NPM Issues (was Re: [Discussion] "Life after FlashPlayer" post @flex lists)

[Alex Harui]

OK, but what file would contain that string such that it shows up then?

Thanks,
-Alex

On 5/1/18, 9:32 PM, "OmPrakash Muppirala"  wrote:

On Tue, May 1, 2018, 8:13 PM Alex Harui  wrote:

> That's my windows machine that I used for 0.9.2.
>
> Where did it say 0.9.3?


It shows up after the npm install completes.


I'm trying to remember if I published the package to NPM or if you did.
> Maybe the scripts are running or were run after some versions got updated
> in the repos.  I have not run a build on my windows machine since the
> release.


>
>
>
>
> -Alex
>
> On 5/1/18, 2:14 PM, "omup...@gmail.com on behalf of OmPrakash Muppirala"
>  wrote:
>
> On Tue, May 1, 2018 at 1:36 PM, Alex Harui 
> wrote:
>
> > Om,
> >
> > I got the warnings when we released 0.9.2 but it appeared to leave a
> > functional SDK.  Now it is hanging for me and Dany.
> >
> > I don't think I've run a publish and I don't think the CI server is
> > either, but you never know.  Try grabbing one of the SWCs.  Unzip it
> and
> > swfdump the library.swf.  There should be debugfile bytecodes with
> paths to
> > the source that give us a clue.
> >
> >
> Good idea.  Here is what I see in swfdump:
>
> 2) + 0:1 debugfile
>
> 
"C:\temp\rc\royale-asjs\frameworks\projects\Collections\src\main\royale;org\apache\royale\collections;FlattenedList.as"
>
> Any idea whose machine this could have been?
>
> Thanks,
> Om
>
>
> > -Alex
> >
> > On 5/1/18, 10:49 AM, "omup...@gmail.com on behalf of OmPrakash
> > Muppirala" 
> wrote:
> >
> > On Tue, May 1, 2018 at 8:57 AM, Alex Harui
> 
> > wrote:
> >
> > > Hi Dany,
> > >
> > > The link is directly to a tar.gz built by our CI server.  No
> humans
> > > involved, and no special packaging for NPM.  NPM says it can
> install
> > any
> > > tar.gz file.  Try downloading the file at that link and using
> Mac
> > Finder to
> > > expand it into a folder.  For me it expands without errors.
> This is
> > why I
> > > believe it is some incompatibility with NPM.
> > >
> > > If the tar.gz expands for you on Mac, you can try to "npm
> install"
> > the
> > > folder.  Running:
> > >
> > > sudo npm install -g 
> > >
> > > seemed to work for me.
> > >
> > > -Alex
> > >
> >
> > I just tried
> > npm install @apache-royale/royale-js -g
> > on windwos and it installed just fine.
> >
> > Although, I notice that  @apache-royale/royale-js@0.9.3 was
> > installed.  It
> > should have installed 0.9.2.  Did we publish 0.9.3 to npm by
> mistake?
> > Could that be related?
> >
> > Thanks,
> > Om
> >
> >
> > >
> > > On 5/1/18, 1:51 AM, "Dany Dhondt"  wrote:
> > >
> > > Hi Alex,
> > >
> > > I tried the direct link you provided: same result
> (hundreds of
> > > ‘invalid entry’ messages)
> > >
> > > My version of npm is 6.0.0
> > >
> > > downgrading to a lower npm version is not an option at 
this
> > moment
> > > because I’m on a production mac with crucial projects on it.
> > >
> > > Amids all the invalid entry messages, now and then is a
> somewhat
> > > longer message like
> > >
> > > WARN⸨   ░░░⸩ ⠸ extract:@apache-royale
> /royale-js:
> > verb
> > > lock using /Users/ tar invalid entry
> > >
> > >
> > >
> > >
> > > Is it Om that created this package? This might give him an
> idea
> > of
> > > what’s going on?
> > >
> > > -Dany
> > >
> > >
> > >
> > > > Op 30 apr. 2018, om 18:39 heeft Alex Harui
> >  > > > het volgende geschreven:
> > > >
> > > > Hmm.  I'm getting that too on my Mac.  It used to
> 

Re: [Discussion] "Life after FlashPlayer" post @flex lists (Was: Royale Frameworks and More)

[OmPrakash Muppirala]

On Tue, May 1, 2018, 8:13 PM Alex Harui  wrote:

> That's my windows machine that I used for 0.9.2.
>
> Where did it say 0.9.3?


It shows up after the npm install completes.


I'm trying to remember if I published the package to NPM or if you did.
> Maybe the scripts are running or were run after some versions got updated
> in the repos.  I have not run a build on my windows machine since the
> release.


>
>
>
>
> -Alex
>
> On 5/1/18, 2:14 PM, "omup...@gmail.com on behalf of OmPrakash Muppirala"
>  wrote:
>
> On Tue, May 1, 2018 at 1:36 PM, Alex Harui 
> wrote:
>
> > Om,
> >
> > I got the warnings when we released 0.9.2 but it appeared to leave a
> > functional SDK.  Now it is hanging for me and Dany.
> >
> > I don't think I've run a publish and I don't think the CI server is
> > either, but you never know.  Try grabbing one of the SWCs.  Unzip it
> and
> > swfdump the library.swf.  There should be debugfile bytecodes with
> paths to
> > the source that give us a clue.
> >
> >
> Good idea.  Here is what I see in swfdump:
>
> 2) + 0:1 debugfile
>
> "C:\temp\rc\royale-asjs\frameworks\projects\Collections\src\main\royale;org\apache\royale\collections;FlattenedList.as"
>
> Any idea whose machine this could have been?
>
> Thanks,
> Om
>
>
> > -Alex
> >
> > On 5/1/18, 10:49 AM, "omup...@gmail.com on behalf of OmPrakash
> > Muppirala" 
> wrote:
> >
> > On Tue, May 1, 2018 at 8:57 AM, Alex Harui
> 
> > wrote:
> >
> > > Hi Dany,
> > >
> > > The link is directly to a tar.gz built by our CI server.  No
> humans
> > > involved, and no special packaging for NPM.  NPM says it can
> install
> > any
> > > tar.gz file.  Try downloading the file at that link and using
> Mac
> > Finder to
> > > expand it into a folder.  For me it expands without errors.
> This is
> > why I
> > > believe it is some incompatibility with NPM.
> > >
> > > If the tar.gz expands for you on Mac, you can try to "npm
> install"
> > the
> > > folder.  Running:
> > >
> > > sudo npm install -g 
> > >
> > > seemed to work for me.
> > >
> > > -Alex
> > >
> >
> > I just tried
> > npm install @apache-royale/royale-js -g
> > on windwos and it installed just fine.
> >
> > Although, I notice that  @apache-royale/royale-js@0.9.3 was
> > installed.  It
> > should have installed 0.9.2.  Did we publish 0.9.3 to npm by
> mistake?
> > Could that be related?
> >
> > Thanks,
> > Om
> >
> >
> > >
> > > On 5/1/18, 1:51 AM, "Dany Dhondt"  wrote:
> > >
> > > Hi Alex,
> > >
> > > I tried the direct link you provided: same result
> (hundreds of
> > > ‘invalid entry’ messages)
> > >
> > > My version of npm is 6.0.0
> > >
> > > downgrading to a lower npm version is not an option at this
> > moment
> > > because I’m on a production mac with crucial projects on it.
> > >
> > > Amids all the invalid entry messages, now and then is a
> somewhat
> > > longer message like
> > >
> > > WARN⸨   ░░░⸩ ⠸ extract:@apache-royale
> /royale-js:
> > verb
> > > lock using /Users/ tar invalid entry
> > >
> > >
> > >
> > >
> > > Is it Om that created this package? This might give him an
> idea
> > of
> > > what’s going on?
> > >
> > > -Dany
> > >
> > >
> > >
> > > > Op 30 apr. 2018, om 18:39 heeft Alex Harui
> >  > > > het volgende geschreven:
> > > >
> > > > Hmm.  I'm getting that too on my Mac.  It used to
> complete
> > anyway
> > > and work, but now it appears to be hanging.
> > > >
> > > > The internet hints that it might have to do with the
> version
> > of NPM
> > > used to create the package vs the  NPM we are using to
> install.  Om,
> > did
> > > you create the packages for publishing to NPM?  If so, what
> version
> > did you
> > > use?
> > > >
> > > > On the other hand, just installing the nightly build also
> > gives a
> > > ton of those warnings but seems to complete and produce a
> runnable
> > compiler.
> > > >
> > > > sudo npm install -g
> 

Re: [Discussion] "Life after FlashPlayer" post @flex lists (Was: Royale Frameworks and More)

[Alex Harui]

That's my windows machine that I used for 0.9.2.

Where did it say 0.9.3?  I'm trying to remember if I published the package to 
NPM or if you did.  Maybe the scripts are running or were run after some 
versions got updated in the repos.  I have not run a build on my windows 
machine since the release.

-Alex

On 5/1/18, 2:14 PM, "omup...@gmail.com on behalf of OmPrakash Muppirala" 
 wrote:

On Tue, May 1, 2018 at 1:36 PM, Alex Harui  wrote:

> Om,
>
> I got the warnings when we released 0.9.2 but it appeared to leave a
> functional SDK.  Now it is hanging for me and Dany.
>
> I don't think I've run a publish and I don't think the CI server is
> either, but you never know.  Try grabbing one of the SWCs.  Unzip it and
> swfdump the library.swf.  There should be debugfile bytecodes with paths 
to
> the source that give us a clue.
>
>
Good idea.  Here is what I see in swfdump:

2) + 0:1 debugfile

"C:\temp\rc\royale-asjs\frameworks\projects\Collections\src\main\royale;org\apache\royale\collections;FlattenedList.as"

Any idea whose machine this could have been?

Thanks,
Om


> -Alex
>
> On 5/1/18, 10:49 AM, "omup...@gmail.com on behalf of OmPrakash
> Muppirala"  wrote:
>
> On Tue, May 1, 2018 at 8:57 AM, Alex Harui 
> wrote:
>
> > Hi Dany,
> >
> > The link is directly to a tar.gz built by our CI server.  No humans
> > involved, and no special packaging for NPM.  NPM says it can install
> any
> > tar.gz file.  Try downloading the file at that link and using Mac
> Finder to
> > expand it into a folder.  For me it expands without errors.  This is
> why I
> > believe it is some incompatibility with NPM.
> >
> > If the tar.gz expands for you on Mac, you can try to "npm install"
> the
> > folder.  Running:
> >
> > sudo npm install -g 
> >
> > seemed to work for me.
> >
> > -Alex
> >
>
> I just tried
> npm install @apache-royale/royale-js -g
> on windwos and it installed just fine.
>
> Although, I notice that  @apache-royale/royale-js@0.9.3 was
> installed.  It
> should have installed 0.9.2.  Did we publish 0.9.3 to npm by mistake?
> Could that be related?
>
> Thanks,
> Om
>
>
> >
> > On 5/1/18, 1:51 AM, "Dany Dhondt"  wrote:
> >
> > Hi Alex,
> >
> > I tried the direct link you provided: same result (hundreds of
> > ‘invalid entry’ messages)
> >
> > My version of npm is 6.0.0
> >
> > downgrading to a lower npm version is not an option at this
> moment
> > because I’m on a production mac with crucial projects on it.
> >
> > Amids all the invalid entry messages, now and then is a somewhat
> > longer message like
> >
> > WARN⸨   ░░░⸩ ⠸ extract:@apache-royale/royale-js:
> verb
> > lock using /Users/ tar invalid entry
> >
> >
> >
> >
> > Is it Om that created this package? This might give him an idea
> of
> > what’s going on?
> >
> > -Dany
> >
> >
> >
> > > Op 30 apr. 2018, om 18:39 heeft Alex Harui
>  > > het volgende geschreven:
> > >
> > > Hmm.  I'm getting that too on my Mac.  It used to complete
> anyway
> > and work, but now it appears to be hanging.
> > >
> > > The internet hints that it might have to do with the version
> of NPM
> > used to create the package vs the  NPM we are using to install.  Om,
> did
> > you create the packages for publishing to NPM?  If so, what version
> did you
> > use?
> > >
> > > On the other hand, just installing the nightly build also
> gives a
> > ton of those warnings but seems to complete and produce a runnable
> compiler.
> > >
> > > sudo npm install -g 
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fna01.safelinks=02%7C01%7Caharui%40adobe.com%7Ca57dd23c4db24e0a18d008d5afa88c6b%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636608060969636900=sgZBuZoOmRwla%2Bti%2FwtU0vRemJMy7MjB5sRC9d0Qdic%3D=0.
> protection.outlook.com/?url=https%3A%2F%2Fna01.safelinks&
> data=02%7C01%7Caharui%40adobe.com%7C00079fd32a194a4caf1b08d5af8bd09e%
> 

Re: [Discussion] "Life after FlashPlayer" post @flex lists (Was: Royale Frameworks and More)

[OmPrakash Muppirala]

On Tue, May 1, 2018 at 1:36 PM, Alex Harui  wrote:

> Om,
>
> I got the warnings when we released 0.9.2 but it appeared to leave a
> functional SDK.  Now it is hanging for me and Dany.
>
> I don't think I've run a publish and I don't think the CI server is
> either, but you never know.  Try grabbing one of the SWCs.  Unzip it and
> swfdump the library.swf.  There should be debugfile bytecodes with paths to
> the source that give us a clue.
>
>
Good idea.  Here is what I see in swfdump:

2) + 0:1 debugfile
"C:\temp\rc\royale-asjs\frameworks\projects\Collections\src\main\royale;org\apache\royale\collections;FlattenedList.as"

Any idea whose machine this could have been?

Thanks,
Om


> -Alex
>
> On 5/1/18, 10:49 AM, "omup...@gmail.com on behalf of OmPrakash
> Muppirala"  wrote:
>
> On Tue, May 1, 2018 at 8:57 AM, Alex Harui 
> wrote:
>
> > Hi Dany,
> >
> > The link is directly to a tar.gz built by our CI server.  No humans
> > involved, and no special packaging for NPM.  NPM says it can install
> any
> > tar.gz file.  Try downloading the file at that link and using Mac
> Finder to
> > expand it into a folder.  For me it expands without errors.  This is
> why I
> > believe it is some incompatibility with NPM.
> >
> > If the tar.gz expands for you on Mac, you can try to "npm install"
> the
> > folder.  Running:
> >
> > sudo npm install -g 
> >
> > seemed to work for me.
> >
> > -Alex
> >
>
> I just tried
> npm install @apache-royale/royale-js -g
> on windwos and it installed just fine.
>
> Although, I notice that  @apache-royale/royale-js@0.9.3 was
> installed.  It
> should have installed 0.9.2.  Did we publish 0.9.3 to npm by mistake?
> Could that be related?
>
> Thanks,
> Om
>
>
> >
> > On 5/1/18, 1:51 AM, "Dany Dhondt"  wrote:
> >
> > Hi Alex,
> >
> > I tried the direct link you provided: same result (hundreds of
> > ‘invalid entry’ messages)
> >
> > My version of npm is 6.0.0
> >
> > downgrading to a lower npm version is not an option at this
> moment
> > because I’m on a production mac with crucial projects on it.
> >
> > Amids all the invalid entry messages, now and then is a somewhat
> > longer message like
> >
> > WARN⸨   ░░░⸩ ⠸ extract:@apache-royale/royale-js:
> verb
> > lock using /Users/ tar invalid entry
> >
> >
> >
> >
> > Is it Om that created this package? This might give him an idea
> of
> > what’s going on?
> >
> > -Dany
> >
> >
> >
> > > Op 30 apr. 2018, om 18:39 heeft Alex Harui
>  > > het volgende geschreven:
> > >
> > > Hmm.  I'm getting that too on my Mac.  It used to complete
> anyway
> > and work, but now it appears to be hanging.
> > >
> > > The internet hints that it might have to do with the version
> of NPM
> > used to create the package vs the  NPM we are using to install.  Om,
> did
> > you create the packages for publishing to NPM?  If so, what version
> did you
> > use?
> > >
> > > On the other hand, just installing the nightly build also
> gives a
> > ton of those warnings but seems to complete and produce a runnable
> compiler.
> > >
> > > sudo npm install -g https://na01.safelinks.
> protection.outlook.com/?url=https%3A%2F%2Fna01.safelinks&
> data=02%7C01%7Caharui%40adobe.com%7C00079fd32a194a4caf1b08d5af8bd09e%
> 7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636607937396983186=D%
> 2BedD6Pv8Eh7GSR8q1dLpGxi6RL5R43uCdchzXoECD0%3D=0.
> > protection.outlook.com/?url=http%3A%2F%2Fapacheroyaleci.
> > westus2.cloudapp.azure.com%3A8080%2Fjob%2Froyale-asjs_
> > jsonly%2FlastSuccessfulBuild%2Fartifact%2Fout%2Fapache-
> > royale-0.9.3-bin-js.tar.gz=02%7C01%7Caharui%40adobe.com%
> > 7C0c405d2ff9134638451608d5af40c2f1%7Cfa7b1b5a7b34438794aed2c178de
> > cee1%7C0%7C0%7C636607615050264364=zKrYmme3fourKgDPSyULPEiiR5%
> > 2FfXBUXmGQ49qdY0aM%3D=0  protection.outlook.com/?url=https%3A%2F%2Fna01.safelinks&
> data=02%7C01%7Caharui%40adobe.com%7C00079fd32a194a4caf1b08d5af8bd09e%
> 7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636607937396983186=D%
> 2BedD6Pv8Eh7GSR8q1dLpGxi6RL5R43uCdchzXoECD0%3D=0.
> > protection.outlook.com/?url=http%3A%2F%2Fapacheroyaleci.
> > westus2.cloudapp.azure.com%3A8080%2Fjob%2Froyale-asjs_
> > jsonly%2FlastSuccessfulBuild%2Fartifact%2Fout%2Fapache-
> > royale-0.9.3-bin-js.tar.gz=02%7C01%7Caharui%40adobe.com%
> > 7C0c405d2ff9134638451608d5af40c2f1%7Cfa7b1b5a7b34438794aed2c178de
> > cee1%7C0%7C0%7C636607615050264364=zKrYmme3fourKgDPSyULPEiiR5%
>  

Re: [Discussion] "Life after FlashPlayer" post @flex lists (Was: Royale Frameworks and More)

[Alex Harui]

Om,

I got the warnings when we released 0.9.2 but it appeared to leave a functional 
SDK.  Now it is hanging for me and Dany.

I don't think I've run a publish and I don't think the CI server is either, but 
you never know.  Try grabbing one of the SWCs.  Unzip it and swfdump the 
library.swf.  There should be debugfile bytecodes with paths to the source that 
give us a clue.

-Alex

On 5/1/18, 10:49 AM, "omup...@gmail.com on behalf of OmPrakash Muppirala" 
 wrote:

On Tue, May 1, 2018 at 8:57 AM, Alex Harui  wrote:

> Hi Dany,
>
> The link is directly to a tar.gz built by our CI server.  No humans
> involved, and no special packaging for NPM.  NPM says it can install any
> tar.gz file.  Try downloading the file at that link and using Mac Finder 
to
> expand it into a folder.  For me it expands without errors.  This is why I
> believe it is some incompatibility with NPM.
>
> If the tar.gz expands for you on Mac, you can try to "npm install" the
> folder.  Running:
>
> sudo npm install -g 
>
> seemed to work for me.
>
> -Alex
>

I just tried
npm install @apache-royale/royale-js -g
on windwos and it installed just fine.

Although, I notice that  @apache-royale/royale-js@0.9.3 was installed.  It
should have installed 0.9.2.  Did we publish 0.9.3 to npm by mistake?
Could that be related?

Thanks,
Om


>
> On 5/1/18, 1:51 AM, "Dany Dhondt"  wrote:
>
> Hi Alex,
>
> I tried the direct link you provided: same result (hundreds of
> ‘invalid entry’ messages)
>
> My version of npm is 6.0.0
>
> downgrading to a lower npm version is not an option at this moment
> because I’m on a production mac with crucial projects on it.
>
> Amids all the invalid entry messages, now and then is a somewhat
> longer message like
>
> WARN⸨   ░░░⸩ ⠸ extract:@apache-royale/royale-js: verb
> lock using /Users/ tar invalid entry
>
>
>
>
> Is it Om that created this package? This might give him an idea of
> what’s going on?
>
> -Dany
>
>
>
> > Op 30 apr. 2018, om 18:39 heeft Alex Harui  > het volgende geschreven:
> >
> > Hmm.  I'm getting that too on my Mac.  It used to complete anyway
> and work, but now it appears to be hanging.
> >
> > The internet hints that it might have to do with the version of NPM
> used to create the package vs the  NPM we are using to install.  Om, did
> you create the packages for publishing to NPM?  If so, what version did 
you
> use?
> >
> > On the other hand, just installing the nightly build also gives a
> ton of those warnings but seems to complete and produce a runnable 
compiler.
> >
> > sudo npm install -g 
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fna01.safelinks=02%7C01%7Caharui%40adobe.com%7C00079fd32a194a4caf1b08d5af8bd09e%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636607937396983186=D%2BedD6Pv8Eh7GSR8q1dLpGxi6RL5R43uCdchzXoECD0%3D=0.
> protection.outlook.com/?url=http%3A%2F%2Fapacheroyaleci.
> westus2.cloudapp.azure.com%3A8080%2Fjob%2Froyale-asjs_
> jsonly%2FlastSuccessfulBuild%2Fartifact%2Fout%2Fapache-
> royale-0.9.3-bin-js.tar.gz=02%7C01%7Caharui%40adobe.com%
> 7C0c405d2ff9134638451608d5af40c2f1%7Cfa7b1b5a7b34438794aed2c178de
> cee1%7C0%7C0%7C636607615050264364=zKrYmme3fourKgDPSyULPEiiR5%
> 2FfXBUXmGQ49qdY0aM%3D=0 
 protection.outlook.com/?url=http%3A%2F%2Fapacheroyaleci.
> westus2.cloudapp.azure.com%3A8080%2Fjob%2Froyale-asjs_
> jsonly%2FlastSuccessfulBuild%2Fartifact%2Fout%2Fapache-
> royale-0.9.3-bin-js.tar.gz=02%7C01%7Caharui%40adobe.com%
> 7C0c405d2ff9134638451608d5af40c2f1%7Cfa7b1b5a7b34438794aed2c178de
> cee1%7C0%7C0%7C636607615050264364=zKrYmme3fourKgDPSyULPEiiR5%
> 2FfXBUXmGQ49qdY0aM%3D=0>
> >
> > I haven't tried downgrading or upgrading my version of NPM (5.6).
> Dany which version are you running?  Can you try different versions of 
NPM?
> >
> > HTH,
> > -Alex
> >
> > On 4/30/18, 1:11 AM, "Dany Dhondt" > wrote:
> >
> >Hi Om,
> >
> >I did that!
> >
> >In the mean time I tried all sorts of things:
> >install 0.9.1 instead of 0.9.3
> >

Re: States doesn't work properly anymore

[Alex Harui]

An element without a royale_wrapper breaks the fundamental contract in UIBase.  
I recommend that you figure that part out.

Good luck,
-Alex

On 5/1/18, 12:33 PM, "Piotr Zarzycki"  wrote:

In the other words - My website is loaded and component have initial state.

Than I'm clicking on the button I'm changing state for that component.

Piotr

On Tue, May 1, 2018, 7:27 PM Piotr Zarzycki 
wrote:

> This component is already on the display list. I'm changing States of that
> component. That's the weirdest thing.
>
> Piotr
>
> On Tue, May 1, 2018, 7:06 PM Alex Harui  wrote:
>
>> Hi Piotr,
>>
>> What is the call stack at that point?  Should the component be on the
>> display list or is it still having its properties set up per the
>> lifecycle?  If it is just having its properties set up, there should be
>> another call to setClassName in addedToParent and parent should not be
>> undefined.
>>
>> -Alex
>>
>> On 5/1/18, 9:58 AM, "piotrz"  wrote:
>>
>> Alex,
>>
>> I just checked getter for parent and royale_wrapper on element is
>> undefined.
>>
>> <
>> 
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fapache-royale-development.20373.n8.nabble.com%2Ffile%2Ft1%2Froyale_wrapper_undefined.png=02%7C01%7Caharui%40adobe.com%7C4df3cf45055440294eb908d5af84cc3e%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636607907300284114=%2BWvX%2B84oVUsWZO4Q3rO4QwPRnR1TcZJKseXYJB1ni8U%3D=0>
>>
>>
>> Piotr
>>
>>
>>
>>
>>
>> --
>> Sent from:
>> 
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fapache-royale-development.20373.n8.nabble.com%2F=02%7C01%7Caharui%40adobe.com%7C4df3cf45055440294eb908d5af84cc3e%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636607907300284114=W4AV3qDPNfO1VlCywDh5bd5u9Oa7Fh0JXDhzrFIp%2FMM%3D=0
>>
>>
>>

Re: [Discussion] "Life after FlashPlayer" post @flex lists (Was: Royale Frameworks and More)

[Dany Dhondt]

Hi Om

I also noticed that 0.9.3 was installeren so I tried ...@0.9.2 but that was non 
existing. ...@0.9.1 did exist but gave the same result/errors

-Dany

Verstuurd vanaf mijn iPhone

> Op 1 mei 2018 om 19:48 heeft OmPrakash Muppirala  het 
> volgende geschreven:
> 
>> On Tue, May 1, 2018 at 8:57 AM, Alex Harui  wrote:
>> 
>> Hi Dany,
>> 
>> The link is directly to a tar.gz built by our CI server.  No humans
>> involved, and no special packaging for NPM.  NPM says it can install any
>> tar.gz file.  Try downloading the file at that link and using Mac Finder to
>> expand it into a folder.  For me it expands without errors.  This is why I
>> believe it is some incompatibility with NPM.
>> 
>> If the tar.gz expands for you on Mac, you can try to "npm install" the
>> folder.  Running:
>> 
>>sudo npm install -g 
>> 
>> seemed to work for me.
>> 
>> -Alex
>> 
> 
> I just tried
> npm install @apache-royale/royale-js -g
> on windwos and it installed just fine.
> 
> Although, I notice that  @apache-royale/royale-js@0.9.3 was installed.  It
> should have installed 0.9.2.  Did we publish 0.9.3 to npm by mistake?
> Could that be related?
> 
> Thanks,
> Om
> 
> 
>> 
>> On 5/1/18, 1:51 AM, "Dany Dhondt"  wrote:
>> 
>>Hi Alex,
>> 
>>I tried the direct link you provided: same result (hundreds of
>> ‘invalid entry’ messages)
>> 
>>My version of npm is 6.0.0
>> 
>>downgrading to a lower npm version is not an option at this moment
>> because I’m on a production mac with crucial projects on it.
>> 
>>Amids all the invalid entry messages, now and then is a somewhat
>> longer message like
>> 
>>WARN⸨   ░░░⸩ ⠸ extract:@apache-royale/royale-js: verb
>> lock using /Users/ tar invalid entry
>> 
>> 
>> 
>> 
>>Is it Om that created this package? This might give him an idea of
>> what’s going on?
>> 
>>-Dany
>> 
>> 
>> 
>>> Op 30 apr. 2018, om 18:39 heeft Alex Harui > > het volgende geschreven:
>>> 
>>> Hmm.  I'm getting that too on my Mac.  It used to complete anyway
>> and work, but now it appears to be hanging.
>>> 
>>> The internet hints that it might have to do with the version of NPM
>> used to create the package vs the  NPM we are using to install.  Om, did
>> you create the packages for publishing to NPM?  If so, what version did you
>> use?
>>> 
>>> On the other hand, just installing the nightly build also gives a
>> ton of those warnings but seems to complete and produce a runnable compiler.
>>> 
>>> sudo npm install -g https://na01.safelinks.
>> protection.outlook.com/?url=http%3A%2F%2Fapacheroyaleci.
>> westus2.cloudapp.azure.com%3A8080%2Fjob%2Froyale-asjs_
>> jsonly%2FlastSuccessfulBuild%2Fartifact%2Fout%2Fapache-
>> royale-0.9.3-bin-js.tar.gz=02%7C01%7Caharui%40adobe.com%
>> 7C0c405d2ff9134638451608d5af40c2f1%7Cfa7b1b5a7b34438794aed2c178de
>> cee1%7C0%7C0%7C636607615050264364=zKrYmme3fourKgDPSyULPEiiR5%
>> 2FfXBUXmGQ49qdY0aM%3D=0 > protection.outlook.com/?url=http%3A%2F%2Fapacheroyaleci.
>> westus2.cloudapp.azure.com%3A8080%2Fjob%2Froyale-asjs_
>> jsonly%2FlastSuccessfulBuild%2Fartifact%2Fout%2Fapache-
>> royale-0.9.3-bin-js.tar.gz=02%7C01%7Caharui%40adobe.com%
>> 7C0c405d2ff9134638451608d5af40c2f1%7Cfa7b1b5a7b34438794aed2c178de
>> cee1%7C0%7C0%7C636607615050264364=zKrYmme3fourKgDPSyULPEiiR5%
>> 2FfXBUXmGQ49qdY0aM%3D=0>
>>> 
>>> I haven't tried downgrading or upgrading my version of NPM (5.6).
>> Dany which version are you running?  Can you try different versions of NPM?
>>> 
>>> HTH,
>>> -Alex
>>> 
>>> On 4/30/18, 1:11 AM, "Dany Dhondt"  archeme...@mac.com>> wrote:
>>> 
>>>   Hi Om,
>>> 
>>>   I did that!
>>> 
>>>   In the mean time I tried all sorts of things:
>>>   install 0.9.1 instead of 0.9.3
>>>   reinstalled npm
>>> 
>>>   nothing works
>>> 
>>>   -Dany
>>> 
>>> 
 Op 30 apr. 2018, om 09:59 heeft OmPrakash Muppirala <
>> bigosma...@gmail.com > het volgende
>> geschreven:
 
 Dany,
 
 I haven't tested this on a Mac.  I am told that Mac users need to
>> use "sudo
 install npm  " depending on how they installed npm.
>> Perhaps
 you can try that out?
 
 Thanks,
 Om
 
 On Mon, Apr 30, 2018 at 12:54 AM, Dany Dhondt > > wrote:
 
> Hi Om,
> 
> Thanks for pointing that out.
> 
> I’ll have a go with that scenario.
> 
> 1. (re) installing royale
> npm install @apache-royale/royale-js -g
> WARN tar invalid entry extract:@apache-royale/royale-js: verb
>> lock using
> /Users/dany/.npm/_locks/staging-3a08f0df5026584d.lock for
> /usr/local/lib/node_modules/.staging
> WARN tar invalid entry
> 
> Then hundreds of ‘invalid entry’ lines in the terminal.
> 
> And finally
> 
> …
> 
> WARN tar invalid entry

Re: States doesn't work properly anymore

[Piotr Zarzycki]

In the other words - My website is loaded and component have initial state.

Than I'm clicking on the button I'm changing state for that component.

Piotr

On Tue, May 1, 2018, 7:27 PM Piotr Zarzycki 
wrote:

> This component is already on the display list. I'm changing States of that
> component. That's the weirdest thing.
>
> Piotr
>
> On Tue, May 1, 2018, 7:06 PM Alex Harui  wrote:
>
>> Hi Piotr,
>>
>> What is the call stack at that point?  Should the component be on the
>> display list or is it still having its properties set up per the
>> lifecycle?  If it is just having its properties set up, there should be
>> another call to setClassName in addedToParent and parent should not be
>> undefined.
>>
>> -Alex
>>
>> On 5/1/18, 9:58 AM, "piotrz"  wrote:
>>
>> Alex,
>>
>> I just checked getter for parent and royale_wrapper on element is
>> undefined.
>>
>> <
>> https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fapache-royale-development.20373.n8.nabble.com%2Ffile%2Ft1%2Froyale_wrapper_undefined.png=02%7C01%7Caharui%40adobe.com%7C4df3cf45055440294eb908d5af84cc3e%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636607907300284114=%2BWvX%2B84oVUsWZO4Q3rO4QwPRnR1TcZJKseXYJB1ni8U%3D=0>
>>
>>
>> Piotr
>>
>>
>>
>>
>>
>> --
>> Sent from:
>> https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fapache-royale-development.20373.n8.nabble.com%2F=02%7C01%7Caharui%40adobe.com%7C4df3cf45055440294eb908d5af84cc3e%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636607907300284114=W4AV3qDPNfO1VlCywDh5bd5u9Oa7Fh0JXDhzrFIp%2FMM%3D=0
>>
>>
>>

Re: Null pointer errors

[Piotr Zarzycki]

Harbs,

+1 for a new committer! :)

Piotr

2018-05-01 19:44 GMT+02:00 Harbs :

> Yeah. I was planning on spending time on the compiler, but time has not
> been recently as forthcoming as I was hoping… ;-)
>
> FWIW, my daughter is finishing school soon and I’m going to have her try
> and learn the compiler. I’ll see how that goes…
>
> Harbs
>
> > On May 1, 2018, at 7:53 PM, Alex Harui  wrote:
> >
> > Sorry,  I thought the link was to the source so I didn't click on it.
> >
> > This could be a good exercise for anyone wanting to learn more about the
> compiler to figure out how to catch this situation (maybe by detecting the
> AST for "void 0") and outputting an error instead of an NPE, or swapping in
> "undefined".
> >
> > -Alex
> >
> > On 5/1/18, 9:35 AM, "Harbs"  wrote:
> >
> >I attached a paste of the stack in my email.
> >
> >I just finished resolving all my errors. It was hard to trace them
> down, but I just discovered that there were uses of "void 0" in the code. I
> replaced all those cases with “undefined" and poof! All my errors went
> away! :-)
> >
> >My theory is that use of void 0 was confusing the compiler and
> causing errors elsewhere… I’m not sure how to demonstrate this…
> >
> >Harbs
> >
> >(P.S. Sorry I’ve been so quiet lately. Life has been pretty busy… ;) )
> >
> >> On May 1, 2018, at 7:03 PM, Alex Harui 
> wrote:
> >>
> >> What is the call stack for the null pointer?
> >>
> >> On 5/1/18, 5:05 AM, "Harbs"  wrote:
> >>
> >>   I’m trying to compile a library using comic and I’m getting a **LOT**
> of null pointer errors. They seem to be related to missing type
> information, although I’m not completely sure why the missing type
> information is causing these errors.
> >>
> >>   Some of the errors are resolved-able by adding type info, but not all.
> >>
> >>   I picked one at random to track down the issue:
> >>   https://na01.safelinks.protection.outlook.com/?url=
> https%3A%2F%2Fpaste.apache.org%2F8pLB=02%7C01%7Caharui%40adobe.com%
> 7C31fb16c8d65948a50da008d5af5bd2c6%7Cfa7b1b5a7b34438794aed2c178de
> cee1%7C0%7C0%7C636607731304191116=ik%2FnmWaooQD%
> 2BwLT6VM1JN1ypa4BdOXe8VS1NfyF5FUw%3D=0  protection.outlook.com/?url=https%3A%2F%2Fpaste.apache.
> org%2F8pLB=02%7C01%7Caharui%40adobe.com%
> 7C31fb16c8d65948a50da008d5af5bd2c6%7Cfa7b1b5a7b34438794aed2c178de
> cee1%7C0%7C0%7C636607731304191116=ik%2FnmWaooQD%
> 2BwLT6VM1JN1ypa4BdOXe8VS1NfyF5FUw%3D=0>
> >>
> >>   The line in question looks like this:
> >>   var codepoint:int = glyphs[i].codePoints[0];
> >>
> >>   glyphs[I] is a “GlyphInfo” instance. Either of the following two
> changes make this particular error go away:
> >>   var codepoint:* = glyphs[i].codePoints[0];
> >>   var codepoint:int = (glyphs[i] as GlyphInfo).codePoints[0];
> >>
> >>   I’m not really sure why these changes are necessary though…
> >>
> >>   Thoughts?
> >>   Harbs
> >>
> >
> >
> >
>
>


-- 

Piotr Zarzycki

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

Re: [Discussion] "Life after FlashPlayer" post @flex lists (Was: Royale Frameworks and More)

[OmPrakash Muppirala]

On Tue, May 1, 2018 at 8:57 AM, Alex Harui  wrote:

> Hi Dany,
>
> The link is directly to a tar.gz built by our CI server.  No humans
> involved, and no special packaging for NPM.  NPM says it can install any
> tar.gz file.  Try downloading the file at that link and using Mac Finder to
> expand it into a folder.  For me it expands without errors.  This is why I
> believe it is some incompatibility with NPM.
>
> If the tar.gz expands for you on Mac, you can try to "npm install" the
> folder.  Running:
>
> sudo npm install -g 
>
> seemed to work for me.
>
> -Alex
>

I just tried
npm install @apache-royale/royale-js -g
on windwos and it installed just fine.

Although, I notice that  @apache-royale/royale-js@0.9.3 was installed.  It
should have installed 0.9.2.  Did we publish 0.9.3 to npm by mistake?
Could that be related?

Thanks,
Om


>
> On 5/1/18, 1:51 AM, "Dany Dhondt"  wrote:
>
> Hi Alex,
>
> I tried the direct link you provided: same result (hundreds of
> ‘invalid entry’ messages)
>
> My version of npm is 6.0.0
>
> downgrading to a lower npm version is not an option at this moment
> because I’m on a production mac with crucial projects on it.
>
> Amids all the invalid entry messages, now and then is a somewhat
> longer message like
>
> WARN⸨   ░░░⸩ ⠸ extract:@apache-royale/royale-js: verb
> lock using /Users/ tar invalid entry
>
>
>
>
> Is it Om that created this package? This might give him an idea of
> what’s going on?
>
> -Dany
>
>
>
> > Op 30 apr. 2018, om 18:39 heeft Alex Harui  > het volgende geschreven:
> >
> > Hmm.  I'm getting that too on my Mac.  It used to complete anyway
> and work, but now it appears to be hanging.
> >
> > The internet hints that it might have to do with the version of NPM
> used to create the package vs the  NPM we are using to install.  Om, did
> you create the packages for publishing to NPM?  If so, what version did you
> use?
> >
> > On the other hand, just installing the nightly build also gives a
> ton of those warnings but seems to complete and produce a runnable compiler.
> >
> > sudo npm install -g https://na01.safelinks.
> protection.outlook.com/?url=http%3A%2F%2Fapacheroyaleci.
> westus2.cloudapp.azure.com%3A8080%2Fjob%2Froyale-asjs_
> jsonly%2FlastSuccessfulBuild%2Fartifact%2Fout%2Fapache-
> royale-0.9.3-bin-js.tar.gz=02%7C01%7Caharui%40adobe.com%
> 7C0c405d2ff9134638451608d5af40c2f1%7Cfa7b1b5a7b34438794aed2c178de
> cee1%7C0%7C0%7C636607615050264364=zKrYmme3fourKgDPSyULPEiiR5%
> 2FfXBUXmGQ49qdY0aM%3D=0  protection.outlook.com/?url=http%3A%2F%2Fapacheroyaleci.
> westus2.cloudapp.azure.com%3A8080%2Fjob%2Froyale-asjs_
> jsonly%2FlastSuccessfulBuild%2Fartifact%2Fout%2Fapache-
> royale-0.9.3-bin-js.tar.gz=02%7C01%7Caharui%40adobe.com%
> 7C0c405d2ff9134638451608d5af40c2f1%7Cfa7b1b5a7b34438794aed2c178de
> cee1%7C0%7C0%7C636607615050264364=zKrYmme3fourKgDPSyULPEiiR5%
> 2FfXBUXmGQ49qdY0aM%3D=0>
> >
> > I haven't tried downgrading or upgrading my version of NPM (5.6).
> Dany which version are you running?  Can you try different versions of NPM?
> >
> > HTH,
> > -Alex
> >
> > On 4/30/18, 1:11 AM, "Dany Dhondt" > wrote:
> >
> >Hi Om,
> >
> >I did that!
> >
> >In the mean time I tried all sorts of things:
> >install 0.9.1 instead of 0.9.3
> >reinstalled npm
> >
> >nothing works
> >
> >-Dany
> >
> >
> >> Op 30 apr. 2018, om 09:59 heeft OmPrakash Muppirala <
> bigosma...@gmail.com > het volgende
> geschreven:
> >>
> >> Dany,
> >>
> >> I haven't tested this on a Mac.  I am told that Mac users need to
> use "sudo
> >> install npm  " depending on how they installed npm.
> Perhaps
> >> you can try that out?
> >>
> >> Thanks,
> >> Om
> >>
> >> On Mon, Apr 30, 2018 at 12:54 AM, Dany Dhondt  > wrote:
> >>
> >>> Hi Om,
> >>>
> >>> Thanks for pointing that out.
> >>>
> >>> I’ll have a go with that scenario.
> >>>
> >>> 1. (re) installing royale
> >>> npm install @apache-royale/royale-js -g
> >>> WARN tar invalid entry extract:@apache-royale/royale-js: verb
> lock using
> >>> /Users/dany/.npm/_locks/staging-3a08f0df5026584d.lock for
> >>> /usr/local/lib/node_modules/.staging
> >>> WARN tar invalid entry
> >>>
> >>> Then hundreds of ‘invalid entry’ lines in the terminal.
> >>>
> >>> And finally
> >>>
> >>> …
> >>>
> >>> WARN tar invalid entry
> >>> /usr/local/bin/asjscompc -> /usr/local/lib/node_modules/@
> >>> apache-royale/royale-js/royale-asjs/js/bin/asjscompcnpm
> >>> /usr/local/bin/asjsc -> 

Re: Null pointer errors

[Harbs]

Yeah. I was planning on spending time on the compiler, but time has not been 
recently as forthcoming as I was hoping… ;-)

FWIW, my daughter is finishing school soon and I’m going to have her try and 
learn the compiler. I’ll see how that goes…

Harbs

> On May 1, 2018, at 7:53 PM, Alex Harui  wrote:
> 
> Sorry,  I thought the link was to the source so I didn't click on it.
> 
> This could be a good exercise for anyone wanting to learn more about the 
> compiler to figure out how to catch this situation (maybe by detecting the 
> AST for "void 0") and outputting an error instead of an NPE, or swapping in 
> "undefined".
> 
> -Alex
> 
> On 5/1/18, 9:35 AM, "Harbs"  wrote:
> 
>I attached a paste of the stack in my email.
> 
>I just finished resolving all my errors. It was hard to trace them down, 
> but I just discovered that there were uses of "void 0" in the code. I 
> replaced all those cases with “undefined" and poof! All my errors went away! 
> :-)
> 
>My theory is that use of void 0 was confusing the compiler and causing 
> errors elsewhere… I’m not sure how to demonstrate this…
> 
>Harbs
> 
>(P.S. Sorry I’ve been so quiet lately. Life has been pretty busy… ;) )
> 
>> On May 1, 2018, at 7:03 PM, Alex Harui  wrote:
>> 
>> What is the call stack for the null pointer?
>> 
>> On 5/1/18, 5:05 AM, "Harbs"  wrote:
>> 
>>   I’m trying to compile a library using comic and I’m getting a **LOT** of 
>> null pointer errors. They seem to be related to missing type information, 
>> although I’m not completely sure why the missing type information is causing 
>> these errors.
>> 
>>   Some of the errors are resolved-able by adding type info, but not all.
>> 
>>   I picked one at random to track down the issue:
>>   
>> https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpaste.apache.org%2F8pLB=02%7C01%7Caharui%40adobe.com%7C31fb16c8d65948a50da008d5af5bd2c6%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636607731304191116=ik%2FnmWaooQD%2BwLT6VM1JN1ypa4BdOXe8VS1NfyF5FUw%3D=0
>>  
>> 
>> 
>>   The line in question looks like this:
>>   var codepoint:int = glyphs[i].codePoints[0];
>> 
>>   glyphs[I] is a “GlyphInfo” instance. Either of the following two changes 
>> make this particular error go away:
>>   var codepoint:* = glyphs[i].codePoints[0];
>>   var codepoint:int = (glyphs[i] as GlyphInfo).codePoints[0];
>> 
>>   I’m not really sure why these changes are necessary though…
>> 
>>   Thoughts?
>>   Harbs
>> 
> 
> 
> 

Re: States doesn't work properly anymore

[Piotr Zarzycki]

This component is already on the display list. I'm changing States of that
component. That's the weirdest thing.

Piotr

On Tue, May 1, 2018, 7:06 PM Alex Harui  wrote:

> Hi Piotr,
>
> What is the call stack at that point?  Should the component be on the
> display list or is it still having its properties set up per the
> lifecycle?  If it is just having its properties set up, there should be
> another call to setClassName in addedToParent and parent should not be
> undefined.
>
> -Alex
>
> On 5/1/18, 9:58 AM, "piotrz"  wrote:
>
> Alex,
>
> I just checked getter for parent and royale_wrapper on element is
> undefined.
>
> <
> https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fapache-royale-development.20373.n8.nabble.com%2Ffile%2Ft1%2Froyale_wrapper_undefined.png=02%7C01%7Caharui%40adobe.com%7C4df3cf45055440294eb908d5af84cc3e%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636607907300284114=%2BWvX%2B84oVUsWZO4Q3rO4QwPRnR1TcZJKseXYJB1ni8U%3D=0>
>
>
> Piotr
>
>
>
>
>
> --
> Sent from:
> https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fapache-royale-development.20373.n8.nabble.com%2F=02%7C01%7Caharui%40adobe.com%7C4df3cf45055440294eb908d5af84cc3e%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636607907300284114=W4AV3qDPNfO1VlCywDh5bd5u9Oa7Fh0JXDhzrFIp%2FMM%3D=0
>
>
>

Re: Null pointer errors

[Alex Harui]

Sorry,  I thought the link was to the source so I didn't click on it.

This could be a good exercise for anyone wanting to learn more about the 
compiler to figure out how to catch this situation (maybe by detecting the AST 
for "void 0") and outputting an error instead of an NPE, or swapping in 
"undefined".

-Alex

On 5/1/18, 9:35 AM, "Harbs"  wrote:

I attached a paste of the stack in my email.

I just finished resolving all my errors. It was hard to trace them down, 
but I just discovered that there were uses of "void 0" in the code. I replaced 
all those cases with “undefined" and poof! All my errors went away! :-)

My theory is that use of void 0 was confusing the compiler and causing 
errors elsewhere… I’m not sure how to demonstrate this…

Harbs

(P.S. Sorry I’ve been so quiet lately. Life has been pretty busy… ;) )

> On May 1, 2018, at 7:03 PM, Alex Harui  wrote:
> 
> What is the call stack for the null pointer?
> 
> On 5/1/18, 5:05 AM, "Harbs"  wrote:
> 
>I’m trying to compile a library using comic and I’m getting a **LOT** 
of null pointer errors. They seem to be related to missing type information, 
although I’m not completely sure why the missing type information is causing 
these errors.
> 
>Some of the errors are resolved-able by adding type info, but not all.
> 
>I picked one at random to track down the issue:
>
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpaste.apache.org%2F8pLB=02%7C01%7Caharui%40adobe.com%7C31fb16c8d65948a50da008d5af5bd2c6%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636607731304191116=ik%2FnmWaooQD%2BwLT6VM1JN1ypa4BdOXe8VS1NfyF5FUw%3D=0
 

> 
>The line in question looks like this:
>var codepoint:int = glyphs[i].codePoints[0];
> 
>glyphs[I] is a “GlyphInfo” instance. Either of the following two 
changes make this particular error go away:
>var codepoint:* = glyphs[i].codePoints[0];
>var codepoint:int = (glyphs[i] as GlyphInfo).codePoints[0];
> 
>I’m not really sure why these changes are necessary though…
> 
>Thoughts?
>Harbs
> 

Re: Null pointer errors

[Harbs]

I attached a paste of the stack in my email.

I just finished resolving all my errors. It was hard to trace them down, but I 
just discovered that there were uses of "void 0" in the code. I replaced all 
those cases with “undefined" and poof! All my errors went away! :-)

My theory is that use of void 0 was confusing the compiler and causing errors 
elsewhere… I’m not sure how to demonstrate this…

Harbs

(P.S. Sorry I’ve been so quiet lately. Life has been pretty busy… ;) )

> On May 1, 2018, at 7:03 PM, Alex Harui  wrote:
> 
> What is the call stack for the null pointer?
> 
> On 5/1/18, 5:05 AM, "Harbs"  wrote:
> 
>I’m trying to compile a library using comic and I’m getting a **LOT** of 
> null pointer errors. They seem to be related to missing type information, 
> although I’m not completely sure why the missing type information is causing 
> these errors.
> 
>Some of the errors are resolved-able by adding type info, but not all.
> 
>I picked one at random to track down the issue:
>
> https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpaste.apache.org%2F8pLB=02%7C01%7Caharui%40adobe.com%7C31fb16c8d65948a50da008d5af5bd2c6%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636607731304191116=ik%2FnmWaooQD%2BwLT6VM1JN1ypa4BdOXe8VS1NfyF5FUw%3D=0
>  
> 
> 
>The line in question looks like this:
>var codepoint:int = glyphs[i].codePoints[0];
> 
>glyphs[I] is a “GlyphInfo” instance. Either of the following two changes 
> make this particular error go away:
>var codepoint:* = glyphs[i].codePoints[0];
>var codepoint:int = (glyphs[i] as GlyphInfo).codePoints[0];
> 
>I’m not really sure why these changes are necessary though…
> 
>Thoughts?
>Harbs
> 

Re: Null pointer errors

[Alex Harui]

What is the call stack for the null pointer?

On 5/1/18, 5:05 AM, "Harbs"  wrote:

I’m trying to compile a library using comic and I’m getting a **LOT** of 
null pointer errors. They seem to be related to missing type information, 
although I’m not completely sure why the missing type information is causing 
these errors.

Some of the errors are resolved-able by adding type info, but not all.

I picked one at random to track down the issue:

https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpaste.apache.org%2F8pLB=02%7C01%7Caharui%40adobe.com%7C31fb16c8d65948a50da008d5af5bd2c6%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636607731304191116=ik%2FnmWaooQD%2BwLT6VM1JN1ypa4BdOXe8VS1NfyF5FUw%3D=0
 


The line in question looks like this:
var codepoint:int = glyphs[i].codePoints[0];

glyphs[I] is a “GlyphInfo” instance. Either of the following two changes 
make this particular error go away:
var codepoint:* = glyphs[i].codePoints[0];
var codepoint:int = (glyphs[i] as GlyphInfo).codePoints[0];

I’m not really sure why these changes are necessary though…

Thoughts?
Harbs

Re: States doesn't work properly anymore

[Alex Harui]

Because setClassNames can be expensive, we only want to do it if we have to.  
The component lifecycle is designed for setting lots of properties before 
adding a component to its parent.  There is a call to setClassNames in 
addedToParent for that reason, so other calls to set the className shouldn't 
bother to do all of that work.

To me, it is far more worrisome that the parent property is undefined.  We need 
to understand why that is the case.  Lots of other things could break due to 
that.  If you look at the component's element and check the element's 
parentNode, is it defined?  If so, then the parent getter should not be 
returning undefined.  Step into the parent getter and try to see why.

-Alex

On 5/1/18, 2:05 AM, "Piotr Zarzycki"  wrote:

Alex,

I did more experiments to that, but not sure whether it's ok. In those two
places I had parent - undefined [1] (UIBase) [2] (ClassSelectorList). As
you can see I commented out those condition and everything started to work
properly.

The question is - Why those condition even there here ? Did you or Carlos
had and scenario where something was crashing without it ?

[1] 
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpaste.apache.org%2FOTab=02%7C01%7Caharui%40adobe.com%7Ce2e3e44ca87a4d025c6208d5af42afdd%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636607623322558881=0qoiIUEIyA4DCoVF1BrHBSFVP1IOrLS9H95d35qwstk%3D=0
[2] 
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpaste.apache.org%2FSv1K=02%7C01%7Caharui%40adobe.com%7Ce2e3e44ca87a4d025c6208d5af42afdd%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636607623322568889=G6T9atRqzQwDOtPEjp6mkGRUXZE%2FZy2yY%2F95leIPQXo%3D=0

Thanks,
Piotr

2018-04-29 20:02 GMT+02:00 Piotr Zarzycki :

> Alex,
>
> I did look into that but unfortunately adding ClassListSelector didn't
> change anything. Still setting className dynamically won't change it
> because parent in UIBase is undefined. I did find that loop while was
> infinite in some circumstances. I have pushed fix, not sure if it's
> correct.
>
> You were saying that parent in your browser was not undefined. I did check
> even on my Mac on Safari and in UIBase it's undefined all the time. Maybe
> states are messing up with something and current system do not handle
> everything as it was before.
>
> I updated website [1] with committed changes. Switching between sub pages
> reproduce the issue.
>
> [1] 
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Ftranspiledactionscript.com%2Fdev%2Fexamples%2F=02%7C01%7Caharui%40adobe.com%7Ce2e3e44ca87a4d025c6208d5af42afdd%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636607623322568889=9vrin7DQVxaTK94dnfZkTpL8s4LIbB7fte8E%2BkjHJKI%3D=0
> TranspiledActionScript/
>
> Thanks,
> Piotr
>
> 2018-04-29 11:25 GMT+02:00 Piotr Zarzycki :
>
>> Hi Alex,
>>
>> Ok I will look into that.
>>
>> Thanks,
>> Piotr
>>
>>
>> On Sun, Apr 29, 2018, 8:12 AM Alex Harui 
>> wrote:
>>
>>> Hi Piotr,
>>>
>>> For me, parent is not undefined when the new value of snippetBackground
>>> is being applied.  The first time I debugged through the code it looked
>>> like was undefined, but when I stepped through the parent getter it
>>> returned a valid result.  I'm not sure if we can trust the value 
tooltips
>>> for our getters.
>>>
>>> I think the problem is that I did not upgrade the className
>>> implementation in MDL when I updated UIBase and the Jewel code.  MDL is
>>> still using the ClassList class and it should probably do what Jewel is
>>> doing and use ClassSelectorList instead.   I think if that works for MDL
>>> then we can drop ClassList from the code base.
>>>
>>> I would like you to do the upgrade since it will be easier for you to
>>> test it and make any further adjustments.
>>>
>>> Thanks,
>>> -Alex
>>>
>>> On 4/28/18, 10:02 PM, "Alex Harui"  wrote:
>>>
>>> I will look into it.
>>>
>>> On 4/28/18, 2:08 PM, "Carlos Rovira" 
>>> wrote:
>>>
>>> Sorry but as I said I'm few days off so I can't look at it. Hope
>>> you can
>>> see the root of the problem since I was planning to use states
>>> in jewel
>>> example to do better organization of different examples
>>>
>>>
>>> El El sáb, 28 abr 2018 a las 9:58, Piotr Zarzycki <
>>> piotrzarzyck...@gmail.com>
>>> escribió:
>>>
>>> > Carlos,
>>> >
>>> > Any thoughts ? Maybe it make sense that parent is undefined
>>> since view is
>>> > changing