You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@royale.apache.org by Josh Tynjala <jo...@bowlerhat.dev> on 2024/10/09 17:09:36 UTC
Super method calls not currently checked for errors
Hey all,
A user of vscode-as3mxml pointed out to me recently that the arguments
passed to super() and super.method() calls are not currently checked for
errors by our compiler. That includes checking the types of the arguments,
and even whether the number of arguments is actually correct! However,
regular function calls without super are correctly type checked, so it is
only in cases of inheritance where super is used.
I have royale-compiler fixes ready to go.
This bug has been in the compiler since the code was donated by Adobe. So
it's definitely not even remotely new, and it doesn't necessarily need to
delay the current release in progress.
However, it may impact a lot of existing code once the compiler can start
detecting the wrong types. Several of our framework libraries in
royale-asjs are affected. It's likely that user projects, especially ones
with a lot of inheritance, will be affected too.
In some cases, it just requires an extra cast. However, in a few cases that
I've seen so far, it is actually showing that there are bugs in our code.
I'm working through the necessary fixes in royale-asjs now.
I just wanted to give a heads up about this, as it may have some impact on
everyone's projects.
Any thoughts on pushing this into the 0.9.12? Or would you rather wait for
the next one?
--
Josh Tynjala
Bowler Hat LLC <https://bowlerhat.dev>
Re: Super method calls not currently checked for errors
Posted by Josh Tynjala <jo...@bowlerhat.dev>.
Quick followup. I finished fixing issues in royale-asjs a bit after
starting this thread.
About 15 files in the royale-asjs/framework libraries are affected. Of
those, 6 contained bugs. The rest were passing a variable of type Object
where a class was expected, and they were simply missing a cast.
All royale-asjs examples build successfully, which is heartening.
Real-world apps may have more complex inheritance than our examples, though.
Piotr, I'll try to push some branches to royale-compiler and royale-asjs
later today.
--
Josh Tynjala
Bowler Hat LLC <https://bowlerhat.dev>
On Wed, Oct 9, 2024 at 10:09 AM Josh Tynjala <jo...@bowlerhat.dev>
wrote:
> Hey all,
>
> A user of vscode-as3mxml pointed out to me recently that the arguments
> passed to super() and super.method() calls are not currently checked for
> errors by our compiler. That includes checking the types of the arguments,
> and even whether the number of arguments is actually correct! However,
> regular function calls without super are correctly type checked, so it is
> only in cases of inheritance where super is used.
>
> I have royale-compiler fixes ready to go.
>
> This bug has been in the compiler since the code was donated by Adobe. So
> it's definitely not even remotely new, and it doesn't necessarily need to
> delay the current release in progress.
>
> However, it may impact a lot of existing code once the compiler can start
> detecting the wrong types. Several of our framework libraries in
> royale-asjs are affected. It's likely that user projects, especially ones
> with a lot of inheritance, will be affected too.
>
> In some cases, it just requires an extra cast. However, in a few cases
> that I've seen so far, it is actually showing that there are bugs in our
> code. I'm working through the necessary fixes in royale-asjs now.
>
> I just wanted to give a heads up about this, as it may have some impact on
> everyone's projects.
>
> Any thoughts on pushing this into the 0.9.12? Or would you rather wait for
> the next one?
>
> --
> Josh Tynjala
> Bowler Hat LLC <https://bowlerhat.dev>
>
Re: Super method calls not currently checked for errors
Posted by Piotr Zarzycki <pi...@gmail.com>.
Hi Josh,
I'm sorry for the delay. It looks like in our application everything looks
fine.
Thanks Harbs for tip with ignorecoercion - it helps. :)
Thanks,
Piotr
sob., 12 paź 2024 o 01:06 Maria Jose Esteve <mj...@iest.com> napisał(a):
> Switching to 0.9.12-SNAPSHOT and a small tweak to a component.
> I compile the SDK with ant and Maven and my projects with Maven.
> Everything compiled OK: debug and release
>
> Hiedra
>
> -----Mensaje original-----
> De: Maria Jose Esteve <mj...@iest.com>
> Enviado el: viernes, 11 de octubre de 2024 20:46
> Para: dev@royale.apache.org
> Asunto: RE: Super method calls not currently checked for errors
>
> Hi, I'll also test my applications. It will also be a good test bench for
> programmers who have more difficulty programming 😝
>
> Hiedra
>
> -----Mensaje original-----
> De: Harbs <ha...@gmail.com>
> Enviado el: viernes, 11 de octubre de 2024 9:53
> Para: Apache Royale Development <de...@royale.apache.org>
> Asunto: Re: Super method calls not currently checked for errors
>
> I’ll test this against my apps next week.
>
> I’m on the fence about whether we should push it into the current release.
>
> > On Oct 9, 2024, at 11:53 PM, Josh Tynjala <jo...@bowlerhat.dev>
> wrote:
> >
> > I too tend to be extra wary to make changes close to a release.
> > Sometimes, even seemingly minor bug fixes can introduce unexpected side
> effects!
> >
> > I totally don't mind waiting to merge this stuff.
> >
> > --
> > Josh Tynjala
> > Bowler Hat LLC <https://bowlerhat.dev>
> >
> >
> > On Wed, Oct 9, 2024 at 11:20 AM Andrew Wetmore <co...@gmail.com>
> wrote:
> >
> >> I am always scared of putting a major, necessary fix that might have
> >> downstream impact into a release that is about to go out the door. I
> >> would vote for making this part, perhaps the major part, of the next
> >> release, and using the time between this one and that one to find and
> >> fix as much as we can in our own code...and maybe prepare some
> >> examples with fixes to help Royale users.
> >>
> >> On Wed, Oct 9, 2024 at 2:11 PM Josh Tynjala
> >> <jo...@bowlerhat.dev>
> >> wrote:
> >>
> >>> Hey all,
> >>>
> >>> A user of vscode-as3mxml pointed out to me recently that the
> >>> arguments passed to super() and super.method() calls are not
> >>> currently checked for errors by our compiler. That includes checking
> >>> the types of the
> >> arguments,
> >>> and even whether the number of arguments is actually correct!
> >>> However, regular function calls without super are correctly type
> >>> checked, so it is only in cases of inheritance where super is used.
> >>>
> >>> I have royale-compiler fixes ready to go.
> >>>
> >>> This bug has been in the compiler since the code was donated by
> >>> Adobe. So it's definitely not even remotely new, and it doesn't
> >>> necessarily need to delay the current release in progress.
> >>>
> >>> However, it may impact a lot of existing code once the compiler can
> >>> start detecting the wrong types. Several of our framework libraries
> >>> in royale-asjs are affected. It's likely that user projects,
> >>> especially ones with a lot of inheritance, will be affected too.
> >>>
> >>> In some cases, it just requires an extra cast. However, in a few
> >>> cases
> >> that
> >>> I've seen so far, it is actually showing that there are bugs in our
> code.
> >>> I'm working through the necessary fixes in royale-asjs now.
> >>>
> >>> I just wanted to give a heads up about this, as it may have some
> >>> impact
> >> on
> >>> everyone's projects.
> >>>
> >>> Any thoughts on pushing this into the 0.9.12? Or would you rather
> >>> wait
> >> for
> >>> the next one?
> >>>
> >>> --
> >>> Josh Tynjala
> >>> Bowler Hat LLC <https://bowlerhat.dev>
> >>>
> >>
> >>
> >> --
> >> Andrew Wetmore
> >>
> >> Editor, Moose House Publications <https://moosehousepress.com/>
> >> Editor-Writer, The Apache Software Foundation <https://apache.org/>
> >>
>
>
--
Piotr Zarzycki
RE: Super method calls not currently checked for errors
Posted by Maria Jose Esteve <mj...@iest.com>.
Switching to 0.9.12-SNAPSHOT and a small tweak to a component.
I compile the SDK with ant and Maven and my projects with Maven.
Everything compiled OK: debug and release
Hiedra
-----Mensaje original-----
De: Maria Jose Esteve <mj...@iest.com>
Enviado el: viernes, 11 de octubre de 2024 20:46
Para: dev@royale.apache.org
Asunto: RE: Super method calls not currently checked for errors
Hi, I'll also test my applications. It will also be a good test bench for programmers who have more difficulty programming 😝
Hiedra
-----Mensaje original-----
De: Harbs <ha...@gmail.com>
Enviado el: viernes, 11 de octubre de 2024 9:53
Para: Apache Royale Development <de...@royale.apache.org>
Asunto: Re: Super method calls not currently checked for errors
I’ll test this against my apps next week.
I’m on the fence about whether we should push it into the current release.
> On Oct 9, 2024, at 11:53 PM, Josh Tynjala <jo...@bowlerhat.dev> wrote:
>
> I too tend to be extra wary to make changes close to a release.
> Sometimes, even seemingly minor bug fixes can introduce unexpected side effects!
>
> I totally don't mind waiting to merge this stuff.
>
> --
> Josh Tynjala
> Bowler Hat LLC <https://bowlerhat.dev>
>
>
> On Wed, Oct 9, 2024 at 11:20 AM Andrew Wetmore <co...@gmail.com> wrote:
>
>> I am always scared of putting a major, necessary fix that might have
>> downstream impact into a release that is about to go out the door. I
>> would vote for making this part, perhaps the major part, of the next
>> release, and using the time between this one and that one to find and
>> fix as much as we can in our own code...and maybe prepare some
>> examples with fixes to help Royale users.
>>
>> On Wed, Oct 9, 2024 at 2:11 PM Josh Tynjala
>> <jo...@bowlerhat.dev>
>> wrote:
>>
>>> Hey all,
>>>
>>> A user of vscode-as3mxml pointed out to me recently that the
>>> arguments passed to super() and super.method() calls are not
>>> currently checked for errors by our compiler. That includes checking
>>> the types of the
>> arguments,
>>> and even whether the number of arguments is actually correct!
>>> However, regular function calls without super are correctly type
>>> checked, so it is only in cases of inheritance where super is used.
>>>
>>> I have royale-compiler fixes ready to go.
>>>
>>> This bug has been in the compiler since the code was donated by
>>> Adobe. So it's definitely not even remotely new, and it doesn't
>>> necessarily need to delay the current release in progress.
>>>
>>> However, it may impact a lot of existing code once the compiler can
>>> start detecting the wrong types. Several of our framework libraries
>>> in royale-asjs are affected. It's likely that user projects,
>>> especially ones with a lot of inheritance, will be affected too.
>>>
>>> In some cases, it just requires an extra cast. However, in a few
>>> cases
>> that
>>> I've seen so far, it is actually showing that there are bugs in our code.
>>> I'm working through the necessary fixes in royale-asjs now.
>>>
>>> I just wanted to give a heads up about this, as it may have some
>>> impact
>> on
>>> everyone's projects.
>>>
>>> Any thoughts on pushing this into the 0.9.12? Or would you rather
>>> wait
>> for
>>> the next one?
>>>
>>> --
>>> Josh Tynjala
>>> Bowler Hat LLC <https://bowlerhat.dev>
>>>
>>
>>
>> --
>> Andrew Wetmore
>>
>> Editor, Moose House Publications <https://moosehousepress.com/>
>> Editor-Writer, The Apache Software Foundation <https://apache.org/>
>>
RE: Super method calls not currently checked for errors
Posted by Maria Jose Esteve <mj...@iest.com>.
Hi, I'll also test my applications. It will also be a good test bench for programmers who have more difficulty programming 😝
Hiedra
-----Mensaje original-----
De: Harbs <ha...@gmail.com>
Enviado el: viernes, 11 de octubre de 2024 9:53
Para: Apache Royale Development <de...@royale.apache.org>
Asunto: Re: Super method calls not currently checked for errors
I’ll test this against my apps next week.
I’m on the fence about whether we should push it into the current release.
> On Oct 9, 2024, at 11:53 PM, Josh Tynjala <jo...@bowlerhat.dev> wrote:
>
> I too tend to be extra wary to make changes close to a release.
> Sometimes, even seemingly minor bug fixes can introduce unexpected side effects!
>
> I totally don't mind waiting to merge this stuff.
>
> --
> Josh Tynjala
> Bowler Hat LLC <https://bowlerhat.dev>
>
>
> On Wed, Oct 9, 2024 at 11:20 AM Andrew Wetmore <co...@gmail.com> wrote:
>
>> I am always scared of putting a major, necessary fix that might have
>> downstream impact into a release that is about to go out the door. I
>> would vote for making this part, perhaps the major part, of the next
>> release, and using the time between this one and that one to find and
>> fix as much as we can in our own code...and maybe prepare some
>> examples with fixes to help Royale users.
>>
>> On Wed, Oct 9, 2024 at 2:11 PM Josh Tynjala
>> <jo...@bowlerhat.dev>
>> wrote:
>>
>>> Hey all,
>>>
>>> A user of vscode-as3mxml pointed out to me recently that the
>>> arguments passed to super() and super.method() calls are not
>>> currently checked for errors by our compiler. That includes checking
>>> the types of the
>> arguments,
>>> and even whether the number of arguments is actually correct!
>>> However, regular function calls without super are correctly type
>>> checked, so it is only in cases of inheritance where super is used.
>>>
>>> I have royale-compiler fixes ready to go.
>>>
>>> This bug has been in the compiler since the code was donated by
>>> Adobe. So it's definitely not even remotely new, and it doesn't
>>> necessarily need to delay the current release in progress.
>>>
>>> However, it may impact a lot of existing code once the compiler can
>>> start detecting the wrong types. Several of our framework libraries
>>> in royale-asjs are affected. It's likely that user projects,
>>> especially ones with a lot of inheritance, will be affected too.
>>>
>>> In some cases, it just requires an extra cast. However, in a few
>>> cases
>> that
>>> I've seen so far, it is actually showing that there are bugs in our code.
>>> I'm working through the necessary fixes in royale-asjs now.
>>>
>>> I just wanted to give a heads up about this, as it may have some
>>> impact
>> on
>>> everyone's projects.
>>>
>>> Any thoughts on pushing this into the 0.9.12? Or would you rather
>>> wait
>> for
>>> the next one?
>>>
>>> --
>>> Josh Tynjala
>>> Bowler Hat LLC <https://bowlerhat.dev>
>>>
>>
>>
>> --
>> Andrew Wetmore
>>
>> Editor, Moose House Publications <https://moosehousepress.com/>
>> Editor-Writer, The Apache Software Foundation <https://apache.org/>
>>
Re: Super method calls not currently checked for errors
Posted by Harbs <ha...@gmail.com>.
I’ll test this against my apps next week.
I’m on the fence about whether we should push it into the current release.
> On Oct 9, 2024, at 11:53 PM, Josh Tynjala <jo...@bowlerhat.dev> wrote:
>
> I too tend to be extra wary to make changes close to a release. Sometimes,
> even seemingly minor bug fixes can introduce unexpected side effects!
>
> I totally don't mind waiting to merge this stuff.
>
> --
> Josh Tynjala
> Bowler Hat LLC <https://bowlerhat.dev>
>
>
> On Wed, Oct 9, 2024 at 11:20 AM Andrew Wetmore <co...@gmail.com> wrote:
>
>> I am always scared of putting a major, necessary fix that might have
>> downstream impact into a release that is about to go out the door. I would
>> vote for making this part, perhaps the major part, of the next release, and
>> using the time between this one and that one to find and fix as much as we
>> can in our own code...and maybe prepare some examples with fixes to help
>> Royale users.
>>
>> On Wed, Oct 9, 2024 at 2:11 PM Josh Tynjala <jo...@bowlerhat.dev>
>> wrote:
>>
>>> Hey all,
>>>
>>> A user of vscode-as3mxml pointed out to me recently that the arguments
>>> passed to super() and super.method() calls are not currently checked for
>>> errors by our compiler. That includes checking the types of the
>> arguments,
>>> and even whether the number of arguments is actually correct! However,
>>> regular function calls without super are correctly type checked, so it is
>>> only in cases of inheritance where super is used.
>>>
>>> I have royale-compiler fixes ready to go.
>>>
>>> This bug has been in the compiler since the code was donated by Adobe. So
>>> it's definitely not even remotely new, and it doesn't necessarily need to
>>> delay the current release in progress.
>>>
>>> However, it may impact a lot of existing code once the compiler can start
>>> detecting the wrong types. Several of our framework libraries in
>>> royale-asjs are affected. It's likely that user projects, especially ones
>>> with a lot of inheritance, will be affected too.
>>>
>>> In some cases, it just requires an extra cast. However, in a few cases
>> that
>>> I've seen so far, it is actually showing that there are bugs in our code.
>>> I'm working through the necessary fixes in royale-asjs now.
>>>
>>> I just wanted to give a heads up about this, as it may have some impact
>> on
>>> everyone's projects.
>>>
>>> Any thoughts on pushing this into the 0.9.12? Or would you rather wait
>> for
>>> the next one?
>>>
>>> --
>>> Josh Tynjala
>>> Bowler Hat LLC <https://bowlerhat.dev>
>>>
>>
>>
>> --
>> Andrew Wetmore
>>
>> Editor, Moose House Publications <https://moosehousepress.com/>
>> Editor-Writer, The Apache Software Foundation <https://apache.org/>
>>
Re: Super method calls not currently checked for errors
Posted by Harbs <ha...@gmail.com>.
If you’re only compiling for JS, using an “as” cast with ignoring coercions should work, although you’re actually lying about the type...
> On Oct 10, 2024, at 11:02 PM, Piotr Zarzycki <pi...@gmail.com> wrote:
>
> Sounds like casting to that fail. After running application I'm getting:
>
> Uncaught TypeError: Error #1034: Type Coercion failed: cannot convert
> view.support.NewTicket to interfaces.ticket.IEmergencySupport
> as
> https://localdev.prominic.net/prominic/bin/js-debug/org/apache/royale/utils/Language.js
> :113
> MediatorNewTicket
> https://localdev.prominic.net/prominic/bin/js-debug/mediator/MediatorNewTicket.js
> :23
>
> I will think how to manage that.
>
> czw., 10 paź 2024 o 21:54 Piotr Zarzycki <pi...@gmail.com>
> napisał(a):
>
>> Yep! This is what I did :) Thank you. Let me test next apps in the next
>> few days and I will get back to you.
>>
>> czw., 10 paź 2024 o 21:53 Josh Tynjala <jo...@bowlerhat.dev>
>> napisał(a):
>>
>>> Yes, that is most likely the result of my compiler fix. You should simply
>>> need to add a cast, probably like this:
>>>
>>> super(IEmergencySupport(component));
>>>
>>> --
>>> Josh Tynjala
>>> Bowler Hat LLC <https://bowlerhat.dev>
>>>
>>>
>>> On Thu, Oct 10, 2024 at 12:44 PM Piotr Zarzycki <
>>> piotrzarzycki21@gmail.com>
>>> wrote:
>>>
>>>> Got it! I just rebuild framework and build our app with your code. I'm
>>>> having only one build error which wasn't present previously. I'm
>>> wondering
>>>> if this is because of your changes:
>>>>
>>>> MediatorNewTicket.as(14): col: 10 Error: Implicit coercion of a value
>>> with
>>>> static type interfaces.ticket.INormalSupport to a possibly unrelated
>>> type
>>>> interfaces.ticket.IEmergencySupport.
>>>> : super(component);
>>>>
>>>> MediatorNewTicket extends MediatorEmergencySupport
>>>>
>>>> IEmergencySupport extends INormalSupport
>>>>
>>>> MediatorNewTicket(component:INormalSupport) - constructor
>>>> MediatorEmergencySupport(component:IEmergencySupport) - constructor
>>>>
>>>> I think I know how to fix this. Is this results of your changes ?
>>>>
>>>> Thanks,
>>>> Piotr
>>>>
>>>>
>>>> czw., 10 paź 2024 o 17:29 Josh Tynjala <jo...@bowlerhat.dev>
>>>> napisał(a):
>>>>
>>>>> Oh, and make sure that the pom.xml files in your app projects are
>>> using
>>>>> version 0.9.12 for the compiler too.
>>>>>
>>>>> --
>>>>> Josh Tynjala
>>>>> Bowler Hat LLC <https://bowlerhat.dev>
>>>>>
>>>>>
>>>>> On Thu, Oct 10, 2024 at 8:15 AM Josh Tynjala <
>>> joshtynjala@bowlerhat.dev>
>>>>> wrote:
>>>>>
>>>>>> Hey Piotr,
>>>>>>
>>>>>> I just pushed one more commit to the fix/super-signature-checks
>>> branch
>>>> of
>>>>>> royale-asjs.
>>>>>>
>>>>>> I noticed that the Maven build was not using my changes in
>>>>> royale-compiler
>>>>>> because we're midway through the process of updating the repos from
>>>>> 0.9.11
>>>>>> to 0.9.12. So you might not have actually tested my changes if you
>>> were
>>>>>> building with Maven, unless you had manually made this tweak to
>>> pom.xml
>>>>> too.
>>>>>>
>>>>>> --
>>>>>> Josh Tynjala
>>>>>> Bowler Hat LLC <https://bowlerhat.dev>
>>>>>>
>>>>>>
>>>>>> On Thu, Oct 10, 2024 at 2:18 AM Piotr Zarzycki <
>>>>> piotrzarzycki21@gmail.com>
>>>>>> wrote:
>>>>>>
>>>>>>> Hi Josh,
>>>>>>>
>>>>>>> I have used your branches to build locally Royale. I have checked
>>> our
>>>>> main
>>>>>>> application Jewel based with your changes and everything seems to
>>> be
>>>>> fine.
>>>>>>> Application is working like a charm debug and release version. No
>>>> errors
>>>>>>> in
>>>>>>> DOM elements on the console etc.
>>>>>>>
>>>>>>> I will check soon another app to confirm, but currently it looks
>>>> really
>>>>>>> good.
>>>>>>>
>>>>>>> Thanks,
>>>>>>> Piotr
>>>>>>>
>>>>>>> śr., 9 paź 2024 o 22:54 Josh Tynjala <jo...@bowlerhat.dev>
>>>>>>> napisał(a):
>>>>>>>
>>>>>>>> I too tend to be extra wary to make changes close to a release.
>>>>>>> Sometimes,
>>>>>>>> even seemingly minor bug fixes can introduce unexpected side
>>>> effects!
>>>>>>>>
>>>>>>>> I totally don't mind waiting to merge this stuff.
>>>>>>>>
>>>>>>>> --
>>>>>>>> Josh Tynjala
>>>>>>>> Bowler Hat LLC <https://bowlerhat.dev>
>>>>>>>>
>>>>>>>>
>>>>>>>> On Wed, Oct 9, 2024 at 11:20 AM Andrew Wetmore <
>>> cottage14@gmail.com
>>>>>
>>>>>>>> wrote:
>>>>>>>>
>>>>>>>>> I am always scared of putting a major, necessary fix that might
>>>> have
>>>>>>>>> downstream impact into a release that is about to go out the
>>>> door. I
>>>>>>>> would
>>>>>>>>> vote for making this part, perhaps the major part, of the next
>>>>>>> release,
>>>>>>>> and
>>>>>>>>> using the time between this one and that one to find and fix as
>>>> much
>>>>>>> as
>>>>>>>> we
>>>>>>>>> can in our own code...and maybe prepare some examples with
>>> fixes
>>>> to
>>>>>>> help
>>>>>>>>> Royale users.
>>>>>>>>>
>>>>>>>>> On Wed, Oct 9, 2024 at 2:11 PM Josh Tynjala <
>>>>>>> joshtynjala@bowlerhat.dev>
>>>>>>>>> wrote:
>>>>>>>>>
>>>>>>>>>> Hey all,
>>>>>>>>>>
>>>>>>>>>> A user of vscode-as3mxml pointed out to me recently that the
>>>>>>> arguments
>>>>>>>>>> passed to super() and super.method() calls are not currently
>>>>> checked
>>>>>>>> for
>>>>>>>>>> errors by our compiler. That includes checking the types of
>>> the
>>>>>>>>> arguments,
>>>>>>>>>> and even whether the number of arguments is actually correct!
>>>>>>> However,
>>>>>>>>>> regular function calls without super are correctly type
>>> checked,
>>>>> so
>>>>>>> it
>>>>>>>> is
>>>>>>>>>> only in cases of inheritance where super is used.
>>>>>>>>>>
>>>>>>>>>> I have royale-compiler fixes ready to go.
>>>>>>>>>>
>>>>>>>>>> This bug has been in the compiler since the code was donated
>>> by
>>>>>>> Adobe.
>>>>>>>> So
>>>>>>>>>> it's definitely not even remotely new, and it doesn't
>>>> necessarily
>>>>>>> need
>>>>>>>> to
>>>>>>>>>> delay the current release in progress.
>>>>>>>>>>
>>>>>>>>>> However, it may impact a lot of existing code once the
>>> compiler
>>>>> can
>>>>>>>> start
>>>>>>>>>> detecting the wrong types. Several of our framework
>>> libraries in
>>>>>>>>>> royale-asjs are affected. It's likely that user projects,
>>>>> especially
>>>>>>>> ones
>>>>>>>>>> with a lot of inheritance, will be affected too.
>>>>>>>>>>
>>>>>>>>>> In some cases, it just requires an extra cast. However, in a
>>> few
>>>>>>> cases
>>>>>>>>> that
>>>>>>>>>> I've seen so far, it is actually showing that there are bugs
>>> in
>>>>> our
>>>>>>>> code.
>>>>>>>>>> I'm working through the necessary fixes in royale-asjs now.
>>>>>>>>>>
>>>>>>>>>> I just wanted to give a heads up about this, as it may have
>>> some
>>>>>>> impact
>>>>>>>>> on
>>>>>>>>>> everyone's projects.
>>>>>>>>>>
>>>>>>>>>> Any thoughts on pushing this into the 0.9.12? Or would you
>>>> rather
>>>>>>> wait
>>>>>>>>> for
>>>>>>>>>> the next one?
>>>>>>>>>>
>>>>>>>>>> --
>>>>>>>>>> Josh Tynjala
>>>>>>>>>> Bowler Hat LLC <https://bowlerhat.dev>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> --
>>>>>>>>> Andrew Wetmore
>>>>>>>>>
>>>>>>>>> Editor, Moose House Publications <https://moosehousepress.com/
>>>>
>>>>>>>>> Editor-Writer, The Apache Software Foundation <
>>>> https://apache.org/>
>>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>>
>>>>>>> Piotr Zarzycki
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>>
>>>> --
>>>>
>>>> Piotr Zarzycki
>>>>
>>>
>>
>>
>> --
>>
>> Piotr Zarzycki
>>
>
>
> --
>
> Piotr Zarzycki
Re: Super method calls not currently checked for errors
Posted by Piotr Zarzycki <pi...@gmail.com>.
Sounds like casting to that fail. After running application I'm getting:
Uncaught TypeError: Error #1034: Type Coercion failed: cannot convert
view.support.NewTicket to interfaces.ticket.IEmergencySupport
as
https://localdev.prominic.net/prominic/bin/js-debug/org/apache/royale/utils/Language.js
:113
MediatorNewTicket
https://localdev.prominic.net/prominic/bin/js-debug/mediator/MediatorNewTicket.js
:23
I will think how to manage that.
czw., 10 paź 2024 o 21:54 Piotr Zarzycki <pi...@gmail.com>
napisał(a):
> Yep! This is what I did :) Thank you. Let me test next apps in the next
> few days and I will get back to you.
>
> czw., 10 paź 2024 o 21:53 Josh Tynjala <jo...@bowlerhat.dev>
> napisał(a):
>
>> Yes, that is most likely the result of my compiler fix. You should simply
>> need to add a cast, probably like this:
>>
>> super(IEmergencySupport(component));
>>
>> --
>> Josh Tynjala
>> Bowler Hat LLC <https://bowlerhat.dev>
>>
>>
>> On Thu, Oct 10, 2024 at 12:44 PM Piotr Zarzycki <
>> piotrzarzycki21@gmail.com>
>> wrote:
>>
>> > Got it! I just rebuild framework and build our app with your code. I'm
>> > having only one build error which wasn't present previously. I'm
>> wondering
>> > if this is because of your changes:
>> >
>> > MediatorNewTicket.as(14): col: 10 Error: Implicit coercion of a value
>> with
>> > static type interfaces.ticket.INormalSupport to a possibly unrelated
>> type
>> > interfaces.ticket.IEmergencySupport.
>> > : super(component);
>> >
>> > MediatorNewTicket extends MediatorEmergencySupport
>> >
>> > IEmergencySupport extends INormalSupport
>> >
>> > MediatorNewTicket(component:INormalSupport) - constructor
>> > MediatorEmergencySupport(component:IEmergencySupport) - constructor
>> >
>> > I think I know how to fix this. Is this results of your changes ?
>> >
>> > Thanks,
>> > Piotr
>> >
>> >
>> > czw., 10 paź 2024 o 17:29 Josh Tynjala <jo...@bowlerhat.dev>
>> > napisał(a):
>> >
>> > > Oh, and make sure that the pom.xml files in your app projects are
>> using
>> > > version 0.9.12 for the compiler too.
>> > >
>> > > --
>> > > Josh Tynjala
>> > > Bowler Hat LLC <https://bowlerhat.dev>
>> > >
>> > >
>> > > On Thu, Oct 10, 2024 at 8:15 AM Josh Tynjala <
>> joshtynjala@bowlerhat.dev>
>> > > wrote:
>> > >
>> > > > Hey Piotr,
>> > > >
>> > > > I just pushed one more commit to the fix/super-signature-checks
>> branch
>> > of
>> > > > royale-asjs.
>> > > >
>> > > > I noticed that the Maven build was not using my changes in
>> > > royale-compiler
>> > > > because we're midway through the process of updating the repos from
>> > > 0.9.11
>> > > > to 0.9.12. So you might not have actually tested my changes if you
>> were
>> > > > building with Maven, unless you had manually made this tweak to
>> pom.xml
>> > > too.
>> > > >
>> > > > --
>> > > > Josh Tynjala
>> > > > Bowler Hat LLC <https://bowlerhat.dev>
>> > > >
>> > > >
>> > > > On Thu, Oct 10, 2024 at 2:18 AM Piotr Zarzycki <
>> > > piotrzarzycki21@gmail.com>
>> > > > wrote:
>> > > >
>> > > >> Hi Josh,
>> > > >>
>> > > >> I have used your branches to build locally Royale. I have checked
>> our
>> > > main
>> > > >> application Jewel based with your changes and everything seems to
>> be
>> > > fine.
>> > > >> Application is working like a charm debug and release version. No
>> > errors
>> > > >> in
>> > > >> DOM elements on the console etc.
>> > > >>
>> > > >> I will check soon another app to confirm, but currently it looks
>> > really
>> > > >> good.
>> > > >>
>> > > >> Thanks,
>> > > >> Piotr
>> > > >>
>> > > >> śr., 9 paź 2024 o 22:54 Josh Tynjala <jo...@bowlerhat.dev>
>> > > >> napisał(a):
>> > > >>
>> > > >> > I too tend to be extra wary to make changes close to a release.
>> > > >> Sometimes,
>> > > >> > even seemingly minor bug fixes can introduce unexpected side
>> > effects!
>> > > >> >
>> > > >> > I totally don't mind waiting to merge this stuff.
>> > > >> >
>> > > >> > --
>> > > >> > Josh Tynjala
>> > > >> > Bowler Hat LLC <https://bowlerhat.dev>
>> > > >> >
>> > > >> >
>> > > >> > On Wed, Oct 9, 2024 at 11:20 AM Andrew Wetmore <
>> cottage14@gmail.com
>> > >
>> > > >> > wrote:
>> > > >> >
>> > > >> > > I am always scared of putting a major, necessary fix that might
>> > have
>> > > >> > > downstream impact into a release that is about to go out the
>> > door. I
>> > > >> > would
>> > > >> > > vote for making this part, perhaps the major part, of the next
>> > > >> release,
>> > > >> > and
>> > > >> > > using the time between this one and that one to find and fix as
>> > much
>> > > >> as
>> > > >> > we
>> > > >> > > can in our own code...and maybe prepare some examples with
>> fixes
>> > to
>> > > >> help
>> > > >> > > Royale users.
>> > > >> > >
>> > > >> > > On Wed, Oct 9, 2024 at 2:11 PM Josh Tynjala <
>> > > >> joshtynjala@bowlerhat.dev>
>> > > >> > > wrote:
>> > > >> > >
>> > > >> > > > Hey all,
>> > > >> > > >
>> > > >> > > > A user of vscode-as3mxml pointed out to me recently that the
>> > > >> arguments
>> > > >> > > > passed to super() and super.method() calls are not currently
>> > > checked
>> > > >> > for
>> > > >> > > > errors by our compiler. That includes checking the types of
>> the
>> > > >> > > arguments,
>> > > >> > > > and even whether the number of arguments is actually correct!
>> > > >> However,
>> > > >> > > > regular function calls without super are correctly type
>> checked,
>> > > so
>> > > >> it
>> > > >> > is
>> > > >> > > > only in cases of inheritance where super is used.
>> > > >> > > >
>> > > >> > > > I have royale-compiler fixes ready to go.
>> > > >> > > >
>> > > >> > > > This bug has been in the compiler since the code was donated
>> by
>> > > >> Adobe.
>> > > >> > So
>> > > >> > > > it's definitely not even remotely new, and it doesn't
>> > necessarily
>> > > >> need
>> > > >> > to
>> > > >> > > > delay the current release in progress.
>> > > >> > > >
>> > > >> > > > However, it may impact a lot of existing code once the
>> compiler
>> > > can
>> > > >> > start
>> > > >> > > > detecting the wrong types. Several of our framework
>> libraries in
>> > > >> > > > royale-asjs are affected. It's likely that user projects,
>> > > especially
>> > > >> > ones
>> > > >> > > > with a lot of inheritance, will be affected too.
>> > > >> > > >
>> > > >> > > > In some cases, it just requires an extra cast. However, in a
>> few
>> > > >> cases
>> > > >> > > that
>> > > >> > > > I've seen so far, it is actually showing that there are bugs
>> in
>> > > our
>> > > >> > code.
>> > > >> > > > I'm working through the necessary fixes in royale-asjs now.
>> > > >> > > >
>> > > >> > > > I just wanted to give a heads up about this, as it may have
>> some
>> > > >> impact
>> > > >> > > on
>> > > >> > > > everyone's projects.
>> > > >> > > >
>> > > >> > > > Any thoughts on pushing this into the 0.9.12? Or would you
>> > rather
>> > > >> wait
>> > > >> > > for
>> > > >> > > > the next one?
>> > > >> > > >
>> > > >> > > > --
>> > > >> > > > Josh Tynjala
>> > > >> > > > Bowler Hat LLC <https://bowlerhat.dev>
>> > > >> > > >
>> > > >> > >
>> > > >> > >
>> > > >> > > --
>> > > >> > > Andrew Wetmore
>> > > >> > >
>> > > >> > > Editor, Moose House Publications <https://moosehousepress.com/
>> >
>> > > >> > > Editor-Writer, The Apache Software Foundation <
>> > https://apache.org/>
>> > > >> > >
>> > > >> >
>> > > >>
>> > > >>
>> > > >> --
>> > > >>
>> > > >> Piotr Zarzycki
>> > > >>
>> > > >
>> > >
>> >
>> >
>> > --
>> >
>> > Piotr Zarzycki
>> >
>>
>
>
> --
>
> Piotr Zarzycki
>
--
Piotr Zarzycki
Re: Super method calls not currently checked for errors
Posted by Piotr Zarzycki <pi...@gmail.com>.
Yep! This is what I did :) Thank you. Let me test next apps in the next
few days and I will get back to you.
czw., 10 paź 2024 o 21:53 Josh Tynjala <jo...@bowlerhat.dev>
napisał(a):
> Yes, that is most likely the result of my compiler fix. You should simply
> need to add a cast, probably like this:
>
> super(IEmergencySupport(component));
>
> --
> Josh Tynjala
> Bowler Hat LLC <https://bowlerhat.dev>
>
>
> On Thu, Oct 10, 2024 at 12:44 PM Piotr Zarzycki <piotrzarzycki21@gmail.com
> >
> wrote:
>
> > Got it! I just rebuild framework and build our app with your code. I'm
> > having only one build error which wasn't present previously. I'm
> wondering
> > if this is because of your changes:
> >
> > MediatorNewTicket.as(14): col: 10 Error: Implicit coercion of a value
> with
> > static type interfaces.ticket.INormalSupport to a possibly unrelated type
> > interfaces.ticket.IEmergencySupport.
> > : super(component);
> >
> > MediatorNewTicket extends MediatorEmergencySupport
> >
> > IEmergencySupport extends INormalSupport
> >
> > MediatorNewTicket(component:INormalSupport) - constructor
> > MediatorEmergencySupport(component:IEmergencySupport) - constructor
> >
> > I think I know how to fix this. Is this results of your changes ?
> >
> > Thanks,
> > Piotr
> >
> >
> > czw., 10 paź 2024 o 17:29 Josh Tynjala <jo...@bowlerhat.dev>
> > napisał(a):
> >
> > > Oh, and make sure that the pom.xml files in your app projects are using
> > > version 0.9.12 for the compiler too.
> > >
> > > --
> > > Josh Tynjala
> > > Bowler Hat LLC <https://bowlerhat.dev>
> > >
> > >
> > > On Thu, Oct 10, 2024 at 8:15 AM Josh Tynjala <
> joshtynjala@bowlerhat.dev>
> > > wrote:
> > >
> > > > Hey Piotr,
> > > >
> > > > I just pushed one more commit to the fix/super-signature-checks
> branch
> > of
> > > > royale-asjs.
> > > >
> > > > I noticed that the Maven build was not using my changes in
> > > royale-compiler
> > > > because we're midway through the process of updating the repos from
> > > 0.9.11
> > > > to 0.9.12. So you might not have actually tested my changes if you
> were
> > > > building with Maven, unless you had manually made this tweak to
> pom.xml
> > > too.
> > > >
> > > > --
> > > > Josh Tynjala
> > > > Bowler Hat LLC <https://bowlerhat.dev>
> > > >
> > > >
> > > > On Thu, Oct 10, 2024 at 2:18 AM Piotr Zarzycki <
> > > piotrzarzycki21@gmail.com>
> > > > wrote:
> > > >
> > > >> Hi Josh,
> > > >>
> > > >> I have used your branches to build locally Royale. I have checked
> our
> > > main
> > > >> application Jewel based with your changes and everything seems to be
> > > fine.
> > > >> Application is working like a charm debug and release version. No
> > errors
> > > >> in
> > > >> DOM elements on the console etc.
> > > >>
> > > >> I will check soon another app to confirm, but currently it looks
> > really
> > > >> good.
> > > >>
> > > >> Thanks,
> > > >> Piotr
> > > >>
> > > >> śr., 9 paź 2024 o 22:54 Josh Tynjala <jo...@bowlerhat.dev>
> > > >> napisał(a):
> > > >>
> > > >> > I too tend to be extra wary to make changes close to a release.
> > > >> Sometimes,
> > > >> > even seemingly minor bug fixes can introduce unexpected side
> > effects!
> > > >> >
> > > >> > I totally don't mind waiting to merge this stuff.
> > > >> >
> > > >> > --
> > > >> > Josh Tynjala
> > > >> > Bowler Hat LLC <https://bowlerhat.dev>
> > > >> >
> > > >> >
> > > >> > On Wed, Oct 9, 2024 at 11:20 AM Andrew Wetmore <
> cottage14@gmail.com
> > >
> > > >> > wrote:
> > > >> >
> > > >> > > I am always scared of putting a major, necessary fix that might
> > have
> > > >> > > downstream impact into a release that is about to go out the
> > door. I
> > > >> > would
> > > >> > > vote for making this part, perhaps the major part, of the next
> > > >> release,
> > > >> > and
> > > >> > > using the time between this one and that one to find and fix as
> > much
> > > >> as
> > > >> > we
> > > >> > > can in our own code...and maybe prepare some examples with fixes
> > to
> > > >> help
> > > >> > > Royale users.
> > > >> > >
> > > >> > > On Wed, Oct 9, 2024 at 2:11 PM Josh Tynjala <
> > > >> joshtynjala@bowlerhat.dev>
> > > >> > > wrote:
> > > >> > >
> > > >> > > > Hey all,
> > > >> > > >
> > > >> > > > A user of vscode-as3mxml pointed out to me recently that the
> > > >> arguments
> > > >> > > > passed to super() and super.method() calls are not currently
> > > checked
> > > >> > for
> > > >> > > > errors by our compiler. That includes checking the types of
> the
> > > >> > > arguments,
> > > >> > > > and even whether the number of arguments is actually correct!
> > > >> However,
> > > >> > > > regular function calls without super are correctly type
> checked,
> > > so
> > > >> it
> > > >> > is
> > > >> > > > only in cases of inheritance where super is used.
> > > >> > > >
> > > >> > > > I have royale-compiler fixes ready to go.
> > > >> > > >
> > > >> > > > This bug has been in the compiler since the code was donated
> by
> > > >> Adobe.
> > > >> > So
> > > >> > > > it's definitely not even remotely new, and it doesn't
> > necessarily
> > > >> need
> > > >> > to
> > > >> > > > delay the current release in progress.
> > > >> > > >
> > > >> > > > However, it may impact a lot of existing code once the
> compiler
> > > can
> > > >> > start
> > > >> > > > detecting the wrong types. Several of our framework libraries
> in
> > > >> > > > royale-asjs are affected. It's likely that user projects,
> > > especially
> > > >> > ones
> > > >> > > > with a lot of inheritance, will be affected too.
> > > >> > > >
> > > >> > > > In some cases, it just requires an extra cast. However, in a
> few
> > > >> cases
> > > >> > > that
> > > >> > > > I've seen so far, it is actually showing that there are bugs
> in
> > > our
> > > >> > code.
> > > >> > > > I'm working through the necessary fixes in royale-asjs now.
> > > >> > > >
> > > >> > > > I just wanted to give a heads up about this, as it may have
> some
> > > >> impact
> > > >> > > on
> > > >> > > > everyone's projects.
> > > >> > > >
> > > >> > > > Any thoughts on pushing this into the 0.9.12? Or would you
> > rather
> > > >> wait
> > > >> > > for
> > > >> > > > the next one?
> > > >> > > >
> > > >> > > > --
> > > >> > > > Josh Tynjala
> > > >> > > > Bowler Hat LLC <https://bowlerhat.dev>
> > > >> > > >
> > > >> > >
> > > >> > >
> > > >> > > --
> > > >> > > Andrew Wetmore
> > > >> > >
> > > >> > > Editor, Moose House Publications <https://moosehousepress.com/>
> > > >> > > Editor-Writer, The Apache Software Foundation <
> > https://apache.org/>
> > > >> > >
> > > >> >
> > > >>
> > > >>
> > > >> --
> > > >>
> > > >> Piotr Zarzycki
> > > >>
> > > >
> > >
> >
> >
> > --
> >
> > Piotr Zarzycki
> >
>
--
Piotr Zarzycki
Re: Super method calls not currently checked for errors
Posted by Josh Tynjala <jo...@bowlerhat.dev>.
Yes, that is most likely the result of my compiler fix. You should simply
need to add a cast, probably like this:
super(IEmergencySupport(component));
--
Josh Tynjala
Bowler Hat LLC <https://bowlerhat.dev>
On Thu, Oct 10, 2024 at 12:44 PM Piotr Zarzycki <pi...@gmail.com>
wrote:
> Got it! I just rebuild framework and build our app with your code. I'm
> having only one build error which wasn't present previously. I'm wondering
> if this is because of your changes:
>
> MediatorNewTicket.as(14): col: 10 Error: Implicit coercion of a value with
> static type interfaces.ticket.INormalSupport to a possibly unrelated type
> interfaces.ticket.IEmergencySupport.
> : super(component);
>
> MediatorNewTicket extends MediatorEmergencySupport
>
> IEmergencySupport extends INormalSupport
>
> MediatorNewTicket(component:INormalSupport) - constructor
> MediatorEmergencySupport(component:IEmergencySupport) - constructor
>
> I think I know how to fix this. Is this results of your changes ?
>
> Thanks,
> Piotr
>
>
> czw., 10 paź 2024 o 17:29 Josh Tynjala <jo...@bowlerhat.dev>
> napisał(a):
>
> > Oh, and make sure that the pom.xml files in your app projects are using
> > version 0.9.12 for the compiler too.
> >
> > --
> > Josh Tynjala
> > Bowler Hat LLC <https://bowlerhat.dev>
> >
> >
> > On Thu, Oct 10, 2024 at 8:15 AM Josh Tynjala <jo...@bowlerhat.dev>
> > wrote:
> >
> > > Hey Piotr,
> > >
> > > I just pushed one more commit to the fix/super-signature-checks branch
> of
> > > royale-asjs.
> > >
> > > I noticed that the Maven build was not using my changes in
> > royale-compiler
> > > because we're midway through the process of updating the repos from
> > 0.9.11
> > > to 0.9.12. So you might not have actually tested my changes if you were
> > > building with Maven, unless you had manually made this tweak to pom.xml
> > too.
> > >
> > > --
> > > Josh Tynjala
> > > Bowler Hat LLC <https://bowlerhat.dev>
> > >
> > >
> > > On Thu, Oct 10, 2024 at 2:18 AM Piotr Zarzycki <
> > piotrzarzycki21@gmail.com>
> > > wrote:
> > >
> > >> Hi Josh,
> > >>
> > >> I have used your branches to build locally Royale. I have checked our
> > main
> > >> application Jewel based with your changes and everything seems to be
> > fine.
> > >> Application is working like a charm debug and release version. No
> errors
> > >> in
> > >> DOM elements on the console etc.
> > >>
> > >> I will check soon another app to confirm, but currently it looks
> really
> > >> good.
> > >>
> > >> Thanks,
> > >> Piotr
> > >>
> > >> śr., 9 paź 2024 o 22:54 Josh Tynjala <jo...@bowlerhat.dev>
> > >> napisał(a):
> > >>
> > >> > I too tend to be extra wary to make changes close to a release.
> > >> Sometimes,
> > >> > even seemingly minor bug fixes can introduce unexpected side
> effects!
> > >> >
> > >> > I totally don't mind waiting to merge this stuff.
> > >> >
> > >> > --
> > >> > Josh Tynjala
> > >> > Bowler Hat LLC <https://bowlerhat.dev>
> > >> >
> > >> >
> > >> > On Wed, Oct 9, 2024 at 11:20 AM Andrew Wetmore <cottage14@gmail.com
> >
> > >> > wrote:
> > >> >
> > >> > > I am always scared of putting a major, necessary fix that might
> have
> > >> > > downstream impact into a release that is about to go out the
> door. I
> > >> > would
> > >> > > vote for making this part, perhaps the major part, of the next
> > >> release,
> > >> > and
> > >> > > using the time between this one and that one to find and fix as
> much
> > >> as
> > >> > we
> > >> > > can in our own code...and maybe prepare some examples with fixes
> to
> > >> help
> > >> > > Royale users.
> > >> > >
> > >> > > On Wed, Oct 9, 2024 at 2:11 PM Josh Tynjala <
> > >> joshtynjala@bowlerhat.dev>
> > >> > > wrote:
> > >> > >
> > >> > > > Hey all,
> > >> > > >
> > >> > > > A user of vscode-as3mxml pointed out to me recently that the
> > >> arguments
> > >> > > > passed to super() and super.method() calls are not currently
> > checked
> > >> > for
> > >> > > > errors by our compiler. That includes checking the types of the
> > >> > > arguments,
> > >> > > > and even whether the number of arguments is actually correct!
> > >> However,
> > >> > > > regular function calls without super are correctly type checked,
> > so
> > >> it
> > >> > is
> > >> > > > only in cases of inheritance where super is used.
> > >> > > >
> > >> > > > I have royale-compiler fixes ready to go.
> > >> > > >
> > >> > > > This bug has been in the compiler since the code was donated by
> > >> Adobe.
> > >> > So
> > >> > > > it's definitely not even remotely new, and it doesn't
> necessarily
> > >> need
> > >> > to
> > >> > > > delay the current release in progress.
> > >> > > >
> > >> > > > However, it may impact a lot of existing code once the compiler
> > can
> > >> > start
> > >> > > > detecting the wrong types. Several of our framework libraries in
> > >> > > > royale-asjs are affected. It's likely that user projects,
> > especially
> > >> > ones
> > >> > > > with a lot of inheritance, will be affected too.
> > >> > > >
> > >> > > > In some cases, it just requires an extra cast. However, in a few
> > >> cases
> > >> > > that
> > >> > > > I've seen so far, it is actually showing that there are bugs in
> > our
> > >> > code.
> > >> > > > I'm working through the necessary fixes in royale-asjs now.
> > >> > > >
> > >> > > > I just wanted to give a heads up about this, as it may have some
> > >> impact
> > >> > > on
> > >> > > > everyone's projects.
> > >> > > >
> > >> > > > Any thoughts on pushing this into the 0.9.12? Or would you
> rather
> > >> wait
> > >> > > for
> > >> > > > the next one?
> > >> > > >
> > >> > > > --
> > >> > > > Josh Tynjala
> > >> > > > Bowler Hat LLC <https://bowlerhat.dev>
> > >> > > >
> > >> > >
> > >> > >
> > >> > > --
> > >> > > Andrew Wetmore
> > >> > >
> > >> > > Editor, Moose House Publications <https://moosehousepress.com/>
> > >> > > Editor-Writer, The Apache Software Foundation <
> https://apache.org/>
> > >> > >
> > >> >
> > >>
> > >>
> > >> --
> > >>
> > >> Piotr Zarzycki
> > >>
> > >
> >
>
>
> --
>
> Piotr Zarzycki
>
Re: Super method calls not currently checked for errors
Posted by Piotr Zarzycki <pi...@gmail.com>.
Got it! I just rebuild framework and build our app with your code. I'm
having only one build error which wasn't present previously. I'm wondering
if this is because of your changes:
MediatorNewTicket.as(14): col: 10 Error: Implicit coercion of a value with
static type interfaces.ticket.INormalSupport to a possibly unrelated type
interfaces.ticket.IEmergencySupport.
: super(component);
MediatorNewTicket extends MediatorEmergencySupport
IEmergencySupport extends INormalSupport
MediatorNewTicket(component:INormalSupport) - constructor
MediatorEmergencySupport(component:IEmergencySupport) - constructor
I think I know how to fix this. Is this results of your changes ?
Thanks,
Piotr
czw., 10 paź 2024 o 17:29 Josh Tynjala <jo...@bowlerhat.dev>
napisał(a):
> Oh, and make sure that the pom.xml files in your app projects are using
> version 0.9.12 for the compiler too.
>
> --
> Josh Tynjala
> Bowler Hat LLC <https://bowlerhat.dev>
>
>
> On Thu, Oct 10, 2024 at 8:15 AM Josh Tynjala <jo...@bowlerhat.dev>
> wrote:
>
> > Hey Piotr,
> >
> > I just pushed one more commit to the fix/super-signature-checks branch of
> > royale-asjs.
> >
> > I noticed that the Maven build was not using my changes in
> royale-compiler
> > because we're midway through the process of updating the repos from
> 0.9.11
> > to 0.9.12. So you might not have actually tested my changes if you were
> > building with Maven, unless you had manually made this tweak to pom.xml
> too.
> >
> > --
> > Josh Tynjala
> > Bowler Hat LLC <https://bowlerhat.dev>
> >
> >
> > On Thu, Oct 10, 2024 at 2:18 AM Piotr Zarzycki <
> piotrzarzycki21@gmail.com>
> > wrote:
> >
> >> Hi Josh,
> >>
> >> I have used your branches to build locally Royale. I have checked our
> main
> >> application Jewel based with your changes and everything seems to be
> fine.
> >> Application is working like a charm debug and release version. No errors
> >> in
> >> DOM elements on the console etc.
> >>
> >> I will check soon another app to confirm, but currently it looks really
> >> good.
> >>
> >> Thanks,
> >> Piotr
> >>
> >> śr., 9 paź 2024 o 22:54 Josh Tynjala <jo...@bowlerhat.dev>
> >> napisał(a):
> >>
> >> > I too tend to be extra wary to make changes close to a release.
> >> Sometimes,
> >> > even seemingly minor bug fixes can introduce unexpected side effects!
> >> >
> >> > I totally don't mind waiting to merge this stuff.
> >> >
> >> > --
> >> > Josh Tynjala
> >> > Bowler Hat LLC <https://bowlerhat.dev>
> >> >
> >> >
> >> > On Wed, Oct 9, 2024 at 11:20 AM Andrew Wetmore <co...@gmail.com>
> >> > wrote:
> >> >
> >> > > I am always scared of putting a major, necessary fix that might have
> >> > > downstream impact into a release that is about to go out the door. I
> >> > would
> >> > > vote for making this part, perhaps the major part, of the next
> >> release,
> >> > and
> >> > > using the time between this one and that one to find and fix as much
> >> as
> >> > we
> >> > > can in our own code...and maybe prepare some examples with fixes to
> >> help
> >> > > Royale users.
> >> > >
> >> > > On Wed, Oct 9, 2024 at 2:11 PM Josh Tynjala <
> >> joshtynjala@bowlerhat.dev>
> >> > > wrote:
> >> > >
> >> > > > Hey all,
> >> > > >
> >> > > > A user of vscode-as3mxml pointed out to me recently that the
> >> arguments
> >> > > > passed to super() and super.method() calls are not currently
> checked
> >> > for
> >> > > > errors by our compiler. That includes checking the types of the
> >> > > arguments,
> >> > > > and even whether the number of arguments is actually correct!
> >> However,
> >> > > > regular function calls without super are correctly type checked,
> so
> >> it
> >> > is
> >> > > > only in cases of inheritance where super is used.
> >> > > >
> >> > > > I have royale-compiler fixes ready to go.
> >> > > >
> >> > > > This bug has been in the compiler since the code was donated by
> >> Adobe.
> >> > So
> >> > > > it's definitely not even remotely new, and it doesn't necessarily
> >> need
> >> > to
> >> > > > delay the current release in progress.
> >> > > >
> >> > > > However, it may impact a lot of existing code once the compiler
> can
> >> > start
> >> > > > detecting the wrong types. Several of our framework libraries in
> >> > > > royale-asjs are affected. It's likely that user projects,
> especially
> >> > ones
> >> > > > with a lot of inheritance, will be affected too.
> >> > > >
> >> > > > In some cases, it just requires an extra cast. However, in a few
> >> cases
> >> > > that
> >> > > > I've seen so far, it is actually showing that there are bugs in
> our
> >> > code.
> >> > > > I'm working through the necessary fixes in royale-asjs now.
> >> > > >
> >> > > > I just wanted to give a heads up about this, as it may have some
> >> impact
> >> > > on
> >> > > > everyone's projects.
> >> > > >
> >> > > > Any thoughts on pushing this into the 0.9.12? Or would you rather
> >> wait
> >> > > for
> >> > > > the next one?
> >> > > >
> >> > > > --
> >> > > > Josh Tynjala
> >> > > > Bowler Hat LLC <https://bowlerhat.dev>
> >> > > >
> >> > >
> >> > >
> >> > > --
> >> > > Andrew Wetmore
> >> > >
> >> > > Editor, Moose House Publications <https://moosehousepress.com/>
> >> > > Editor-Writer, The Apache Software Foundation <https://apache.org/>
> >> > >
> >> >
> >>
> >>
> >> --
> >>
> >> Piotr Zarzycki
> >>
> >
>
--
Piotr Zarzycki
Re: Super method calls not currently checked for errors
Posted by Josh Tynjala <jo...@bowlerhat.dev>.
Oh, and make sure that the pom.xml files in your app projects are using
version 0.9.12 for the compiler too.
--
Josh Tynjala
Bowler Hat LLC <https://bowlerhat.dev>
On Thu, Oct 10, 2024 at 8:15 AM Josh Tynjala <jo...@bowlerhat.dev>
wrote:
> Hey Piotr,
>
> I just pushed one more commit to the fix/super-signature-checks branch of
> royale-asjs.
>
> I noticed that the Maven build was not using my changes in royale-compiler
> because we're midway through the process of updating the repos from 0.9.11
> to 0.9.12. So you might not have actually tested my changes if you were
> building with Maven, unless you had manually made this tweak to pom.xml too.
>
> --
> Josh Tynjala
> Bowler Hat LLC <https://bowlerhat.dev>
>
>
> On Thu, Oct 10, 2024 at 2:18 AM Piotr Zarzycki <pi...@gmail.com>
> wrote:
>
>> Hi Josh,
>>
>> I have used your branches to build locally Royale. I have checked our main
>> application Jewel based with your changes and everything seems to be fine.
>> Application is working like a charm debug and release version. No errors
>> in
>> DOM elements on the console etc.
>>
>> I will check soon another app to confirm, but currently it looks really
>> good.
>>
>> Thanks,
>> Piotr
>>
>> śr., 9 paź 2024 o 22:54 Josh Tynjala <jo...@bowlerhat.dev>
>> napisał(a):
>>
>> > I too tend to be extra wary to make changes close to a release.
>> Sometimes,
>> > even seemingly minor bug fixes can introduce unexpected side effects!
>> >
>> > I totally don't mind waiting to merge this stuff.
>> >
>> > --
>> > Josh Tynjala
>> > Bowler Hat LLC <https://bowlerhat.dev>
>> >
>> >
>> > On Wed, Oct 9, 2024 at 11:20 AM Andrew Wetmore <co...@gmail.com>
>> > wrote:
>> >
>> > > I am always scared of putting a major, necessary fix that might have
>> > > downstream impact into a release that is about to go out the door. I
>> > would
>> > > vote for making this part, perhaps the major part, of the next
>> release,
>> > and
>> > > using the time between this one and that one to find and fix as much
>> as
>> > we
>> > > can in our own code...and maybe prepare some examples with fixes to
>> help
>> > > Royale users.
>> > >
>> > > On Wed, Oct 9, 2024 at 2:11 PM Josh Tynjala <
>> joshtynjala@bowlerhat.dev>
>> > > wrote:
>> > >
>> > > > Hey all,
>> > > >
>> > > > A user of vscode-as3mxml pointed out to me recently that the
>> arguments
>> > > > passed to super() and super.method() calls are not currently checked
>> > for
>> > > > errors by our compiler. That includes checking the types of the
>> > > arguments,
>> > > > and even whether the number of arguments is actually correct!
>> However,
>> > > > regular function calls without super are correctly type checked, so
>> it
>> > is
>> > > > only in cases of inheritance where super is used.
>> > > >
>> > > > I have royale-compiler fixes ready to go.
>> > > >
>> > > > This bug has been in the compiler since the code was donated by
>> Adobe.
>> > So
>> > > > it's definitely not even remotely new, and it doesn't necessarily
>> need
>> > to
>> > > > delay the current release in progress.
>> > > >
>> > > > However, it may impact a lot of existing code once the compiler can
>> > start
>> > > > detecting the wrong types. Several of our framework libraries in
>> > > > royale-asjs are affected. It's likely that user projects, especially
>> > ones
>> > > > with a lot of inheritance, will be affected too.
>> > > >
>> > > > In some cases, it just requires an extra cast. However, in a few
>> cases
>> > > that
>> > > > I've seen so far, it is actually showing that there are bugs in our
>> > code.
>> > > > I'm working through the necessary fixes in royale-asjs now.
>> > > >
>> > > > I just wanted to give a heads up about this, as it may have some
>> impact
>> > > on
>> > > > everyone's projects.
>> > > >
>> > > > Any thoughts on pushing this into the 0.9.12? Or would you rather
>> wait
>> > > for
>> > > > the next one?
>> > > >
>> > > > --
>> > > > Josh Tynjala
>> > > > Bowler Hat LLC <https://bowlerhat.dev>
>> > > >
>> > >
>> > >
>> > > --
>> > > Andrew Wetmore
>> > >
>> > > Editor, Moose House Publications <https://moosehousepress.com/>
>> > > Editor-Writer, The Apache Software Foundation <https://apache.org/>
>> > >
>> >
>>
>>
>> --
>>
>> Piotr Zarzycki
>>
>
Re: Super method calls not currently checked for errors
Posted by Josh Tynjala <jo...@bowlerhat.dev>.
Hey Piotr,
I just pushed one more commit to the fix/super-signature-checks branch of
royale-asjs.
I noticed that the Maven build was not using my changes in royale-compiler
because we're midway through the process of updating the repos from 0.9.11
to 0.9.12. So you might not have actually tested my changes if you were
building with Maven, unless you had manually made this tweak to pom.xml too.
--
Josh Tynjala
Bowler Hat LLC <https://bowlerhat.dev>
On Thu, Oct 10, 2024 at 2:18 AM Piotr Zarzycki <pi...@gmail.com>
wrote:
> Hi Josh,
>
> I have used your branches to build locally Royale. I have checked our main
> application Jewel based with your changes and everything seems to be fine.
> Application is working like a charm debug and release version. No errors in
> DOM elements on the console etc.
>
> I will check soon another app to confirm, but currently it looks really
> good.
>
> Thanks,
> Piotr
>
> śr., 9 paź 2024 o 22:54 Josh Tynjala <jo...@bowlerhat.dev>
> napisał(a):
>
> > I too tend to be extra wary to make changes close to a release.
> Sometimes,
> > even seemingly minor bug fixes can introduce unexpected side effects!
> >
> > I totally don't mind waiting to merge this stuff.
> >
> > --
> > Josh Tynjala
> > Bowler Hat LLC <https://bowlerhat.dev>
> >
> >
> > On Wed, Oct 9, 2024 at 11:20 AM Andrew Wetmore <co...@gmail.com>
> > wrote:
> >
> > > I am always scared of putting a major, necessary fix that might have
> > > downstream impact into a release that is about to go out the door. I
> > would
> > > vote for making this part, perhaps the major part, of the next release,
> > and
> > > using the time between this one and that one to find and fix as much as
> > we
> > > can in our own code...and maybe prepare some examples with fixes to
> help
> > > Royale users.
> > >
> > > On Wed, Oct 9, 2024 at 2:11 PM Josh Tynjala <joshtynjala@bowlerhat.dev
> >
> > > wrote:
> > >
> > > > Hey all,
> > > >
> > > > A user of vscode-as3mxml pointed out to me recently that the
> arguments
> > > > passed to super() and super.method() calls are not currently checked
> > for
> > > > errors by our compiler. That includes checking the types of the
> > > arguments,
> > > > and even whether the number of arguments is actually correct!
> However,
> > > > regular function calls without super are correctly type checked, so
> it
> > is
> > > > only in cases of inheritance where super is used.
> > > >
> > > > I have royale-compiler fixes ready to go.
> > > >
> > > > This bug has been in the compiler since the code was donated by
> Adobe.
> > So
> > > > it's definitely not even remotely new, and it doesn't necessarily
> need
> > to
> > > > delay the current release in progress.
> > > >
> > > > However, it may impact a lot of existing code once the compiler can
> > start
> > > > detecting the wrong types. Several of our framework libraries in
> > > > royale-asjs are affected. It's likely that user projects, especially
> > ones
> > > > with a lot of inheritance, will be affected too.
> > > >
> > > > In some cases, it just requires an extra cast. However, in a few
> cases
> > > that
> > > > I've seen so far, it is actually showing that there are bugs in our
> > code.
> > > > I'm working through the necessary fixes in royale-asjs now.
> > > >
> > > > I just wanted to give a heads up about this, as it may have some
> impact
> > > on
> > > > everyone's projects.
> > > >
> > > > Any thoughts on pushing this into the 0.9.12? Or would you rather
> wait
> > > for
> > > > the next one?
> > > >
> > > > --
> > > > Josh Tynjala
> > > > Bowler Hat LLC <https://bowlerhat.dev>
> > > >
> > >
> > >
> > > --
> > > Andrew Wetmore
> > >
> > > Editor, Moose House Publications <https://moosehousepress.com/>
> > > Editor-Writer, The Apache Software Foundation <https://apache.org/>
> > >
> >
>
>
> --
>
> Piotr Zarzycki
>
Re: Super method calls not currently checked for errors
Posted by Piotr Zarzycki <pi...@gmail.com>.
Hi Josh,
I have used your branches to build locally Royale. I have checked our main
application Jewel based with your changes and everything seems to be fine.
Application is working like a charm debug and release version. No errors in
DOM elements on the console etc.
I will check soon another app to confirm, but currently it looks really
good.
Thanks,
Piotr
śr., 9 paź 2024 o 22:54 Josh Tynjala <jo...@bowlerhat.dev> napisał(a):
> I too tend to be extra wary to make changes close to a release. Sometimes,
> even seemingly minor bug fixes can introduce unexpected side effects!
>
> I totally don't mind waiting to merge this stuff.
>
> --
> Josh Tynjala
> Bowler Hat LLC <https://bowlerhat.dev>
>
>
> On Wed, Oct 9, 2024 at 11:20 AM Andrew Wetmore <co...@gmail.com>
> wrote:
>
> > I am always scared of putting a major, necessary fix that might have
> > downstream impact into a release that is about to go out the door. I
> would
> > vote for making this part, perhaps the major part, of the next release,
> and
> > using the time between this one and that one to find and fix as much as
> we
> > can in our own code...and maybe prepare some examples with fixes to help
> > Royale users.
> >
> > On Wed, Oct 9, 2024 at 2:11 PM Josh Tynjala <jo...@bowlerhat.dev>
> > wrote:
> >
> > > Hey all,
> > >
> > > A user of vscode-as3mxml pointed out to me recently that the arguments
> > > passed to super() and super.method() calls are not currently checked
> for
> > > errors by our compiler. That includes checking the types of the
> > arguments,
> > > and even whether the number of arguments is actually correct! However,
> > > regular function calls without super are correctly type checked, so it
> is
> > > only in cases of inheritance where super is used.
> > >
> > > I have royale-compiler fixes ready to go.
> > >
> > > This bug has been in the compiler since the code was donated by Adobe.
> So
> > > it's definitely not even remotely new, and it doesn't necessarily need
> to
> > > delay the current release in progress.
> > >
> > > However, it may impact a lot of existing code once the compiler can
> start
> > > detecting the wrong types. Several of our framework libraries in
> > > royale-asjs are affected. It's likely that user projects, especially
> ones
> > > with a lot of inheritance, will be affected too.
> > >
> > > In some cases, it just requires an extra cast. However, in a few cases
> > that
> > > I've seen so far, it is actually showing that there are bugs in our
> code.
> > > I'm working through the necessary fixes in royale-asjs now.
> > >
> > > I just wanted to give a heads up about this, as it may have some impact
> > on
> > > everyone's projects.
> > >
> > > Any thoughts on pushing this into the 0.9.12? Or would you rather wait
> > for
> > > the next one?
> > >
> > > --
> > > Josh Tynjala
> > > Bowler Hat LLC <https://bowlerhat.dev>
> > >
> >
> >
> > --
> > Andrew Wetmore
> >
> > Editor, Moose House Publications <https://moosehousepress.com/>
> > Editor-Writer, The Apache Software Foundation <https://apache.org/>
> >
>
--
Piotr Zarzycki
Re: Super method calls not currently checked for errors
Posted by Josh Tynjala <jo...@bowlerhat.dev>.
I too tend to be extra wary to make changes close to a release. Sometimes,
even seemingly minor bug fixes can introduce unexpected side effects!
I totally don't mind waiting to merge this stuff.
--
Josh Tynjala
Bowler Hat LLC <https://bowlerhat.dev>
On Wed, Oct 9, 2024 at 11:20 AM Andrew Wetmore <co...@gmail.com> wrote:
> I am always scared of putting a major, necessary fix that might have
> downstream impact into a release that is about to go out the door. I would
> vote for making this part, perhaps the major part, of the next release, and
> using the time between this one and that one to find and fix as much as we
> can in our own code...and maybe prepare some examples with fixes to help
> Royale users.
>
> On Wed, Oct 9, 2024 at 2:11 PM Josh Tynjala <jo...@bowlerhat.dev>
> wrote:
>
> > Hey all,
> >
> > A user of vscode-as3mxml pointed out to me recently that the arguments
> > passed to super() and super.method() calls are not currently checked for
> > errors by our compiler. That includes checking the types of the
> arguments,
> > and even whether the number of arguments is actually correct! However,
> > regular function calls without super are correctly type checked, so it is
> > only in cases of inheritance where super is used.
> >
> > I have royale-compiler fixes ready to go.
> >
> > This bug has been in the compiler since the code was donated by Adobe. So
> > it's definitely not even remotely new, and it doesn't necessarily need to
> > delay the current release in progress.
> >
> > However, it may impact a lot of existing code once the compiler can start
> > detecting the wrong types. Several of our framework libraries in
> > royale-asjs are affected. It's likely that user projects, especially ones
> > with a lot of inheritance, will be affected too.
> >
> > In some cases, it just requires an extra cast. However, in a few cases
> that
> > I've seen so far, it is actually showing that there are bugs in our code.
> > I'm working through the necessary fixes in royale-asjs now.
> >
> > I just wanted to give a heads up about this, as it may have some impact
> on
> > everyone's projects.
> >
> > Any thoughts on pushing this into the 0.9.12? Or would you rather wait
> for
> > the next one?
> >
> > --
> > Josh Tynjala
> > Bowler Hat LLC <https://bowlerhat.dev>
> >
>
>
> --
> Andrew Wetmore
>
> Editor, Moose House Publications <https://moosehousepress.com/>
> Editor-Writer, The Apache Software Foundation <https://apache.org/>
>
Re: Super method calls not currently checked for errors
Posted by Josh Tynjala <jo...@bowlerhat.dev>.
In both the royale-compiler and royale-asjs repos, there is a new
"fix/super-signature-checks" branch with my changes.
Here are my commits to royale-asjs in the fix/super-signature-checks branch:
https://github.com/apache/royale-asjs/compare/28d35e6398e1c4093fa6475e0951cca2b5ffeac8...fix/super-signature-checks
I tried to separate each of the bug fixes into a separate commit, and I
combined all of the simple changes that required only a cast into a single
commit.
--
Josh Tynjala
Bowler Hat LLC <https://bowlerhat.dev>
On Wed, Oct 9, 2024 at 11:26 AM Piotr Zarzycki <pi...@gmail.com>
wrote:
> Hi Josh,
>
> Any chance put this on a branch so I could for example build framework by
> Maven and rebuild all our company Royale applications to see how it goes ?
>
> It would be quite reliable test I believe.
>
> Thanks,
> Piotr
>
> śr., 9 paź 2024 o 20:20 Andrew Wetmore <co...@gmail.com> napisał(a):
>
> > I am always scared of putting a major, necessary fix that might have
> > downstream impact into a release that is about to go out the door. I
> would
> > vote for making this part, perhaps the major part, of the next release,
> and
> > using the time between this one and that one to find and fix as much as
> we
> > can in our own code...and maybe prepare some examples with fixes to help
> > Royale users.
> >
> > On Wed, Oct 9, 2024 at 2:11 PM Josh Tynjala <jo...@bowlerhat.dev>
> > wrote:
> >
> > > Hey all,
> > >
> > > A user of vscode-as3mxml pointed out to me recently that the arguments
> > > passed to super() and super.method() calls are not currently checked
> for
> > > errors by our compiler. That includes checking the types of the
> > arguments,
> > > and even whether the number of arguments is actually correct! However,
> > > regular function calls without super are correctly type checked, so it
> is
> > > only in cases of inheritance where super is used.
> > >
> > > I have royale-compiler fixes ready to go.
> > >
> > > This bug has been in the compiler since the code was donated by Adobe.
> So
> > > it's definitely not even remotely new, and it doesn't necessarily need
> to
> > > delay the current release in progress.
> > >
> > > However, it may impact a lot of existing code once the compiler can
> start
> > > detecting the wrong types. Several of our framework libraries in
> > > royale-asjs are affected. It's likely that user projects, especially
> ones
> > > with a lot of inheritance, will be affected too.
> > >
> > > In some cases, it just requires an extra cast. However, in a few cases
> > that
> > > I've seen so far, it is actually showing that there are bugs in our
> code.
> > > I'm working through the necessary fixes in royale-asjs now.
> > >
> > > I just wanted to give a heads up about this, as it may have some impact
> > on
> > > everyone's projects.
> > >
> > > Any thoughts on pushing this into the 0.9.12? Or would you rather wait
> > for
> > > the next one?
> > >
> > > --
> > > Josh Tynjala
> > > Bowler Hat LLC <https://bowlerhat.dev>
> > >
> >
> >
> > --
> > Andrew Wetmore
> >
> > Editor, Moose House Publications <https://moosehousepress.com/>
> > Editor-Writer, The Apache Software Foundation <https://apache.org/>
> >
>
>
> --
>
> Piotr Zarzycki
>
Re: Super method calls not currently checked for errors
Posted by Piotr Zarzycki <pi...@gmail.com>.
Hi Josh,
Any chance put this on a branch so I could for example build framework by
Maven and rebuild all our company Royale applications to see how it goes ?
It would be quite reliable test I believe.
Thanks,
Piotr
śr., 9 paź 2024 o 20:20 Andrew Wetmore <co...@gmail.com> napisał(a):
> I am always scared of putting a major, necessary fix that might have
> downstream impact into a release that is about to go out the door. I would
> vote for making this part, perhaps the major part, of the next release, and
> using the time between this one and that one to find and fix as much as we
> can in our own code...and maybe prepare some examples with fixes to help
> Royale users.
>
> On Wed, Oct 9, 2024 at 2:11 PM Josh Tynjala <jo...@bowlerhat.dev>
> wrote:
>
> > Hey all,
> >
> > A user of vscode-as3mxml pointed out to me recently that the arguments
> > passed to super() and super.method() calls are not currently checked for
> > errors by our compiler. That includes checking the types of the
> arguments,
> > and even whether the number of arguments is actually correct! However,
> > regular function calls without super are correctly type checked, so it is
> > only in cases of inheritance where super is used.
> >
> > I have royale-compiler fixes ready to go.
> >
> > This bug has been in the compiler since the code was donated by Adobe. So
> > it's definitely not even remotely new, and it doesn't necessarily need to
> > delay the current release in progress.
> >
> > However, it may impact a lot of existing code once the compiler can start
> > detecting the wrong types. Several of our framework libraries in
> > royale-asjs are affected. It's likely that user projects, especially ones
> > with a lot of inheritance, will be affected too.
> >
> > In some cases, it just requires an extra cast. However, in a few cases
> that
> > I've seen so far, it is actually showing that there are bugs in our code.
> > I'm working through the necessary fixes in royale-asjs now.
> >
> > I just wanted to give a heads up about this, as it may have some impact
> on
> > everyone's projects.
> >
> > Any thoughts on pushing this into the 0.9.12? Or would you rather wait
> for
> > the next one?
> >
> > --
> > Josh Tynjala
> > Bowler Hat LLC <https://bowlerhat.dev>
> >
>
>
> --
> Andrew Wetmore
>
> Editor, Moose House Publications <https://moosehousepress.com/>
> Editor-Writer, The Apache Software Foundation <https://apache.org/>
>
--
Piotr Zarzycki
Re: Super method calls not currently checked for errors
Posted by Andrew Wetmore <co...@gmail.com>.
I am always scared of putting a major, necessary fix that might have
downstream impact into a release that is about to go out the door. I would
vote for making this part, perhaps the major part, of the next release, and
using the time between this one and that one to find and fix as much as we
can in our own code...and maybe prepare some examples with fixes to help
Royale users.
On Wed, Oct 9, 2024 at 2:11 PM Josh Tynjala <jo...@bowlerhat.dev>
wrote:
> Hey all,
>
> A user of vscode-as3mxml pointed out to me recently that the arguments
> passed to super() and super.method() calls are not currently checked for
> errors by our compiler. That includes checking the types of the arguments,
> and even whether the number of arguments is actually correct! However,
> regular function calls without super are correctly type checked, so it is
> only in cases of inheritance where super is used.
>
> I have royale-compiler fixes ready to go.
>
> This bug has been in the compiler since the code was donated by Adobe. So
> it's definitely not even remotely new, and it doesn't necessarily need to
> delay the current release in progress.
>
> However, it may impact a lot of existing code once the compiler can start
> detecting the wrong types. Several of our framework libraries in
> royale-asjs are affected. It's likely that user projects, especially ones
> with a lot of inheritance, will be affected too.
>
> In some cases, it just requires an extra cast. However, in a few cases that
> I've seen so far, it is actually showing that there are bugs in our code.
> I'm working through the necessary fixes in royale-asjs now.
>
> I just wanted to give a heads up about this, as it may have some impact on
> everyone's projects.
>
> Any thoughts on pushing this into the 0.9.12? Or would you rather wait for
> the next one?
>
> --
> Josh Tynjala
> Bowler Hat LLC <https://bowlerhat.dev>
>
--
Andrew Wetmore
Editor, Moose House Publications <https://moosehousepress.com/>
Editor-Writer, The Apache Software Foundation <https://apache.org/>