

[Alex Harui]

Renaming thread.  On next post I will drop the [ANNOUNCE] tag so we don't
fill the inboxes of those using an Announce filter.

So, I have heard back from my internal contacts.  There is no plan to
allow the packaging of a 64-bit app the way we package the 32-bit
Installer.  Essentially, that means that we have to write our own
installer for the 64-bit Apache Flex SDK Installer.

Also, the much larger size of AIR 25 is intentional.  There is a feature
called "bitcode" that is responsible for the increase in size.

Here are some possible solutions, of which we don't have to choose just
one:
1) New Installer release.  Try to add the call to clear() and maybe
rewrite the way we do MD5 checking to not suck in the entire file at once.
 We are overdue for an Installer release.  We really should replace the
way we get the mirror data with a JSON call.  But on the other hand, that
will take us several days or weeks to do this, and I'd rather do other
things.  But if volunteers can step up to do this, that would be great.
It just won't solve the problem in the short term.

2) Remove AIR 25 from the list of choices until we do #1

3) Make the 64-bit version of the installer available as a zip file.
Folks would have to download it, unpack it and run it instead of the usual
installer.

4) Add the installation of the 64-bit installer to the 32-bit Installer!
I don't know for sure if we can do that, but we should be able to script
the downloading and unpacking described in #3.

Note that both #3 and #4 would probably not be official Apache releases.
The 3.2 build script doesn't create a bundle, it only creates an installer
EXE which will be 32-bit.  Usually, the binary artifact is the result of
running the build script.  In order to create the 64-bit bundle, I had to
tweak the build script which effectively changes the source package.  In
order to do #4, we'd have to create the installer.xml for it.  So, we
would not replace the existing 3.2 release or deploy the zip on the
mirrors.  We'd have to host the zip somewhere else, more like making
nightly builds available to those having problems.

Thoughts?
-Alex

On 3/19/17, 10:02 PM, "Alex Harui"  wrote:

>An update:  I was able to generate a 64-bit bundle, but not an installer
>EXE that installs a 64-bit app.  It seems to generate a 32-bit app.
>
>Has anyone seen any info on creating an installer EXE for a 64-bit app?
>I'll ask internally at Adobe as well.
>
>Thanks,
>-Alex
>
>On 3/18/17, 9:33 PM, "Alex Harui"  wrote:
>
>>I haven't tried clear().  I'm trying to see if 64-bit will work because
>>then I think we can just deploy a new binary instead of having to go
>>through an entire release process.
>>
>>I spent some time this afternoon and could not get the AIR 25 SDK Beta
>>with 64-bit support to actually generate a 64-bit exe.  Has anyone seen
>>any documentation on the process?
>>
>>Thanks,
>>-Alex
>>
>>On 3/18/17, 8:17 AM, "Josh Tynjala"  wrote:
>>
>>>I wonder if it would help to call clear() on the URLLoader's ByteArray
>>>in
>>>the Get AntOnAIR task after we save it to a file. While it's supposed to
>>>be
>>>GCed eventually, clearing the memory immediately is probably a good idea
>>>with these large files. I've had succuss doing something similar in
>>>cases
>>>where BitmapData used too much cumulative memory before it could be
>>>GCed.
>>>
>>>- Josh
>>>
>>>On Mar 18, 2017 7:24 AM, "Alex Harui"  wrote:
>>>
 Do any of the Windows platforms we support not handle 64-bit apps?
 Windows Server 2008 for example?

 Otherwise, it seems like we should rebuild the Windows Installer on a
 64-bit AIR version and replace the 32-bit one that we've deployed?

 Thoughts?
 -Alex

 On 3/17/17, 3:46 PM, "piotrz"  wrote:

 >Hi :),
 >
 >Can someone post also this information to our Facebook fan page ?
 >
 >Thanks,
 >Piotr
 >
 >
 >
 >-
 >Apache Flex PMC
 >piotrzarzyck...@gmail.com
 >--
 >View this message in context:
 >https://na01.safelinks.protection.outlook.com/?url=
 http%3A%2F%2Fapache-fle
 >x-development.247.n4.nabble.com%2FANNOUNCE-Apache-
 Flex-4-16-0-released
 >-tp60427p60557.html=02%7C01%7C%7C526823ea8c5e49aa413108d46d88
 85d2%7Cf
 >a7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636253880511588257=
 VGA%2B5
 >OrpJvyyOLnKX9G4wEv587gNa%2Fa4NInd9G0Auuk%3D=0
 >Sent from the Apache Flex Development mailing list archive at
Nabble.com.


>>
>

Re: 4.16.0 Install issues (was Re: [ANNOUNCE] Apache Flex 4.16.0 released)

[OmPrakash Muppirala]

You can use
FP/AIR 23 and older then use one of the upgrade scripts.


+1 for this approach.

I think we should add these steps to a wiki page and keep it handy for
folks who want this.

In fact,  simply copying over a different AIR SDK should work as well.

Thanks,
Om

Re: 4.16.0 Install issues (was Re: [ANNOUNCE] Apache Flex 4.16.0 released)

[Alex Harui]

On 3/16/17, 12:25 PM, "Nicholas Kwiatkowski"
 wrote:

>I think there is some reason to use very old versions of AIR ( < 4), but
>between AIR 4 and 23 it should be ok.   Do you know if it effects the old
>ones (which still support some older OSs)?

Probably.  The issue is the installer script tries to see AIR SDK is a
tbz2 or dmg file.  The Installer uses ant_on_air which uses Ant 1.7 logic
and it fails to detect tbz2 and always tries the dmg route.  That's why
actually using Ant 1.9 works.

So, AFAICT, the situation is that the Installer can install 4.16.0 with
FP/AIR 24 and newer, and 4.15.0 and earlier with FP/AIR 23 and older.
4.16.0 with FP/AIR 23 and older can be installed via Ant 1.9 or newer, and
4.15.0 with FP/AIR 24 and newer cannot be installed at all.  You can use
FP/AIR 23 and older then use one of the upgrade scripts.

-Alex

Re: 4.16.0 Install issues (was Re: [ANNOUNCE] Apache Flex 4.16.0 released)

[Nicholas Kwiatkowski]

I think there is some reason to use very old versions of AIR ( < 4), but
between AIR 4 and 23 it should be ok.   Do you know if it effects the old
ones (which still support some older OSs)?

-Nick

On Thu, Mar 16, 2017 at 2:54 PM, Alex Harui  wrote:

> Adding dev@...
>
> I just noticed that there is a bug in the installer.xml and the Installer
> will not install 4.16.0 on Mac with FP/AIR 23 and older.  It will only
> work with FP/AIR 24 and newer.  I haven't tried Windows, but it should be
> ok.
>
> I am hopeful very few people will want to install 4.16 with FP/AIR 23 and
> older on Mac.  Does anybody disagree?  Otherwise I will make FP/AIR 24 the
> default.  Folks who have to have 4.16 with FP/AIR 23 and older can use a
> recent version of Ant and install via Ant.
>
> Thoughts?
> -Alex
>
> On 3/16/17, 10:48 AM, "Alex Harui"  wrote:
>
> >
> >
> >On 3/16/17, 10:20 AM, "pol2095"  wrote:
> >
> >>When I choose Air 25, Air 24 is downloaded ?
> >>
> >
> >Good catch.  I mis-typed something.  Please try again.
> >
> >-Alex
> >
>
>

4.16.0 Install issues (was Re: [ANNOUNCE] Apache Flex 4.16.0 released)

[Alex Harui]

Adding dev@...

I just noticed that there is a bug in the installer.xml and the Installer
will not install 4.16.0 on Mac with FP/AIR 23 and older.  It will only
work with FP/AIR 24 and newer.  I haven't tried Windows, but it should be
ok.

I am hopeful very few people will want to install 4.16 with FP/AIR 23 and
older on Mac.  Does anybody disagree?  Otherwise I will make FP/AIR 24 the
default.  Folks who have to have 4.16 with FP/AIR 23 and older can use a
recent version of Ant and install via Ant.

Thoughts?
-Alex

On 3/16/17, 10:48 AM, "Alex Harui"  wrote:

>
>
>On 3/16/17, 10:20 AM, "pol2095"  wrote:
>
>>When I choose Air 25, Air 24 is downloaded ?
>>
>
>Good catch.  I mis-typed something.  Please try again.
>
>-Alex
>

Re: [FlexJS} CreateJS Externs (was Re: Current FlexJS license/notice issues)

[Alex Harui]

On 3/14/17, 11:33 PM, "Justin Mclean"  wrote:

>Hi,
>
>> In order to determine which steps I had to go through to try to donate
>>the
>> CreateJS Externs to the CreateJS community, it turns out I had to get a
>> ruling from the Adobe legal team about who owned the "externs" files we
>> generate from CreateJS source.  The Adobe legal team determined that
>>these
>> externs files are copyright Adobe
>
>If that is the case that may need to be pointed out in the LICENSE/NOTICE?
>
>Do the ASF headers need to change on those files?
>
>I’d suggest confirming what is a correct course of action on legal
>discuss.

I'm not sure I understand your point.  Every line of code I have committed
to an ASF repo is copyright Adobe and donated to the ASF and doesn't
require any other special handling in LICENSE and NOTICE.

Thanks,
-Alex

Re: [FlexJS} CreateJS Externs (was Re: Current FlexJS license/notice issues)

[Justin Mclean]

Hi,

> In order to determine which steps I had to go through to try to donate the
> CreateJS Externs to the CreateJS community, it turns out I had to get a
> ruling from the Adobe legal team about who owned the "externs" files we
> generate from CreateJS source.  The Adobe legal team determined that these
> externs files are copyright Adobe

It would probably simplify things if you got Adobe to supply a LICENSE and 
NOTICE file for the files copyright by Adobe.

Thanks,
Justin

Re: [FlexJS} CreateJS Externs (was Re: Current FlexJS license/notice issues)

[Justin Mclean]

Hi,

> In order to determine which steps I had to go through to try to donate the
> CreateJS Externs to the CreateJS community, it turns out I had to get a
> ruling from the Adobe legal team about who owned the "externs" files we
> generate from CreateJS source.  The Adobe legal team determined that these
> externs files are copyright Adobe

If that is the case that may need to be pointed out in the LICENSE/NOTICE?

Do the ASF headers need to change on those files?

I’d suggest confirming what is a correct course of action on legal discuss.

Thanks,
Justin

[FlexJS} CreateJS Externs (was Re: Current FlexJS license/notice issues)

[Alex Harui]

Justin's fix for CreateJS typedefs on Unix reminded me...

In order to determine which steps I had to go through to try to donate the
CreateJS Externs to the CreateJS community, it turns out I had to get a
ruling from the Adobe legal team about who owned the "externs" files we
generate from CreateJS source.  The Adobe legal team determined that these
externs files are copyright Adobe because I wrote the scripts that
generate them, and that externs files are a list of APIs instead of
functional code and thus not a derivative work.  Apparently there is some
sort of "list" exception.  I think that's why other externs authors also
claim copyrights and licenses different from the library the externs
represent.

So, now the question is, given these files are correctly donated to Apache
is it worth it to try to donate them to CreateJS?  I've got plenty of
other things to spend my time on.

Thoughts?
-Alex 

On 10/21/16, 10:31 PM, "Alex Harui"  wrote:

>
>
>On 10/21/16, 9:59 PM, "Justin Mclean"  wrote:
>
>>
>>> Still trying to get a response from my legal department.  My employment
>>> agreement requires a lot of hoops to jump through before anything I
>>>create
>>> goes to another entity.
>>
>>Understand that but I'd assume if it's already under an Apache license
>>and at the ASF then that would not apply?
>
>I'm only authorized to donate Flex-related code to the ASF.  I'm not
>currently authorized to donate to any other communities.  Adobe owns the
>copyright of anything considered to be my work.
>
>>
>>Otherwise it wouldn’t be compatible with the Apache license as it would
>>have further restrictions on it.
>
>The ASF doesn't like taking code that isn't donated.  I gotta go through
>this process so CreateJS isn't seen as taking code.  You can keep asking
>questions about it, but that's the steps I have to take.
>
>-Alex
>

Re: [Flex JS] Font weight issues

[Alex Harui]

SimpleCSSStylesImpl is Simple.  FontWeight is expecting bold or normal
since that's what Flash supports for free and is less code.

Hopefully someday someone will write AdvancedCSSStylesImpl.

Thanks for pointing it out though,
-Alex

On 3/2/17, 9:45 PM, "Justin Mclean"  wrote:

>Hi,
>
>Just a FYI in case this trips anyone else up.
>
>The code in applyStyle in SimpleCSSValuesImpl assumes that most CSS style
>numbers are pixels - this is not be the case.
>
>The code in questions is:
>if (typeof(value) == 'number') {
>if (colorStyles[p])
>value = CSSUtils.attributeFromColor(value);
>else
>value = value.toString() + 'px';
>}
>
>Which for font-weight means you can't use values
>100,200,300,400,500,600,700,800 or 900.
>
>Thanks,
>Justin
>
>

[Flex JS] Font weight issues

[Justin Mclean]

Hi,

Just a FYI in case this trips anyone else up.

The code in applyStyle in SimpleCSSValuesImpl assumes that most CSS style 
numbers are pixels - this is not be the case.

The code in questions is:
if (typeof(value) == 'number') {
if (colorStyles[p])
value = CSSUtils.attributeFromColor(value);
else
value = value.toString() + 'px';
}

Which for font-weight means you can't use values 
100,200,300,400,500,600,700,800 or 900.

Thanks,
Justin