Hi,
I just committed the fix along with some code clean up [1] and attached the
fixed fdb.jar [2] for testing purpose before the 4.12 release.
Enjoy,
Frédéric THOMAS
PS: Sorry for having mixed the fix with the code clean up though.
[1]https://issues.apache.org/jira/browse/FLEX-34062
[2] https
Since the files are signed you can host them on your domain too. This is
what we did with Flex 4.6 RSLs. We had the config check both adobe and then
fall back to our servers. Since they are signed it would put them in the RSL
cache either way.
Now obviously it can only be hosted on yo
Hi,
Really hard to review this given the amount of changes - most which seem while
space?
Justin
I ordered an asus fhd 10 (x86 android) and noticed the same thing while
testing my air android game. My app runs slow on these devices. The problem
is not just air either, some asus apps won't start and windows RDP for
android runs faster in software mode. The new frankenstien devices that
people a
Yes, there could be some white spaces too, I apologized already I committed the
fix AND the code clean up, the more important changes are in DebugCLI though,
relative to the waitForMetaData function, basically I removed its usage, the
rest is code clean up.
Thanks,
Frédéric THOMAS
> Subject: R
If you use IntelliJ IDEA - its Diff window has an option to ignore all
changes in whitespaces.
On 27.01.2014 15:40, Justin Mclean wrote:
Hi,
Really hard to review this given the amount of changes - most which seem while
space?
Justin
Hi,
> If you use IntelliJ IDEA - its Diff window has an option to ignore all
> changes in whitespaces.
The SVN change log have nothing to do with which IDE you use - you can ignore
whitespace changes in eclipse /flash builder as well but it doesn't help in
this case.
Justin
Hi,
Sorry going to have to vote -1 on this change.
As well as the white space changes making it hard to see what the changes are
there's quite a few few unrelated changes in here.
For instance:
- else if (expect == have)
- yes = true;
- else
-
Hi,
And for instance another concerning change - I can see than addFrame return
type was modified but is this really needed?
- if (addFrame(c))
- ++depth;
-// System.out.println
The SVN change log have nothing to do with which IDE you use - you can
ignore whitespace changes in eclipse /flash builder as well but it
doesn't help in this case. Justin
Ok, so Eclipse / Flash Builder + some SVN support plugin will do the
same trick.
No more flooding, sorry :)
Hi Justin,
Ok, maybe better I revert and re-commit only the fix
as and don't care about the useless, unused and confusing code, I'll do
it tonight, no time at the moment.
> I can see than addFrame return type was modified but is this really needed?
IIRC, addFrame was returning always true, I
Np and thanks for the tip, anyway , I have been a bit quick to commit.
The fix is still applied, can you check on your side everything works as
expected ?
Thanks,
Frédéric THOMAS
> Date: Mon, 27 Jan 2014 17:24:23 +0400
> From: alexander.doros...@jetbrains.com
> To: dev@flex.apache.org
> Subject
Fred,
You might want to hold off on code clean up on FDB. Not cleaning it up
will make it a bit easier to compare the new version of FDB that should be
cleared for donation soon (with workers support) against the existing one.
Then you can clean up afterwards.
-Alex
On 1/27/14 2:44 AM, "Frédér
I think I'll submit a FlexJS session. I saw they are looking for more
proposals. Anything else we should be talking about? Should I also
submit a session on the current SDK?
On 1/17/14 1:06 PM, "Jun Heider" wrote:
>On Jan 17, 2014, at 2:01 PM, Alex Harui wrote:
>
>> Do you know if speakerss
Hi Alex,
Yep, I thought about that too at the begining but then, given my changes was
not structural, I thought it was a good thing (less confusion, better code
reading, etc...), the bad one is I committed it along with the fix.
Anyway, maybe I'll be the one who will merge the upcoming one, may
I was able to get everything installed but when compiling I get a error:
Exception in thread "main" java.lang.UnsupportedClassVersionError:
org/apache/flex/compiler/clients/MXMLC : Unsupported major.minor version
51.0
at java.lang.ClassLoader.defineClass1(Native Method)
at java.lang.ClassLoader.
The launch configurations created by the installer picked a version of
Java by looking for JAVA_HOME environment variable or running which on Mac
and where.exe on Windows.
It sounds like it picked an older version of Java, which could be a bug in
how the installer found the Java version. Check th
Btw, it has been reverted and re-committed with the fix only !
Frédéric THOMAS
> From: ftho...@apache.org
> To: comm...@flex.apache.org
> Date: Mon, 27 Jan 2014 17:23:57 +
> Subject: [5/5] git commit: [flex-sdk] [refs/heads/develop] - FLEX-34062: FDB
> waits up to 20s per SWF before we can s
I just tried to clean up some of the JSHint warnings in the JS code we've
written for FlexJS. The first thing I ran into was the way we declare
uninitialized variables. An example is:
/**
* @expose
* @type {string}
*/
org.apache.flex.binding.BindingBase.prototype.sourceID;
The closure linter
A quick update:
* There are a few Mustella test failures that I am trying to fix
unsuccessfully. If someone can help me deal with those, that would be
great. Getting 100% success of mustella tests is key before we can start
cloning the machines.
* My Windows Azure credits for this month has e
Send out a reminder when the VM is available and I will try to take a look.
On 1/27/14 12:13 PM, "OmPrakash Muppirala" wrote:
>A quick update:
>
>* There are a few Mustella test failures that I am trying to fix
>unsuccessfully. If someone can help me deal with those, that would be
>great. Get
Justin suggested that the installer should save downloads in a central
location on the users computer so that if multiple SDKs are installed that
use, for instance, the huge AIR SDK, it only gets downloaded once.
That seems like a reasonable idea to me. Any thoughts on where to store
these downlo
On Mon, Jan 27, 2014 at 12:33 PM, Alex Harui wrote:
> Justin suggested that the installer should save downloads in a central
> location on the users computer so that if multiple SDKs are installed that
> use, for instance, the huge AIR SDK, it only gets downloaded once.
>
> That seems like a reas
I can report that the Installer worked perfectly well (had to uninstall
current installer first to get rid of the certificate related AIR error)
The example app compiles and runs fine (with a compilation warning, though)
Fantastic job, Alex (and others) This is a huge huge milestone for
FlexJS.
Hi,
Much nicer and easier to understand. Thanks for that.
If you want to clean up the code go for it but try and commit in small bits
explaining each change so the changes can be reviewed easily. We currently not
running unit test for this code so need to be a little careful in making
changes.
I needed to install 1.7 51 SDK again and everything started working
perfect. I had to add the bindable tag to the Strings getter but after that
it compiled perfectly.
[*Bindable*(*"stringsChanged"*)]
*public* *function* *get* strings():Array
{
*return* _strings;
On 1/27/14 12:37 PM, "OmPrakash Muppirala" wrote:
>On Mon, Jan 27, 2014 at 12:33 PM, Alex Harui wrote:
>
>> Justin suggested that the installer should save downloads in a central
>> location on the users computer so that if multiple SDKs are installed
>>that
>> use, for instance, the huge AIR
In my "other" projects I always assign an initial value to these properties.
Closure is OK with empty strings and -1 (or whatever) for numbers.
For complex types you will have to take into account that the type
annotation must match the initial value, e.g. if you want null to be
the initial value
OK thanks. I assume Closure is smart about these assignments? Otherwise
it seems wasteful to assign the same default value.
-Alex?
On 1/27/14 1:13 PM, "Erik de Bruin" wrote:
>In my "other" projects I always assign an initial value to these
>properties.
>
>Closure is OK with empty strings and -
Hi Justin,
I can't process as you say, there are small changes all over the classes,
create a JIRA for each change would be a nightmare, if I do it again, it will
be in one shot again, there are no unit tests but I do manual functional tests
on my company projects and ask the team to do the sam
Hi,
> I can't process as you say, there are small changes all over the classes,
> create a JIRA for each change would be a nightmare
No need for a JIRA for each change - just a few more commits explaining the
changes so they can be understood step by step.
Some changes while valid (eg changing
Justin,
yes, indeed, I'm trying to, making tests in IJ and FB, in many ways I can.
Frédéric THOMAS
> Subject: Re: [5/5] git commit: [flex-sdk] [refs/heads/develop] - FLEX-34062:
> FDB waits up to 20s per SWF before we can start debugging (without code clean
> up)
> From: jus...@classsoftware.c
This question is probably best answered by Om, but I think it would be
better to replace RuntimeLocale.as with a simple but extensible properties
files that get loaded as needed based on the locale. That way, new
locales can be added without re-compiling the installer. Is there some
aspect of Run
On Mon, Jan 27, 2014 at 1:41 PM, Alex Harui wrote:
> This question is probably best answered by Om, but I think it would be
> better to replace RuntimeLocale.as with a simple but extensible properties
> files that get loaded as needed based on the locale. That way, new
> locales can be added wit
On Mon, Jan 27, 2014 at 1:09 PM, Alex Harui wrote:
>
>
> On 1/27/14 12:37 PM, "OmPrakash Muppirala" wrote:
>
> >On Mon, Jan 27, 2014 at 12:33 PM, Alex Harui wrote:
> >
> >> Justin suggested that the installer should save downloads in a central
> >> location on the users computer so that if mult
On 1/27/14 2:24 PM, "OmPrakash Muppirala" wrote:
>On Mon, Jan 27, 2014 at 1:09 PM, Alex Harui wrote:
>
>>
>>
>> On 1/27/14 12:37 PM, "OmPrakash Muppirala" wrote:
>>
>> >On Mon, Jan 27, 2014 at 12:33 PM, Alex Harui wrote:
>> >
>> >> Justin suggested that the installer should save downloads in
The GCC is all about optimisation. If it can discard or inline a value
or assignment, it will. The default value may even help give the
compiler (and the developer) additional clues that will allow it to
better optimise the code.
EdB
On Mon, Jan 27, 2014 at 10:17 PM, Alex Harui wrote:
> OK tha
On Mon, Jan 27, 2014 at 2:35 PM, Alex Harui wrote:
>
>
> On 1/27/14 2:24 PM, "OmPrakash Muppirala" wrote:
>
> >On Mon, Jan 27, 2014 at 1:09 PM, Alex Harui wrote:
> >
> >>
> >>
> >> On 1/27/14 12:37 PM, "OmPrakash Muppirala"
> wrote:
> >>
> >> >On Mon, Jan 27, 2014 at 12:33 PM, Alex Harui wrot
38 matches
Mail list logo
