AW: Setting up a CI server for Mustella
Thanks for that tip Clint, Strangely Azure was never on my radar ... now I'm trying to figure out how to setup the Flexmojos CI Agent on that ;-) Cool thing is incoming traffic is free ... this was the part making my EC2 agent rather expensive (Had to download 16 complete Flex FDK versions with every full build). Free incoming traffic sounds great :-) Unfortunately I couldn't find any TeamCity plugin that allows starting and stopping of an Azure VM the way there was for EC2 Build Agents. Couldn't find something like that for Hudson/Jenkins either :-( Fortunately the API for starting and stopping Azure VMs seems to be pretty trivial using these REST services: http://msdn.microsoft.com/en-us/library/windowsazure/jj157206.aspx So I guess I know what I'll be whipping up while the other machine is preparing FM6 for release ;-) Chris -Ursprüngliche Nachricht- Von: omup...@gmail.com [mailto:omup...@gmail.com] Im Auftrag von Om Gesendet: Freitag, 22. Februar 2013 22:49 An: dev@flex.apache.org Betreff: Re: Setting up a CI server for Mustella On Fri, Feb 22, 2013 at 8:40 AM, Clint Modien cmod...@gmail.com wrote: Hmmm. I wonder if that means we can get some Azure services? http://www.windowsazure.com/en-us/pricing/member-offers/msdn-benefits/ That is a fantastic idea, Clint. I will try out this option this weekend! Om On Feb 20, 2013, at 4:51 PM, Justin Mclean jus...@classsoftware.com wrote: Hi, the only licenses I have for Windows is Windows 2008 server most probably Standard, let me know if this is OK because I'll have to install it. As part of a program Microsoft has that gives a MSDN subscription to Apache commiters I have just about any MS licence we may need. Justin
Setting up an Azure instance as TeamCity agent
Ok ... it took me a little while, but I finally managed to have my TeamCity agent for Flexmojos to run on an Azure VM: Documented all things here: https://dev.c-ware.de/confluence/display/PUBLIC/Setting+up+a+TeamCity+agent+on+an+Azure+VM So if you guys want to give it a try, perhaps this will help you. Chris
Re: Linux support for Apache Flex - it works!
Ah now he says so :-) On Feb 23, 2013 5:07 AM, Alex Harui aha...@adobe.com wrote: On 2/22/13 6:02 PM, Justin Mclean jus...@classsoftware.com wrote: Hi, I copied theSDK across to my Linus SDK folder and plugged it into FDT5 (expecting it to fail). Non, c'est parfait. Worked like a dream. So it seems that there is the option to build the SDK in windows and just copy the SDK into Linux and that's that Great even easier than I thought. Next step is to see if it passes all of the mustella tests and I guess package it up in a more friendly way for use by various Linux distributions. Well, that's what a binary distribution is: the results of compilation on mac or win and copied to Linux. That is also what Adobe distributed. -- Alex Harui Flex SDK Team Adobe Systems, Inc. http://blogs.adobe.com/aharui
Re: [FalconJx] writeToken()
Mike, I have moved all emitter strings to enums and added a few overloaded 'write' methods to easily deal with that. I also moved all the 'write' method declarations to a separate interface, which I then implemented in both the regular and 'doc' emitters. I hope this was what you had in mind? EdB On Mon, Feb 18, 2013 at 3:26 PM, Michael Schmalle apa...@teotigraphix.com wrote: Sure, For now we can do that and refactor later(if needed), we just don't want these constants in a concrete class if they are specific to a production type IE goog, amd etc. There is some other things I am thinking about but, I will talk to you later about it. Mike Quoting Erik de Bruin e...@ixsoftware.nl: Would you create a JSTokens, JSAMDTokens and JSGoogTokens for the specific and shared (if any) tokens for these implementations? EdB On Mon, Feb 18, 2013 at 3:18 PM, Michael Schmalle apa...@teotigraphix.com wrote: Just make new strings as we don't want to create dependencies on anything in the compiler package, this is just for emitters. Mike Quoting Erik de Bruin e...@ixsoftware.nl: Mike, Do we introduce new strings, or do we reference existing ones from ASTNodeID, IASKeywordConstants, IASLanguageConstants etc.? EdB On Mon, Feb 18, 2013 at 3:03 PM, Michael Schmalle apa...@teotigraphix.com wrote: Hope everything is OK. Just get better, don't worry about it. :) I changed a couple things with the writeToken(), the main thing is getting ASTokens fully implemented and using that. It feels like it should be an enum but, maybe was can make an enum based off that interface's strings down the road. Mike Quoting Erik de Bruin e...@ixsoftware.nl: Mike, Just returned from the hospital, nothing big, but it is taking a couple of days to get my brain back up to speed after the anesthesia and painkillers. I will take the re-tokenization of the emitters on in the next couple of days. EdB On Fri, Feb 15, 2013 at 11:46 AM, Michael Schmalle apa...@teotigraphix.com wrote: Erik, Just wanted to run this by you but I wanted to do this awhile ago, the writeToken() method's original intention was to pass the token to write and automatically append a space AFTER the token. So writeToken(ASTokens.SET); Will produce; 'set ' This way when ever we have a token, the rule is tokens Always have a space to the right. This will eliminate a hug amount of code and make things more readable. Since we put it in some places, before this change happens either you or I have to do all of it and make sure the tests pass because whitespace will get changed on the writeToken() calls right now. I'm thinking ASTokens should be an enum and we pass the enum to writeToken() for type safety. What do you think? Mike -- Michael Schmalle - Teoti Graphix, LLC http://www.teotigraphix.com http://blog.teotigraphix.com -- Ix Multimedia Software Jan Luykenstraat 27 3521 VB Utrecht T. 06-51952295 I. www.ixsoftware.nl -- Michael Schmalle - Teoti Graphix, LLC http://www.teotigraphix.com http://blog.teotigraphix.com -- Ix Multimedia Software Jan Luykenstraat 27 3521 VB Utrecht T. 06-51952295 I. www.ixsoftware.nl -- Michael Schmalle - Teoti Graphix, LLC http://www.teotigraphix.com http://blog.teotigraphix.com -- Ix Multimedia Software Jan Luykenstraat 27 3521 VB Utrecht T. 06-51952295 I. www.ixsoftware.nl -- Michael Schmalle - Teoti Graphix, LLC http://www.teotigraphix.com http://blog.teotigraphix.com -- Ix Multimedia Software Jan Luykenstraat 27 3521 VB Utrecht T. 06-51952295 I. www.ixsoftware.nl
Re: OneClickMustella on MacOS (was: Re: Apache Flex 4.9.1)
Sorry Erik, for some reason I thought I replied to it. Here is the path for Mustella: https://svn.apache.org/repos/asf/flex/sdk/branches/develop/mustella/utilities/OneClickMustella Let me know how it goes. Thanks, Om On Sat, Feb 23, 2013 at 9:33 AM, Erik de Bruin e...@ixsoftware.nl wrote: bump... EdB On Wed, Feb 20, 2013 at 9:57 AM, Erik de Bruin e...@ixsoftware.nl wrote: Om, Just checking: where is OnClickMustella? Is it the one in sdk/branches/develop/mustella/utilities? There are some 'subtle' differences between the Cygwin and Mac command lines, so you might need to make some UI changes to accomodate these. I'll put in what I can/need to make it run and hand it back over to you. EdB On Fri, Feb 15, 2013 at 8:20 PM, Om bigosma...@gmail.com wrote: On Wed, Feb 13, 2013 at 11:47 PM, Erik de Bruin e...@ixsoftware.nl wrote: Sure, what do you need me to do? EdB Thanks for offering to help! The path to the terminal is currently hardcoded to a windows specific path. If you remove the hardcoding and make it work on Mac as well, that would be great. Let me know if you need more details. Thanks, Om On Wed, Feb 13, 2013 at 7:47 PM, Om bigosma...@gmail.com wrote: On Tue, Feb 12, 2013 at 9:16 PM, Frédéric THOMAS webdoubl...@hotmail.comwrote: Btw, Om, I tried one click mustella and it has been in error as soon I tried to select the sdk path or even tried to select my command line interpreter, didn't had time to tell you before :$ -Message d'origine- From: Justin Mclean Currently downloading, installing and running the tests takes a fair amount of manual setup and effort, have an easier way to run to run tests (one click mustella?) should help. Justin I would like some help with getting OCM to work on a Mac. I have some stuff hardcoded (ex. path to the terminal) which is probably blowing up on a Mac. But these should be very straightforward to change. Can one of you help? Thanks, Om -- Ix Multimedia Software Jan Luykenstraat 27 3521 VB Utrecht T. 06-51952295 I. www.ixsoftware.nl -- Ix Multimedia Software Jan Luykenstraat 27 3521 VB Utrecht T. 06-51952295 I. www.ixsoftware.nl -- Ix Multimedia Software Jan Luykenstraat 27 3521 VB Utrecht T. 06-51952295 I. www.ixsoftware.nl
Re: [FalconJx] writeToken()
Looks fine except it looks like you forgot to have IASEmitter extend IEmitter. I have a bunch of classes that proxy write() methods in sub emitters of a project I am working on that use the interface API. I am going to commit IASEmitter extends IEmitter Mike Quoting Erik de Bruin e...@ixsoftware.nl: Mike, I have moved all emitter strings to enums and added a few overloaded 'write' methods to easily deal with that. I also moved all the 'write' method declarations to a separate interface, which I then implemented in both the regular and 'doc' emitters. I hope this was what you had in mind? EdB On Mon, Feb 18, 2013 at 3:26 PM, Michael Schmalle apa...@teotigraphix.com wrote: Sure, For now we can do that and refactor later(if needed), we just don't want these constants in a concrete class if they are specific to a production type IE goog, amd etc. There is some other things I am thinking about but, I will talk to you later about it. Mike Quoting Erik de Bruin e...@ixsoftware.nl: Would you create a JSTokens, JSAMDTokens and JSGoogTokens for the specific and shared (if any) tokens for these implementations? EdB On Mon, Feb 18, 2013 at 3:18 PM, Michael Schmalle apa...@teotigraphix.com wrote: Just make new strings as we don't want to create dependencies on anything in the compiler package, this is just for emitters. Mike Quoting Erik de Bruin e...@ixsoftware.nl: Mike, Do we introduce new strings, or do we reference existing ones from ASTNodeID, IASKeywordConstants, IASLanguageConstants etc.? EdB On Mon, Feb 18, 2013 at 3:03 PM, Michael Schmalle apa...@teotigraphix.com wrote: Hope everything is OK. Just get better, don't worry about it. :) I changed a couple things with the writeToken(), the main thing is getting ASTokens fully implemented and using that. It feels like it should be an enum but, maybe was can make an enum based off that interface's strings down the road. Mike Quoting Erik de Bruin e...@ixsoftware.nl: Mike, Just returned from the hospital, nothing big, but it is taking a couple of days to get my brain back up to speed after the anesthesia and painkillers. I will take the re-tokenization of the emitters on in the next couple of days. EdB On Fri, Feb 15, 2013 at 11:46 AM, Michael Schmalle apa...@teotigraphix.com wrote: Erik, Just wanted to run this by you but I wanted to do this awhile ago, the writeToken() method's original intention was to pass the token to write and automatically append a space AFTER the token. So writeToken(ASTokens.SET); Will produce; 'set ' This way when ever we have a token, the rule is tokens Always have a space to the right. This will eliminate a hug amount of code and make things more readable. Since we put it in some places, before this change happens either you or I have to do all of it and make sure the tests pass because whitespace will get changed on the writeToken() calls right now. I'm thinking ASTokens should be an enum and we pass the enum to writeToken() for type safety. What do you think? Mike -- Michael Schmalle - Teoti Graphix, LLC http://www.teotigraphix.com http://blog.teotigraphix.com -- Ix Multimedia Software Jan Luykenstraat 27 3521 VB Utrecht T. 06-51952295 I. www.ixsoftware.nl -- Michael Schmalle - Teoti Graphix, LLC http://www.teotigraphix.com http://blog.teotigraphix.com -- Ix Multimedia Software Jan Luykenstraat 27 3521 VB Utrecht T. 06-51952295 I. www.ixsoftware.nl -- Michael Schmalle - Teoti Graphix, LLC http://www.teotigraphix.com http://blog.teotigraphix.com -- Ix Multimedia Software Jan Luykenstraat 27 3521 VB Utrecht T. 06-51952295 I. www.ixsoftware.nl -- Michael Schmalle - Teoti Graphix, LLC http://www.teotigraphix.com http://blog.teotigraphix.com -- Ix Multimedia Software Jan Luykenstraat 27 3521 VB Utrecht T. 06-51952295 I. www.ixsoftware.nl -- Michael Schmalle - Teoti Graphix, LLC http://www.teotigraphix.com http://blog.teotigraphix.com
Re: [FalconJS/Jx] move FlexJS to FalconJx
Falcons already called closure compiler. I just moved where it was called and got it to handle multiple files. We need to get jx to handle mxml in order to switch. Do you have time for that? I want to get basic css working and do some as refactoring first. Sent via the PANTECH Discover, an ATT 4G LTE smartphone. Erik de Bruin e...@ixsoftware.nl wrote: Alex, From your latest commit, it looks like we're doing a lot of work twice. I finished the inclusion of the GCC (incl. deps etc.) into FalconJx a couple of weeks ago. I have some cycles open in the near future. What do you think, time to move FlexJS from FalconJS to FalconJx? EdB -- Ix Multimedia Software Jan Luykenstraat 27 3521 VB Utrecht T. 06-51952295 I. www.ixsoftware.nl
Re: [DISCUSS] Apache Flex 4.9.1 RC3
On Fri, Feb 22, 2013 at 9:22 PM, Alex Harui aha...@adobe.com wrote: On 2/22/13 5:22 PM, Justin Mclean jus...@classsoftware.com wrote: HI, What would it take to get the CI server to generate the RCs? We currently a release build from the develop branch but I think it may be against policy to sign and release them. And what would its path look like? A little ugly and Hudson would get the blame :-) f:\hudson\hudson-slave\workspace\Flex_SDK_build\ We don't have a 4.9.1 build currently on the CI machine, Would there be any legal issues re using hudson in path as Oracle owns it? OK, sounds like we need to build a tool to make the paths pretty. Is it really worth it? What do you want the path to look like? -- Alex Harui Flex SDK Team Adobe Systems, Inc. http://blogs.adobe.com/aharui Why do we need a tool for this? Why cant the release manager just create the release build from a directory like this: c:/apacheflex_4_9_1/ ?
Re: [DISCUSS] Apache Flex 4.9.1 RC3
Mac users shouldn't have to work outside the user folder Sent via the PANTECH Discover, an ATT 4G LTE smartphone. Om bigosma...@gmail.com wrote: On Fri, Feb 22, 2013 at 9:22 PM, Alex Harui aha...@adobe.com wrote: On 2/22/13 5:22 PM, Justin Mclean jus...@classsoftware.com wrote: HI, What would it take to get the CI server to generate the RCs? We currently a release build from the develop branch but I think it may be against policy to sign and release them. And what would its path look like? A little ugly and Hudson would get the blame :-) f:\hudson\hudson-slave\workspace\Flex_SDK_build\ We don't have a 4.9.1 build currently on the CI machine, Would there be any legal issues re using hudson in path as Oracle owns it? OK, sounds like we need to build a tool to make the paths pretty. Is it really worth it? What do you want the path to look like? -- Alex Harui Flex SDK Team Adobe Systems, Inc. http://blogs.adobe.com/aharui Why do we need a tool for this? Why cant the release manager just create the release build from a directory like this: c:/apacheflex_4_9_1/ ?
Re: [DISCUSS] Apache Flex 4.9.1 RC3
Sounds like a dubious reason to me. Writing a script to copy the sdk to a special directory and then building the release artifact would violate this rule? This would be much easier than writing a tool to fix the paths after the build process. Om On Feb 23, 2013 3:29 PM, Alex Harui aha...@adobe.com wrote: Mac users shouldn't have to work outside the user folder Sent via the PANTECH Discover, an ATT 4G LTE smartphone. Om bigosma...@gmail.com wrote: On Fri, Feb 22, 2013 at 9:22 PM, Alex Harui aha...@adobe.com wrote: On 2/22/13 5:22 PM, Justin Mclean jus...@classsoftware.com wrote: HI, What would it take to get the CI server to generate the RCs? We currently a release build from the develop branch but I think it may be against policy to sign and release them. And what would its path look like? A little ugly and Hudson would get the blame :-) f:\hudson\hudson-slave\workspace\Flex_SDK_build\ We don't have a 4.9.1 build currently on the CI machine, Would there be any legal issues re using hudson in path as Oracle owns it? OK, sounds like we need to build a tool to make the paths pretty. Is it really worth it? What do you want the path to look like? -- Alex Harui Flex SDK Team Adobe Systems, Inc. http://blogs.adobe.com/aharui Why do we need a tool for this? Why cant the release manager just create the release build from a directory like this: c:/apacheflex_4_9_1/ ?
Re: [DISCUSS] Apache Flex 4.9.1 RC3
Hi, Sounds like a dubious reason to me. Writing a script to copy the sdk to a special directory and then building the release artifact would violate this rule? Well there's no way we going to make it match on windows and OSX. Thinking about again I actually like the release managers name in there and added incentive to get it right and their 15 minutes of fame (in error messages):-) I've received a couple emails asking me what did I break in 4.9 cos my names on it :-) I think best solution is if we could change it to be in thr form flex.apage.com/... but no idea if this is actually possible or would work. Justin
Re: Donation of TLF word processor
Do you have something other than a video that we can read to learn more about this code? Nothing else unless I build or write something for you. I'm juggling my time between a few things right now - what do you require? It's pure actionscript, and the UI is completely custom built. Do you actually own the code or was this done for a customer or employer? This was a self-indulgent and time-consuming unfunded project. (Like MadComponents is!). Did anyone other than you contribute code? No. Any other third-party components? It uses Adobe's TLF, the Squiggly spell check, and as3corelib.swc What is the current license for the code? I've never released the source to anyone - so no license right now. On Sat, Feb 23, 2013 at 3:31 PM, Alex Harui aha...@adobe.com wrote: On 2/22/13 8:54 PM, Daniel Freeman madcompone...@gmail.com wrote: I've started a new thread for this, as the discussion started off-topic. I agreed to donate my TLF-based word processor to this group. @Harbs said he'd be interested to help getting it ready for donation - and I wondered if anyone else would be interested in helping to get it ready? It was quite a sophisticated and self-indugent project. Text Wrapping wasn't its only novel feature. I've decided to bundle it with e2vector, as they share many classes, and the two apps work together. http://e2easy.wordpress.com/2010/04/28/e2publish-video-tutorials/ http://www.youtube.com/watch?v=VvakUrX9Hvc What licensing do I use to hand it over? MadComponents is MIT licensed - do I use the same? Any particular text comment to put at the top of the source code? Do you have something other than a video that we can read to learn more about this code? Do you actually own the code or was this done for a customer or employer? Did anyone other than you contribute code? Any other third-party components? What is the current license for the code? -- Alex Harui Flex SDK Team Adobe Systems, Inc. http://blogs.adobe.com/aharui
Re: [DISCUSS] Apache Flex 4.9.1 RC3
On Feb 23, 2013 4:53 PM, Justin Mclean jus...@classsoftware.com wrote: Hi, Sounds like a dubious reason to me. Writing a script to copy the sdk to a special directory and then building the release artifact would violate this rule? Well there's no way we going to make it match on windows and OSX. I don't think anyone is saying that the paths have to match. Thinking about again I actually like the release managers name in there and added incentive to get it right and their 15 minutes of fame (in error messages):-) I've received a couple emails asking me what did I break in 4.9 cos my names on it :-) I think best solution is if we could change it to be in thr form flex.apage.com/... but no idea if this is actually possible or would work. Justin I am indifferent to what exactly we call it as long as it sounds official and does not have weird names in the path. Om
[jira] [Created] (FLEX-33412) NativeDragManagerImpl prevents objects from being garbage collected.
Michael Rochelle created FLEX-33412: --- Summary: NativeDragManagerImpl prevents objects from being garbage collected. Key: FLEX-33412 URL: https://issues.apache.org/jira/browse/FLEX-33412 Project: Apache Flex Issue Type: Bug Components: Drag and Drop Affects Versions: Apache Flex 4.9.0 Reporter: Michael Rochelle Priority: Minor The mx.managers.NativeDragManagerImpl class adds a reference to the target in the private _relatedObject property during a dragDrop event. It appears that it does this so that it can construct a dragComplete event with this property. During dragComplete after the _relatedObject is passed to the event, the _relatedObject is never set to null. This causes the view to remain in memory until another native drag occurs to push it out with another object. This may never occur preventing the views from being garbage collected. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
Re: Donation of TLF word processor
On 2/23/13 5:08 PM, Daniel Freeman madcompone...@gmail.com wrote: Do you have something other than a video that we can read to learn more about this code? Nothing else unless I build or write something for you. I'm juggling my time between a few things right now - what do you require? It's pure actionscript, and the UI is completely custom built. I see, so is it Flex-compatible? Any other third-party components? It uses Adobe's TLF, the Squiggly spell check, and as3corelib.swc Personally, I'm not in a hurry to execute on a donation of a word processor and I think the donation would be blocked until Squiggly gets donated, so it might be best to put this on the table for a bit if you are short on time. I want to hear from more folks to judge the priority of this donation, but thanks for your generous offer. -- Alex Harui Flex SDK Team Adobe Systems, Inc. http://blogs.adobe.com/aharui
Re: Donation of TLF word processor
Well, the way I see it is that there's the question of the project as a whole, and there's the question of pieces of it. Some of the functionality (such as text wrap) is something that has been requested quite a bit in the past. Maybe pieces of it can be added to the Flex components independent of the whole project. Tables is something that is missing in Flex as well. If that can be incorporated into Flex components, that should be very interesting too. I actually did not realize that the project was so well developed. (Very impressive!) I wonder what could be done with the project as a whole. It looks quite interesting. On Feb 24, 2013, at 8:36 AM, Alex Harui wrote: On 2/23/13 5:08 PM, Daniel Freeman madcompone...@gmail.com wrote: Do you have something other than a video that we can read to learn more about this code? Nothing else unless I build or write something for you. I'm juggling my time between a few things right now - what do you require? It's pure actionscript, and the UI is completely custom built. I see, so is it Flex-compatible? Any other third-party components? It uses Adobe's TLF, the Squiggly spell check, and as3corelib.swc Personally, I'm not in a hurry to execute on a donation of a word processor and I think the donation would be blocked until Squiggly gets donated, so it might be best to put this on the table for a bit if you are short on time. I want to hear from more folks to judge the priority of this donation, but thanks for your generous offer. -- Alex Harui Flex SDK Team Adobe Systems, Inc. http://blogs.adobe.com/aharui