Hi,
> Then we need to decide whether we want the install scripts to prompt folks
> to accept Saxon or not, and whether we should continue to have folks
> approve OSMF and SWFObject like we currently do.
We should probably be consistent. If we ask for OSMF we should for Saxon. It
seems clear to
I just made another commit. Now Down to 3. :-)
[flexunit] Tests run: 428, Failures: 3, Errors: 0, Skipped: 42, Time elapsed:
155.729 sec
On Dec 29, 2014, at 9:53 PM, Erik de Bruin wrote:
> Oh, there are plenty of Mustella tests failing ;-) But those failures
> don't seem to be related to TLF,
Other than Saxon, I think I’m mostly done fixing up LICENSE and NOTICE so
that it matches the way we packaged prior releases. I hope to finish
Saxon and check it all in tonight or tomorrow. Then it needs review and
we might find a few more things we need to change.
Then we need to decide whether
Good to know you got it working. Patches to fix case-sensitivity issues
are welcome.
-Alex
On 12/29/14, 8:43 AM, "Left Right" wrote:
>Case sensitivity also causes some tests to fail:
>
>[junit] /home/wvxvw/projects/flex-tlf/compile-config.xml:37
>[junit] Error: unable to open
>'/home/w
Hi,
> Do you have the cycles to create a suggested fix for both?
As I said I will have later this week, but Alex is currently looking into it
and would need to check in what he's done first I think.
One of my earlier posts in this thread had details on what the likely changes
would be.
Thanks
> Currently the LICENSE and NOTICE files are incorrect, the simple solution is
> to fix them.
Do you have the cycles to create a suggested fix for both?
EdB
--
Ix Multimedia Software
Jan Luykenstraat 27
3521 VB Utrecht
T. 06-51952295
I. www.ixsoftware.nl
Hi,
> OK. A little bit dramatic, and half those options exist anyway, as
> previous releases have the same issues.
Yep as discussed we need to fix those as well.
Can be summed up in one line:
Currently the LICENSE and NOTICE files are incorrect, the simple solution is to
fix them.
Justin
>> No, I meant what would happen if we went ahead a binary release
>> without fixing this...
>
> We can't as we now know it would be against Apache policy. Worse cases
> include the board taking action and ask us to remove the releases or
> disbanding the PMC, a 3rd party could take us to court,
Hi,
> That was my point. Unless we find out that Saxon has some unfriendly
> dependencies, the question of whether a binary dependency in the package
> was compiled from sources or not or whether the Installer should ask about
> MPL and MIT dependencies probably doesn't put the foundation at risk
Hi,
> No, I meant what would happen if we went ahead a binary release
> without fixing this...
We can't as we now know it would be against Apache policy. Worse cases include
the board taking action and ask us to remove the releases or disbanding the
PMC, a 3rd party could take us to court, busi
Oh, there are plenty of Mustella tests failing ;-) But those failures
don't seem to be related to TLF, as far as I can tell.
So, you're 0.5 errors and 4 failures away from "being done"!
EdB
On Mon, Dec 29, 2014 at 8:46 PM, Harbs wrote:
> I thought I remembered that besides the unit test, ther
I thought I remembered that besides the unit test, there were mustella tests
failing.
If not, all the better… ;-)
On Dec 29, 2014, at 9:34 PM, Erik de Bruin wrote:
> Just curious, why are you looking at Mustella tests? There don't seem
> to be Mustella tests failing in a manner that would indi
Just curious, why are you looking at Mustella tests? There don't seem
to be Mustella tests failing in a manner that would indicate a problem
with TLF.
EdB
On Mon, Dec 29, 2014 at 7:41 PM, Harbs wrote:
> Is there a way to see what Mustella tests are failing on the CI server? I’d
> rather know
It passes when I run the TabTest alone.
It seems to be testing for a Flash Player bug rather than a TLF bug and the
current test does not seem to be doing anything:
//the long string value is the correct test data. Since this is a
Player bug, we can't check in the correct test data
Is there a way to see what Mustella tests are failing on the CI server? I’d
rather know what I’m dealing with before trying to run the full suite of tests
on a local machine…
On Dec 29, 2014, at 8:08 PM, Erik de Bruin wrote:
> Just ran again with 1.8, got the error. Then with 1.6: no error. I
It sounds like a timing issue…
I’ll see what I can do about tracking it down.
On Dec 29, 2014, at 8:08 PM, Erik de Bruin wrote:
> Just ran again with 1.8, got the error. Then with 1.6: no error. I
> threw in 1.7 for good measure: zero errors.
>
> EdB
>
>
>
> On Mon, Dec 29, 2014 at 6:04 PM,
Just ran again with 1.8, got the error. Then with 1.6: no error. I
threw in 1.7 for good measure: zero errors.
EdB
On Mon, Dec 29, 2014 at 6:04 PM, Harbs wrote:
> Weird. I think my machine (Mac 10.9.5) uses Java 1.6.
>
> I get the error.
>
> On Dec 29, 2014, at 3:47 PM, Erik de Bruin wrote:
>
Oops... a little - minor? - 'major/minor' error there ;-)
EdB
On Mon, Dec 29, 2014 at 5:14 PM, Erik de Bruin wrote:
> Can you give me a (realistic) worst case scenario if we were to
> release without making any changes?
IMO Worse case we can't make a release until this is res
Weird. I think my machine (Mac 10.9.5) uses Java 1.6.
I get the error.
On Dec 29, 2014, at 3:47 PM, Erik de Bruin wrote:
> Fun fact: when run with Java (1.)6, I get:
>
> Tests run: 428, Failures: 5, Errors: 0, Skipped: 41, Time elapsed: 246.937 sec
>
> Since the Windows build machine also run
Case sensitivity also causes some tests to fail:
[junit] /home/wvxvw/projects/flex-tlf/compile-config.xml:37
[junit] Error: unable to open
'/home/wvxvw/projects/flex-tlf/textlayout/manifest.xml'.
[junit] /home/wvxvw/projects/flex-tlf/compile-config.xml (line: 37)
[junit]
On 12/29/14, 8:12 AM, "Left Right" wrote:
>Apparently it copied it to a wrong place... where should it be? After
>the build I have it in $FALCON_HOME/../lib/ but that doesn't seem to
>be the proper place for it, or is it?
>Actually, the build doesn't finish properly because of the tests, is
>th
PS. There just aren't debug players on Linux any more, launching a
browser won't help :) Wine is the last option (though it seems
reasonably solid, except the problem with multiple instances of the
flash player all drawing regardless of the browser tab, weird
fullscreen mode and intermittent but ra
Can you give me a (realistic) worst case scenario if we were to
release without making any changes?
>>>
>>> IMO Worse case we can't make a release until this is resolved.
>>
>>No, I meant what would happen if we went ahead a binary release
>>without fixing this... Tempting, because we did
Apparently it copied it to a wrong place... where should it be? After
the build I have it in $FALCON_HOME/../lib/ but that doesn't seem to
be the proper place for it, or is it?
Actually, the build doesn't finish properly because of the tests, is
there any simple way to skip the tests? Maybe it woul
On 12/29/14, 1:40 AM, "Erik de Bruin" wrote:
>>> Can you give me a (realistic) worst case scenario if we were to
>>> release without making any changes?
>>
>> IMO Worse case we can't make a release until this is resolved.
>
>No, I meant what would happen if we went ahead a binary release
>witho
On 12/29/14, 1:16 AM, "Justin Mclean" wrote:
>Hi,
>
>> -Can a binary package bundle a binary without its source?
>
>The answer is yes to that, and in fact we a little unusual as we do
>include the SDK source in a binary release.
According to what documents? I see that my question was ambiguo
The download script should have downloaded flex-tool-api.jar. The ant
script does run on OSX and not just Windows. Is there some difference in
how Ant works on Linux that changes when tasks are resolved?
I don’t know off-hand what the SWF tests might find that the unit tests
don’t. It might als
Fun fact: when run with Java (1.)6, I get:
Tests run: 428, Failures: 5, Errors: 0, Skipped: 41, Time elapsed: 246.937 sec
Since the Windows build machine also runs Java (1.)6, the error seems
to correlate to the Java edition?
Ed
On Mon, Dec 29, 2014 at 2:28 PM, Erik de Bruin wrote:
> On my Mac
On my Mac, this is the current status:
Tests run: 428, Failures: 4, Errors: 1, Skipped: 41, Time elapsed: 253.360 sec
EdB
On Mon, Dec 29, 2014 at 11:01 AM, Harbs wrote:
> I’ve made some more progress. We’re down to 5 failures and 1 error:
> [flexunit] Suite: UnitTest.Tests.OperationTest
> [
Hi Harbs,
It's even better - In jenkins build we do not have these 1 error :)
Great job!
Piotr
-
Apache Flex PMC
piotrzarzyck...@gmail.com
--
View this message in context:
http://apache-flex-development.247.n4.nabble.com/DISCUSS-Release-Apache-Flex-4-14-0-tp43390p43943.html
Sent from
I'm trying to build a project previous built with SDK 4.6 and I'm getting this:
[mxmlc] /home/wvxvw/projects/flex-sdk/frameworks/flex-config.xml(65):
Error: unable to open '{playerglobalHome}/11.1/playerglobal.swc'
[mxmlc]
[mxmlc]
The "missing" SWC is here:
/home/wvxvw/projects
Another question: what do I miss if some junit tests fail
(specifically those related to running SWF in debugger - there simply
isn't a Flash debug player that runs on Linux, unless it's in Wine, so
I don't really know how to set up those tests).
Thanks
On Mon, Dec 29, 2014 at 11:22 AM, Left Righ
I’ve made some more progress. We’re down to 5 failures and 1 error:
[flexunit] Suite: UnitTest.Tests.OperationTest
[flexunit] Tests run: 47, Failures: 2, Errors: 0, Skipped: 2, Time elapsed:
5.002 sec
[flexunit] Suite: UnitTest.Tests.GeneralFunctionsTest
[flexunit] Tests run: 36, Failures: 0,
>> Can you give me a (realistic) worst case scenario if we were to
>> release without making any changes?
>
> IMO Worse case we can't make a release until this is resolved.
No, I meant what would happen if we went ahead a binary release
without fixing this... Tempting, because we did that nearly a
And trying to run falcon gives me:
/home/wvxvw/projects/flex-falcon/compiler/commandline/mxmlc --help
Using Flex SDK: /home/wvxvw/.wine/drive_c/flex/sdk/4.6.0
Exception in thread "main" java.lang.NoClassDefFoundError:
org/apache/flex/tools/FlexTool
at java.lang.ClassLoader.defineClass1(Native Meth
HI,
> Can you give me a (realistic) worst case scenario if we were to
> release without making any changes?
IMO Worse case we can't make a release until this is resolved. We could make a
source release without any changes but given the installer uses the binary
that's hardly ideal.
Justin
Hi,
> -Can a binary package bundle a binary without its source?
The answer is yes to that, and in fact we a little unusual as we do include the
SDK source in a binary release.
> -Was there any past discussion that caused the Installer to ask about
> accepting SWFObject’s MIT License?
At worse
OK, I figured out that par of the problem.
Ant defines macro to have JFlex.jar, but downloads jflex.jar,
that would probably work on Windows, but not on Linux. With that
fixed (why not fail when generating parser files fails?) I get to
this:
[javac]
/home/wvxvw/projects/flex-falcon/compile
The protocol I follow is this:
I get 'flex-falcon' AND 'flex-sdk' from Git. Then, in 'flex-falcon',
in the root of that directory - NOT in e.g. the 'compiler'
subdirectory - I run 'ant main'. This takes care of all the dependency
downloads as well as the copying of the relevant parts of the SDK,
e
I had what git clone brought me.. but now I replaced it with whatever
yum had installed and it finds Guava, so that's ok.
Next problem, however:
[javac]
/home/wvxvw/projects/flex-falcon/compiler/generated/src/org/apache/flex/compiler/internal/parsing/mxml/RawMXMLTokenizer.java:339:
error: ca
Can you give me a (realistic) worst case scenario if we were to
release without making any changes? In other words: is this issue a
blocker for the 4.14 release or just an issue that we need to take
care of for a future release?
EdB
On Mon, Dec 29, 2014 at 9:42 AM, Alex Harui wrote:
>
>
> On 1
Could you have an old guava? I think we’re using Guava17. My copy has
that class.
-Alex
On 12/29/14, 12:41 AM, "Left Right" wrote:
>Actually, I've checked, CacheBuilder isn't in that guava.jar.
>
>On Mon, Dec 29, 2014 at 10:38 AM, Left Right
>wrote:
>> Hi Alex,
>>
>> I've tried to build from
You can download the 'bad' pngs from this URL (mind the
'[buildnumber]' replacer):
http://flex-mustella.cloudapp.net/job/flex-sdk_mustella/[buildnumber]/artifact/*zip*/archive.zip
We store the bad pngs from the last 10 builds from all Mustella jobs
(main, AIR and mobile; regular and RC).
EdB
On 12/28/14, 12:29 PM, "Justin Mclean" wrote:
>Hi,
>
>> Since binary packages are not an act of the foundation, other than the
>>explicit
>> statement that LICENSE and NOTICE must match the contents of the binary
>> package, I can’t imagine that it puts the foundation at risk if we guess
>> wr
Actually, I've checked, CacheBuilder isn't in that guava.jar.
On Mon, Dec 29, 2014 at 10:38 AM, Left Right wrote:
> Hi Alex,
>
> I've tried to build from develop branch, and I think I could get
> further than before, however, there're problems with dependencies:
>
> [javac] Compiling 1777 sou
Hi Alex,
I've tried to build from develop branch, and I think I could get
further than before, however, there're problems with dependencies:
[javac] Compiling 1777 source files to
/home/wvxvw/projects/flex-falcon/compiler/generated/classes
[javac] warning: [options] bootstrap class path n
46 matches
Mail list logo
