Re: serving extra css

2014-02-04 Thread Michal Mocny
On Mon, Feb 3, 2014 at 6:58 PM, Josh Soref jso...@blackberry.com wrote:

 Michal wrote:
  Wouldn't this be better handled by Brian's dream of
 browser-as-a-platform?

 I haven¹t worked through all the details, but I don¹t think so. If you
 want to run the blackberry specific hooks and you browse to the app from
 your blackberry browser, then you¹re going to be unhappy if it¹s a
 different platform.


If you want BB-specific hooks with BB-browser, then target BB.

If you want to test your app using cordova plugins on a desktop browser
simulating a mobile phone, browser-as-a-platform is the I think the way to
go.

If you are done building your app on desktop, and now want to test more
device specific styles (such as fonts), I would say that you should test on
device.  Is this prohibitively difficult on BB (I have not tried so not
sure what the turnaround time is)?  Perhaps the app harness + some of the
auto-reload-on-device work that Braden is doing will help?



 Browser-as-an-engine might work, maybe.


I don't think so -- engine is the concept of replacing the use of the
system webview, but thats still running on-device.



  I understand that this solves your immediate problem, but I'm not sure
  extending serve is really the right long term solution for this.


  ..also, your idea of automatically look for a well known file could be
  done within the application by feature detecting that its running in a
  desktop browser.

 We¹re the SDK, not the application. I could tell authors ³if you want to
 see the fonts, put in this magic², but I¹m hoping to avoid making each
 author do that (where possible). There¹s also licensing fun w/ the fonts.

 I suppose that I could insert some code into cordova.js instead of having
 serve manage the css.

  Would it be prehibitive to dynamically inject the style
  on startup from the app itself?

 Doing it with a link: header is a lot cleaner than inserting strange logic
 into cordova.js, butŠ


I didn't mean cordova.js -- I meant the specific app you are using for
testing.  You can ship cordova-cli hooks to add this if you want to share
the functionality between several apps.

I'm asking: do we really need to change the tools?



 We can and probably will provide templates which include a style saying
 ³use font: Slate Pro². That¹s all you need for the platform engine.

 If we did use cordova.js instead of having serve provide the .css file,
 we¹d still need a way to get the fonts to the browser-as-engine case
 (which is handled by the .extra code).

 In case you¹re curious, installing fonts to a user¹s system is a no-go for
 a number of reasons (* it¹s painful to try to manage ‹ often requiring
 administrative privileges which isn¹t something we require, * Licensing is
 wrong, * it pollutes the user¹s font-space or their computer¹s font-space
 if shared, * it doesn¹t work if the user is visiting serve from a device
 other than localhost).




Re: Make cordova platform add ios work on Windows

2014-02-04 Thread Michal Mocny
With the app-harness, you actually can test on iOS from windows, so this is
becoming a use case.  So long as the pre-packaged native plugins running on
the app-harness match version requirements for your app, you won't need to
build the native bits locally.

All we need is to serve web assets from your development machine to the AH,
and the easiest way to make sure we serve the right assets is to use the
platform targeted prepare step.  If for no other reason, than just to make
sure we serve the right cordova_plugins.js (though Braden will know what
other platform bits the AH consumes, at some point it was using the app's
cordova.js and plugins' js-modules, but I think it changed to used the AH
bundled versions).

-Michal


On Mon, Feb 3, 2014 at 12:50 PM, Jesse purplecabb...@gmail.com wrote:

 Personally, I would not expect a multi-platform team to even checkin the
 platforms folder.

 Addressing Axel's use case: It would only require a few commands to be
 ready to build an iOS Cordova project:
 npm install -g cordova
 cordova create ~/HelloCordova
 cd ~/HelloCordova
 cordova platform add ios

 As an alternative, the phonegap-cli [1] allows actual building of iOS
 targets from Windows, by using the phonegap-build service. Or you could
 just use the PhoneGap build site directly. You would need to setup an
 account either way,

 My reason for creating the issue CB-5963 was NOT to enable the
 functionality Baz referred to with CB-5706, but to provide an easy way to
 test plugins in WP7/8 and Windows8.  I will/am changing the functionality a
 bit in the nodejs scripts because it doesn't make sense to platform add ios
 on windows because you won't be able to run the tests anyway.

 My plan is that the node script for testing mobile spec will only platform
 add whatever platform is specified in the arguments.
 ie.
 node createmobilespec.js --platform wp8
 and possibly in the future, allow the testing of specific plugins, from
 specific locations, using a specific cordova, and a specific platform
 location.

 This really is all I have to say on the CB-5706 subject, if you still want
 to do it, then go ahead. However, if it is a significant change, and I
 think I'll have to support the added code in wp8/windows8, then I will be
 very strict in accepting pull requests as I work on these projects every
 day. You may want to choose a less controversial issue to work on.

 [1] https://github.com/phonegap/phonegap-cli
 [2] https://build.phonegap.com/


 @purplecabbage
 risingj.com


 On Mon, Feb 3, 2014 at 8:09 AM, Andrew Grieve agri...@chromium.org
 wrote:

  Converting the scripts to node is exactly what we'd need to do to have
 them
  work on Win, I think. We can just remove the call to check_reqs made from
  create().
 
  A use-case I can see for this is PhoneGap Build / AppHarness development.
 
  If you'd like to work on this issue, then I think that's great!
 
 
  On Sun, Feb 2, 2014 at 9:20 AM, Bas Bosman m...@nazgul.nu wrote:
 
   I do think it's beneficial to have that option available on Windows.
  
   As far as i know Cordova strives for feature parity across all
 deployment
   platforms (where possible) and I don't think it's strange to extend
 that
   feature parity to the development platform (where possible). I agree
 that
   you need XCode to build, test and deploy an iOS app, but you don't need
  it
   for large parts of developing the app. For that same reason I think
 that
   the ability to do a cordova add wp8 (not build) on a Mac for example
  would
   be a good thing to have as well.
  
   The 'platform add' code currently doesn't actually use any of the
 things
   checked for in check_req as far as I can tell, so there is no technical
   reason not to do it. And we do allow a 'plugin add' to add iOS specific
   stuff on a Windows machine.
  
   If a group of people start developing a multi platform app using
 Cordova
  a
   valid use case in my opinion would be that there is one developer who
  does
   the initial setup. So adding the base platforms and plugins for the
 team
   to use. Currently he has to use multiple development platforms, which
 he
   may not have, to do so while there is a good chance he'll only be
 working
   on the on Mac or Windows only development parts of the app after that
   initial setup. Giving him the opportunity to do that all from the
   environment he's familiar with seems like a nice thing to have.
  
   Another reason I started looking into this is mobile-spec. There is a
  JIRA
   issue (CB-5963, submitted by a core contributor) about
  createmobilespec.sh
   only working on a Mac. That script wouldn't be to hard to port over to
   JavaScript to have it run on other platforms as well, but it also has a
   dependency on the platform add ios stuff. I think that having something
   there that works on all platforms and produces the same end result
 would
   be a good thing as well.
  
   And I agree with Axel, although I hadn't thought of it before, that
  

Re: Make cordova platform add ios work on Windows

2014-02-04 Thread Michal Mocny
..matter of fact, for those same reasons we should change the other
platforms to support installation when check_reqs fail.  I should be able
to test Android using AH the exact same way without a full dev environment.


On Tue, Feb 4, 2014 at 9:27 AM, Michal Mocny mmo...@chromium.org wrote:

 With the app-harness, you actually can test on iOS from windows, so this
 is becoming a use case.  So long as the pre-packaged native plugins running
 on the app-harness match version requirements for your app, you won't need
 to build the native bits locally.

 All we need is to serve web assets from your development machine to the
 AH, and the easiest way to make sure we serve the right assets is to use
 the platform targeted prepare step.  If for no other reason, than just to
 make sure we serve the right cordova_plugins.js (though Braden will know
 what other platform bits the AH consumes, at some point it was using the
 app's cordova.js and plugins' js-modules, but I think it changed to used
 the AH bundled versions).

 -Michal


 On Mon, Feb 3, 2014 at 12:50 PM, Jesse purplecabb...@gmail.com wrote:

 Personally, I would not expect a multi-platform team to even checkin the
 platforms folder.

 Addressing Axel's use case: It would only require a few commands to be
 ready to build an iOS Cordova project:
 npm install -g cordova
 cordova create ~/HelloCordova
 cd ~/HelloCordova
 cordova platform add ios

 As an alternative, the phonegap-cli [1] allows actual building of iOS
 targets from Windows, by using the phonegap-build service. Or you could
 just use the PhoneGap build site directly. You would need to setup an
 account either way,

 My reason for creating the issue CB-5963 was NOT to enable the
 functionality Baz referred to with CB-5706, but to provide an easy way to
 test plugins in WP7/8 and Windows8.  I will/am changing the functionality
 a
 bit in the nodejs scripts because it doesn't make sense to platform add
 ios
 on windows because you won't be able to run the tests anyway.

 My plan is that the node script for testing mobile spec will only platform
 add whatever platform is specified in the arguments.
 ie.
 node createmobilespec.js --platform wp8
 and possibly in the future, allow the testing of specific plugins, from
 specific locations, using a specific cordova, and a specific platform
 location.

 This really is all I have to say on the CB-5706 subject, if you still want
 to do it, then go ahead. However, if it is a significant change, and I
 think I'll have to support the added code in wp8/windows8, then I will be
 very strict in accepting pull requests as I work on these projects every
 day. You may want to choose a less controversial issue to work on.

 [1] https://github.com/phonegap/phonegap-cli
 [2] https://build.phonegap.com/


 @purplecabbage
 risingj.com


 On Mon, Feb 3, 2014 at 8:09 AM, Andrew Grieve agri...@chromium.org
 wrote:

  Converting the scripts to node is exactly what we'd need to do to have
 them
  work on Win, I think. We can just remove the call to check_reqs made
 from
  create().
 
  A use-case I can see for this is PhoneGap Build / AppHarness
 development.
 
  If you'd like to work on this issue, then I think that's great!
 
 
  On Sun, Feb 2, 2014 at 9:20 AM, Bas Bosman m...@nazgul.nu wrote:
 
   I do think it's beneficial to have that option available on Windows.
  
   As far as i know Cordova strives for feature parity across all
 deployment
   platforms (where possible) and I don't think it's strange to extend
 that
   feature parity to the development platform (where possible). I agree
 that
   you need XCode to build, test and deploy an iOS app, but you don't
 need
  it
   for large parts of developing the app. For that same reason I think
 that
   the ability to do a cordova add wp8 (not build) on a Mac for example
  would
   be a good thing to have as well.
  
   The 'platform add' code currently doesn't actually use any of the
 things
   checked for in check_req as far as I can tell, so there is no
 technical
   reason not to do it. And we do allow a 'plugin add' to add iOS
 specific
   stuff on a Windows machine.
  
   If a group of people start developing a multi platform app using
 Cordova
  a
   valid use case in my opinion would be that there is one developer who
  does
   the initial setup. So adding the base platforms and plugins for the
 team
   to use. Currently he has to use multiple development platforms, which
 he
   may not have, to do so while there is a good chance he'll only be
 working
   on the on Mac or Windows only development parts of the app after that
   initial setup. Giving him the opportunity to do that all from the
   environment he's familiar with seems like a nice thing to have.
  
   Another reason I started looking into this is mobile-spec. There is a
  JIRA
   issue (CB-5963, submitted by a core contributor) about
  createmobilespec.sh
   only working on a Mac. That script wouldn't be to hard to port over to
   JavaScript to have 

Re: Make cordova platform add ios work on Windows

2014-02-04 Thread Ian Clelland
In that case, I feel like we should be calling check_reqs on create, but
not throwing an error; just display a warning if the build tools aren't
installed.

We've had reports from people using the cca toolchain that `create` lets
them start the project, but then they can't build because they're missing
javac or something else essential. I think there's an expectation that we
at least let people know early if they're going to have a bad time later on.


On Tue, Feb 4, 2014 at 9:30 AM, Michal Mocny mmo...@chromium.org wrote:

 ..matter of fact, for those same reasons we should change the other
 platforms to support installation when check_reqs fail.  I should be able
 to test Android using AH the exact same way without a full dev environment.


 On Tue, Feb 4, 2014 at 9:27 AM, Michal Mocny mmo...@chromium.org wrote:

  With the app-harness, you actually can test on iOS from windows, so this
  is becoming a use case.  So long as the pre-packaged native plugins
 running
  on the app-harness match version requirements for your app, you won't
 need
  to build the native bits locally.
 
  All we need is to serve web assets from your development machine to the
  AH, and the easiest way to make sure we serve the right assets is to use
  the platform targeted prepare step.  If for no other reason, than just to
  make sure we serve the right cordova_plugins.js (though Braden will know
  what other platform bits the AH consumes, at some point it was using the
  app's cordova.js and plugins' js-modules, but I think it changed to used
  the AH bundled versions).
 
  -Michal
 
 
  On Mon, Feb 3, 2014 at 12:50 PM, Jesse purplecabb...@gmail.com wrote:
 
  Personally, I would not expect a multi-platform team to even checkin the
  platforms folder.
 
  Addressing Axel's use case: It would only require a few commands to be
  ready to build an iOS Cordova project:
  npm install -g cordova
  cordova create ~/HelloCordova
  cd ~/HelloCordova
  cordova platform add ios
 
  As an alternative, the phonegap-cli [1] allows actual building of iOS
  targets from Windows, by using the phonegap-build service. Or you could
  just use the PhoneGap build site directly. You would need to setup an
  account either way,
 
  My reason for creating the issue CB-5963 was NOT to enable the
  functionality Baz referred to with CB-5706, but to provide an easy way
 to
  test plugins in WP7/8 and Windows8.  I will/am changing the
 functionality
  a
  bit in the nodejs scripts because it doesn't make sense to platform add
  ios
  on windows because you won't be able to run the tests anyway.
 
  My plan is that the node script for testing mobile spec will only
 platform
  add whatever platform is specified in the arguments.
  ie.
  node createmobilespec.js --platform wp8
  and possibly in the future, allow the testing of specific plugins, from
  specific locations, using a specific cordova, and a specific platform
  location.
 
  This really is all I have to say on the CB-5706 subject, if you still
 want
  to do it, then go ahead. However, if it is a significant change, and I
  think I'll have to support the added code in wp8/windows8, then I will
 be
  very strict in accepting pull requests as I work on these projects every
  day. You may want to choose a less controversial issue to work on.
 
  [1] https://github.com/phonegap/phonegap-cli
  [2] https://build.phonegap.com/
 
 
  @purplecabbage
  risingj.com
 
 
  On Mon, Feb 3, 2014 at 8:09 AM, Andrew Grieve agri...@chromium.org
  wrote:
 
   Converting the scripts to node is exactly what we'd need to do to have
  them
   work on Win, I think. We can just remove the call to check_reqs made
  from
   create().
  
   A use-case I can see for this is PhoneGap Build / AppHarness
  development.
  
   If you'd like to work on this issue, then I think that's great!
  
  
   On Sun, Feb 2, 2014 at 9:20 AM, Bas Bosman m...@nazgul.nu wrote:
  
I do think it's beneficial to have that option available on Windows.
   
As far as i know Cordova strives for feature parity across all
  deployment
platforms (where possible) and I don't think it's strange to extend
  that
feature parity to the development platform (where possible). I agree
  that
you need XCode to build, test and deploy an iOS app, but you don't
  need
   it
for large parts of developing the app. For that same reason I think
  that
the ability to do a cordova add wp8 (not build) on a Mac for example
   would
be a good thing to have as well.
   
The 'platform add' code currently doesn't actually use any of the
  things
checked for in check_req as far as I can tell, so there is no
  technical
reason not to do it. And we do allow a 'plugin add' to add iOS
  specific
stuff on a Windows machine.
   
If a group of people start developing a multi platform app using
  Cordova
   a
valid use case in my opinion would be that there is one developer
 who
   does
the initial setup. So adding the base 

Re: Make cordova platform add ios work on Windows

2014-02-04 Thread Michal Mocny
I think we've previously also thrown around the idea of Installed vs
Enabled (though the context I think was about plugins and being
conditionally disabled on various platforms).

Perhaps we could always install the platforms you ask for, but leave then
disabled for local builds until check_req's pass?


On Tue, Feb 4, 2014 at 9:37 AM, Ian Clelland iclell...@chromium.org wrote:

 In that case, I feel like we should be calling check_reqs on create, but
 not throwing an error; just display a warning if the build tools aren't
 installed.

 We've had reports from people using the cca toolchain that `create` lets
 them start the project, but then they can't build because they're missing
 javac or something else essential. I think there's an expectation that we
 at least let people know early if they're going to have a bad time later on.


 On Tue, Feb 4, 2014 at 9:30 AM, Michal Mocny mmo...@chromium.org wrote:

 ..matter of fact, for those same reasons we should change the other
 platforms to support installation when check_reqs fail.  I should be able
 to test Android using AH the exact same way without a full dev
 environment.


 On Tue, Feb 4, 2014 at 9:27 AM, Michal Mocny mmo...@chromium.org wrote:

  With the app-harness, you actually can test on iOS from windows, so this
  is becoming a use case.  So long as the pre-packaged native plugins
 running
  on the app-harness match version requirements for your app, you won't
 need
  to build the native bits locally.
 
  All we need is to serve web assets from your development machine to the
  AH, and the easiest way to make sure we serve the right assets is to use
  the platform targeted prepare step.  If for no other reason, than just
 to
  make sure we serve the right cordova_plugins.js (though Braden will know
  what other platform bits the AH consumes, at some point it was using the
  app's cordova.js and plugins' js-modules, but I think it changed to used
  the AH bundled versions).
 
  -Michal
 
 
  On Mon, Feb 3, 2014 at 12:50 PM, Jesse purplecabb...@gmail.com wrote:
 
  Personally, I would not expect a multi-platform team to even checkin
 the
  platforms folder.
 
  Addressing Axel's use case: It would only require a few commands to be
  ready to build an iOS Cordova project:
  npm install -g cordova
  cordova create ~/HelloCordova
  cd ~/HelloCordova
  cordova platform add ios
 
  As an alternative, the phonegap-cli [1] allows actual building of iOS
  targets from Windows, by using the phonegap-build service. Or you could
  just use the PhoneGap build site directly. You would need to setup an
  account either way,
 
  My reason for creating the issue CB-5963 was NOT to enable the
  functionality Baz referred to with CB-5706, but to provide an easy way
 to
  test plugins in WP7/8 and Windows8.  I will/am changing the
 functionality
  a
  bit in the nodejs scripts because it doesn't make sense to platform add
  ios
  on windows because you won't be able to run the tests anyway.
 
  My plan is that the node script for testing mobile spec will only
 platform
  add whatever platform is specified in the arguments.
  ie.
  node createmobilespec.js --platform wp8
  and possibly in the future, allow the testing of specific plugins, from
  specific locations, using a specific cordova, and a specific platform
  location.
 
  This really is all I have to say on the CB-5706 subject, if you still
 want
  to do it, then go ahead. However, if it is a significant change, and I
  think I'll have to support the added code in wp8/windows8, then I will
 be
  very strict in accepting pull requests as I work on these projects
 every
  day. You may want to choose a less controversial issue to work on.
 
  [1] https://github.com/phonegap/phonegap-cli
  [2] https://build.phonegap.com/
 
 
  @purplecabbage
  risingj.com
 
 
  On Mon, Feb 3, 2014 at 8:09 AM, Andrew Grieve agri...@chromium.org
  wrote:
 
   Converting the scripts to node is exactly what we'd need to do to
 have
  them
   work on Win, I think. We can just remove the call to check_reqs made
  from
   create().
  
   A use-case I can see for this is PhoneGap Build / AppHarness
  development.
  
   If you'd like to work on this issue, then I think that's great!
  
  
   On Sun, Feb 2, 2014 at 9:20 AM, Bas Bosman m...@nazgul.nu wrote:
  
I do think it's beneficial to have that option available on
 Windows.
   
As far as i know Cordova strives for feature parity across all
  deployment
platforms (where possible) and I don't think it's strange to extend
  that
feature parity to the development platform (where possible). I
 agree
  that
you need XCode to build, test and deploy an iOS app, but you don't
  need
   it
for large parts of developing the app. For that same reason I think
  that
the ability to do a cordova add wp8 (not build) on a Mac for
 example
   would
be a good thing to have as well.
   
The 'platform add' code currently doesn't actually use any of the
  things

Re: New release of the cordova-plugin-file

2014-02-04 Thread Ian Clelland
File still has a couple of rough edges, but it's been getting use in the
wild recently. Exposure to a wider audience would probably help us work out
the remainder of the issues.

Ian


On Mon, Feb 3, 2014 at 11:11 AM, Andrew Grieve agri...@chromium.org wrote:

 Ian's off today, but I'm definitely in support of doing a plugins release
 this week with it included.


 On Mon, Feb 3, 2014 at 9:41 AM, Maxim Ermilov
 maxim.ermi...@canonical.comwrote:

  Hey guys,
 
  I noticed the cordova-plugin-file was released 3 months ago.
  Could we plan for a new release soon? This would be to enable Ubuntu
  support, which has been merged for a bit.
 
  _
  Best regards,
  Maxim Ermilov
 



Re: Make cordova platform add ios work on Windows

2014-02-04 Thread Braden Shepherdson
For the record, the App Harness makes child apps use the cordova.js,
cordova_plugins.js, and plugins/* which are included with the App Harness,
not the ones that come with the child app. This is because it ensures
compatibility with the native code and the Javascript side, though it can
break the JS API for a plugin if there's a version mismatch (which the App
Harness can warn about). Generally these APIs are more stable than the
bridge calls the plugin is making under the hood, hence the decision to use
the bundled JS code.

Braden


On Tue, Feb 4, 2014 at 9:44 AM, Michal Mocny mmo...@chromium.org wrote:

 I think we've previously also thrown around the idea of Installed vs
 Enabled (though the context I think was about plugins and being
 conditionally disabled on various platforms).

 Perhaps we could always install the platforms you ask for, but leave then
 disabled for local builds until check_req's pass?


 On Tue, Feb 4, 2014 at 9:37 AM, Ian Clelland iclell...@chromium.org
 wrote:

  In that case, I feel like we should be calling check_reqs on create, but
  not throwing an error; just display a warning if the build tools aren't
  installed.
 
  We've had reports from people using the cca toolchain that `create` lets
  them start the project, but then they can't build because they're missing
  javac or something else essential. I think there's an expectation that we
  at least let people know early if they're going to have a bad time later
 on.
 
 
  On Tue, Feb 4, 2014 at 9:30 AM, Michal Mocny mmo...@chromium.org
 wrote:
 
  ..matter of fact, for those same reasons we should change the other
  platforms to support installation when check_reqs fail.  I should be
 able
  to test Android using AH the exact same way without a full dev
  environment.
 
 
  On Tue, Feb 4, 2014 at 9:27 AM, Michal Mocny mmo...@chromium.org
 wrote:
 
   With the app-harness, you actually can test on iOS from windows, so
 this
   is becoming a use case.  So long as the pre-packaged native plugins
  running
   on the app-harness match version requirements for your app, you won't
  need
   to build the native bits locally.
  
   All we need is to serve web assets from your development machine to
 the
   AH, and the easiest way to make sure we serve the right assets is to
 use
   the platform targeted prepare step.  If for no other reason, than just
  to
   make sure we serve the right cordova_plugins.js (though Braden will
 know
   what other platform bits the AH consumes, at some point it was using
 the
   app's cordova.js and plugins' js-modules, but I think it changed to
 used
   the AH bundled versions).
  
   -Michal
  
  
   On Mon, Feb 3, 2014 at 12:50 PM, Jesse purplecabb...@gmail.com
 wrote:
  
   Personally, I would not expect a multi-platform team to even checkin
  the
   platforms folder.
  
   Addressing Axel's use case: It would only require a few commands to
 be
   ready to build an iOS Cordova project:
   npm install -g cordova
   cordova create ~/HelloCordova
   cd ~/HelloCordova
   cordova platform add ios
  
   As an alternative, the phonegap-cli [1] allows actual building of iOS
   targets from Windows, by using the phonegap-build service. Or you
 could
   just use the PhoneGap build site directly. You would need to setup an
   account either way,
  
   My reason for creating the issue CB-5963 was NOT to enable the
   functionality Baz referred to with CB-5706, but to provide an easy
 way
  to
   test plugins in WP7/8 and Windows8.  I will/am changing the
  functionality
   a
   bit in the nodejs scripts because it doesn't make sense to platform
 add
   ios
   on windows because you won't be able to run the tests anyway.
  
   My plan is that the node script for testing mobile spec will only
  platform
   add whatever platform is specified in the arguments.
   ie.
   node createmobilespec.js --platform wp8
   and possibly in the future, allow the testing of specific plugins,
 from
   specific locations, using a specific cordova, and a specific platform
   location.
  
   This really is all I have to say on the CB-5706 subject, if you still
  want
   to do it, then go ahead. However, if it is a significant change, and
 I
   think I'll have to support the added code in wp8/windows8, then I
 will
  be
   very strict in accepting pull requests as I work on these projects
  every
   day. You may want to choose a less controversial issue to work on.
  
   [1] https://github.com/phonegap/phonegap-cli
   [2] https://build.phonegap.com/
  
  
   @purplecabbage
   risingj.com
  
  
   On Mon, Feb 3, 2014 at 8:09 AM, Andrew Grieve agri...@chromium.org
   wrote:
  
Converting the scripts to node is exactly what we'd need to do to
  have
   them
work on Win, I think. We can just remove the call to check_reqs
 made
   from
create().
   
A use-case I can see for this is PhoneGap Build / AppHarness
   development.
   
If you'd like to work on this issue, then I think that's great!
   
   
On Sun, Feb 

Re: Make cordova platform add ios work on Windows

2014-02-04 Thread Michal Mocny
It doesn't use the app's cordova_plugins.js?  I thought we wanted to verify
that the right set of plugins are installed and to do that we need to load
the apps' version.  If we are doing that for now thats fine, but I'm
curious if that is the long term plan?

Are we loading *any* assets outside of whats available in the top level
www/?  Once upon a time serve served up the top level www/, then changed to
platform www/.. should we change it back?

I'm with Jesse that devs should not need (but can if they want) to version
control platforms/ or plugins/, and his workaround seems quite agreeable to
me.

-Michal


On Tue, Feb 4, 2014 at 11:23 AM, Braden Shepherdson bra...@chromium.orgwrote:

 For the record, the App Harness makes child apps use the cordova.js,
 cordova_plugins.js, and plugins/* which are included with the App Harness,
 not the ones that come with the child app. This is because it ensures
 compatibility with the native code and the Javascript side, though it can
 break the JS API for a plugin if there's a version mismatch (which the App
 Harness can warn about). Generally these APIs are more stable than the
 bridge calls the plugin is making under the hood, hence the decision to use
 the bundled JS code.

 Braden



 On Tue, Feb 4, 2014 at 9:44 AM, Michal Mocny mmo...@chromium.org wrote:

 I think we've previously also thrown around the idea of Installed vs
 Enabled (though the context I think was about plugins and being
 conditionally disabled on various platforms).

 Perhaps we could always install the platforms you ask for, but leave then
 disabled for local builds until check_req's pass?


 On Tue, Feb 4, 2014 at 9:37 AM, Ian Clelland iclell...@chromium.org
 wrote:

  In that case, I feel like we should be calling check_reqs on create, but
  not throwing an error; just display a warning if the build tools aren't
  installed.
 
  We've had reports from people using the cca toolchain that `create` lets
  them start the project, but then they can't build because they're
 missing
  javac or something else essential. I think there's an expectation that
 we
  at least let people know early if they're going to have a bad time
 later on.
 
 
  On Tue, Feb 4, 2014 at 9:30 AM, Michal Mocny mmo...@chromium.org
 wrote:
 
  ..matter of fact, for those same reasons we should change the other
  platforms to support installation when check_reqs fail.  I should be
 able
  to test Android using AH the exact same way without a full dev
  environment.
 
 
  On Tue, Feb 4, 2014 at 9:27 AM, Michal Mocny mmo...@chromium.org
 wrote:
 
   With the app-harness, you actually can test on iOS from windows, so
 this
   is becoming a use case.  So long as the pre-packaged native plugins
  running
   on the app-harness match version requirements for your app, you won't
  need
   to build the native bits locally.
  
   All we need is to serve web assets from your development machine to
 the
   AH, and the easiest way to make sure we serve the right assets is to
 use
   the platform targeted prepare step.  If for no other reason, than
 just
  to
   make sure we serve the right cordova_plugins.js (though Braden will
 know
   what other platform bits the AH consumes, at some point it was using
 the
   app's cordova.js and plugins' js-modules, but I think it changed to
 used
   the AH bundled versions).
  
   -Michal
  
  
   On Mon, Feb 3, 2014 at 12:50 PM, Jesse purplecabb...@gmail.com
 wrote:
  
   Personally, I would not expect a multi-platform team to even checkin
  the
   platforms folder.
  
   Addressing Axel's use case: It would only require a few commands to
 be
   ready to build an iOS Cordova project:
   npm install -g cordova
   cordova create ~/HelloCordova
   cd ~/HelloCordova
   cordova platform add ios
  
   As an alternative, the phonegap-cli [1] allows actual building of
 iOS
   targets from Windows, by using the phonegap-build service. Or you
 could
   just use the PhoneGap build site directly. You would need to setup
 an
   account either way,
  
   My reason for creating the issue CB-5963 was NOT to enable the
   functionality Baz referred to with CB-5706, but to provide an easy
 way
  to
   test plugins in WP7/8 and Windows8.  I will/am changing the
  functionality
   a
   bit in the nodejs scripts because it doesn't make sense to platform
 add
   ios
   on windows because you won't be able to run the tests anyway.
  
   My plan is that the node script for testing mobile spec will only
  platform
   add whatever platform is specified in the arguments.
   ie.
   node createmobilespec.js --platform wp8
   and possibly in the future, allow the testing of specific plugins,
 from
   specific locations, using a specific cordova, and a specific
 platform
   location.
  
   This really is all I have to say on the CB-5706 subject, if you
 still
  want
   to do it, then go ahead. However, if it is a significant change,
 and I
   think I'll have to support the added code in wp8/windows8, then I
 will
  be
   very 

Re: Make cordova platform add ios work on Windows

2014-02-04 Thread Braden Shepherdson
We use the App Harness' versions of cordova.js, cordova_plugins.js, and
plugins/*, as I stated. We *fetch* the child app's cordova_plugins.js in
order to read the plugin version data and compare with the versions
installed in the App Harness, but when you launch the app it uses the App
Harness' version, because we want to ensure compatibility at the more
fragile and dangerous of the two layers: exec() calls.

Braden


On Tue, Feb 4, 2014 at 11:53 AM, Michal Mocny mmo...@chromium.org wrote:

 It doesn't use the app's cordova_plugins.js?  I thought we wanted to
 verify that the right set of plugins are installed and to do that we need
 to load the apps' version.  If we are doing that for now thats fine, but
 I'm curious if that is the long term plan?

 Are we loading *any* assets outside of whats available in the top level
 www/?  Once upon a time serve served up the top level www/, then changed to
 platform www/.. should we change it back?

 I'm with Jesse that devs should not need (but can if they want) to version
 control platforms/ or plugins/, and his workaround seems quite agreeable to
 me.

 -Michal


 On Tue, Feb 4, 2014 at 11:23 AM, Braden Shepherdson 
 bra...@chromium.orgwrote:

 For the record, the App Harness makes child apps use the cordova.js,
 cordova_plugins.js, and plugins/* which are included with the App Harness,
 not the ones that come with the child app. This is because it ensures
 compatibility with the native code and the Javascript side, though it can
 break the JS API for a plugin if there's a version mismatch (which the App
 Harness can warn about). Generally these APIs are more stable than the
 bridge calls the plugin is making under the hood, hence the decision to use
 the bundled JS code.

 Braden



 On Tue, Feb 4, 2014 at 9:44 AM, Michal Mocny mmo...@chromium.org wrote:

 I think we've previously also thrown around the idea of Installed vs
 Enabled (though the context I think was about plugins and being
 conditionally disabled on various platforms).

 Perhaps we could always install the platforms you ask for, but leave then
 disabled for local builds until check_req's pass?


 On Tue, Feb 4, 2014 at 9:37 AM, Ian Clelland iclell...@chromium.org
 wrote:

  In that case, I feel like we should be calling check_reqs on create,
 but
  not throwing an error; just display a warning if the build tools aren't
  installed.
 
  We've had reports from people using the cca toolchain that `create`
 lets
  them start the project, but then they can't build because they're
 missing
  javac or something else essential. I think there's an expectation that
 we
  at least let people know early if they're going to have a bad time
 later on.
 
 
  On Tue, Feb 4, 2014 at 9:30 AM, Michal Mocny mmo...@chromium.org
 wrote:
 
  ..matter of fact, for those same reasons we should change the other
  platforms to support installation when check_reqs fail.  I should be
 able
  to test Android using AH the exact same way without a full dev
  environment.
 
 
  On Tue, Feb 4, 2014 at 9:27 AM, Michal Mocny mmo...@chromium.org
 wrote:
 
   With the app-harness, you actually can test on iOS from windows, so
 this
   is becoming a use case.  So long as the pre-packaged native plugins
  running
   on the app-harness match version requirements for your app, you
 won't
  need
   to build the native bits locally.
  
   All we need is to serve web assets from your development machine to
 the
   AH, and the easiest way to make sure we serve the right assets is
 to use
   the platform targeted prepare step.  If for no other reason, than
 just
  to
   make sure we serve the right cordova_plugins.js (though Braden will
 know
   what other platform bits the AH consumes, at some point it was
 using the
   app's cordova.js and plugins' js-modules, but I think it changed to
 used
   the AH bundled versions).
  
   -Michal
  
  
   On Mon, Feb 3, 2014 at 12:50 PM, Jesse purplecabb...@gmail.com
 wrote:
  
   Personally, I would not expect a multi-platform team to even
 checkin
  the
   platforms folder.
  
   Addressing Axel's use case: It would only require a few commands
 to be
   ready to build an iOS Cordova project:
   npm install -g cordova
   cordova create ~/HelloCordova
   cd ~/HelloCordova
   cordova platform add ios
  
   As an alternative, the phonegap-cli [1] allows actual building of
 iOS
   targets from Windows, by using the phonegap-build service. Or you
 could
   just use the PhoneGap build site directly. You would need to setup
 an
   account either way,
  
   My reason for creating the issue CB-5963 was NOT to enable the
   functionality Baz referred to with CB-5706, but to provide an easy
 way
  to
   test plugins in WP7/8 and Windows8.  I will/am changing the
  functionality
   a
   bit in the nodejs scripts because it doesn't make sense to
 platform add
   ios
   on windows because you won't be able to run the tests anyway.
  
   My plan is that the node script for testing mobile spec will only
  platform
   

Re: serving extra css

2014-02-04 Thread Brian LeRoux
Do not feel this is our territory to override. The end dev should be adding
the fonts they want in their app.


On Tue, Feb 4, 2014 at 6:13 AM, Michal Mocny mmo...@chromium.org wrote:

 On Mon, Feb 3, 2014 at 6:58 PM, Josh Soref jso...@blackberry.com wrote:

  Michal wrote:
   Wouldn't this be better handled by Brian's dream of
  browser-as-a-platform?
 
  I haven¹t worked through all the details, but I don¹t think so. If you
  want to run the blackberry specific hooks and you browse to the app from
  your blackberry browser, then you¹re going to be unhappy if it¹s a
  different platform.
 

 If you want BB-specific hooks with BB-browser, then target BB.

 If you want to test your app using cordova plugins on a desktop browser
 simulating a mobile phone, browser-as-a-platform is the I think the way to
 go.

 If you are done building your app on desktop, and now want to test more
 device specific styles (such as fonts), I would say that you should test on
 device.  Is this prohibitively difficult on BB (I have not tried so not
 sure what the turnaround time is)?  Perhaps the app harness + some of the
 auto-reload-on-device work that Braden is doing will help?


 
  Browser-as-an-engine might work, maybe.
 

 I don't think so -- engine is the concept of replacing the use of the
 system webview, but thats still running on-device.


 
   I understand that this solves your immediate problem, but I'm not sure
   extending serve is really the right long term solution for this.
 
 
   ..also, your idea of automatically look for a well known file could
 be
   done within the application by feature detecting that its running in a
   desktop browser.
 
  We¹re the SDK, not the application. I could tell authors ³if you want to
  see the fonts, put in this magic², but I¹m hoping to avoid making each
  author do that (where possible). There¹s also licensing fun w/ the fonts.
 
  I suppose that I could insert some code into cordova.js instead of having
  serve manage the css.
 
   Would it be prehibitive to dynamically inject the style
   on startup from the app itself?
 
  Doing it with a link: header is a lot cleaner than inserting strange
 logic
  into cordova.js, butŠ
 

 I didn't mean cordova.js -- I meant the specific app you are using for
 testing.  You can ship cordova-cli hooks to add this if you want to share
 the functionality between several apps.

 I'm asking: do we really need to change the tools?


 
  We can and probably will provide templates which include a style saying
  ³use font: Slate Pro². That¹s all you need for the platform engine.
 
  If we did use cordova.js instead of having serve provide the .css file,
  we¹d still need a way to get the fonts to the browser-as-engine case
  (which is handled by the .extra code).
 
  In case you¹re curious, installing fonts to a user¹s system is a no-go
 for
  a number of reasons (* it¹s painful to try to manage ‹ often requiring
  administrative privileges which isn¹t something we require, * Licensing
 is
  wrong, * it pollutes the user¹s font-space or their computer¹s font-space
  if shared, * it doesn¹t work if the user is visiting serve from a device
  other than localhost).
 
 



Jira permissions

2014-02-04 Thread Marcel Kinard
We have a new intern joining our group, Martin Canuto Gonzalez Gonzalez. Could 
someone with Jira admin grant him the permission to own Jira items? His Jira id 
is:

martin.c.glez.glez

He'll start by working on some doc issues. Thanks!

Re: Jira permissions

2014-02-04 Thread Anis KADRI
done


On Tue, Feb 4, 2014 at 10:35 AM, Marcel Kinard cmarc...@gmail.com wrote:

 We have a new intern joining our group, Martin Canuto Gonzalez Gonzalez.
 Could someone with Jira admin grant him the permission to own Jira items?
 His Jira id is:

 martin.c.glez.glez

 He'll start by working on some doc issues. Thanks!


Android/amazon-fireos: web view callbacks

2014-02-04 Thread Naik, Archana
Hi, Devs

There are 2 classes - CordovaWebViewClient and CordovaChromeClient in 
amazon-fireos repo. They basically handle WebView callbacks. Recently, while 
using Cordova internally for our projects we needed to handle few more 
callbacks and noticed they were missing. I was wondering if there is any reason 
why some callbacks were not handled? If not, I am cool.
I would like to add the ones we need for now so want to make sure I am not 
missing anything. :)

Also, one more thing, we would like to pass on these callbacks to plugins 
incase they want to handle it. If no plugin wants handle it then callback 
handling will fallback to the default cordova one.

Please let me know your thoughts.
Thanks
Archana


Re: Plugins Release!

2014-02-04 Thread Steven Gill
I am going to take the silence as lazy consensus. I will make sure to
include the new file plugin as well.

I will make sure to have a blog post of changes to review before I publish.

-Steve



On Mon, Feb 3, 2014 at 10:33 AM, Steven Gill stevengil...@gmail.com wrote:

 Hey All,

 What is the general feeling on me moving forward with a plugins release
 today? I could start the process this afternoon if there aren't any
 objections or concerns.

 Are there any plugins that shouldn't be released?



Re: Android/amazon-fireos: web view callbacks

2014-02-04 Thread Anis KADRI
Which callbacks are we talking about?


On Tue, Feb 4, 2014 at 11:20 AM, Naik, Archana na...@lab126.com wrote:

 Hi, Devs

 There are 2 classes - CordovaWebViewClient and CordovaChromeClient in
 amazon-fireos repo. They basically handle WebView callbacks. Recently,
 while using Cordova internally for our projects we needed to handle few
 more callbacks and noticed they were missing. I was wondering if there is
 any reason why some callbacks were not handled? If not, I am cool.
 I would like to add the ones we need for now so want to make sure I am not
 missing anything. :)

 Also, one more thing, we would like to pass on these callbacks to plugins
 incase they want to handle it. If no plugin wants handle it then callback
 handling will fallback to the default cordova one.

 Please let me know your thoughts.
 Thanks
 Archana



Re: Android/amazon-fireos: web view callbacks

2014-02-04 Thread Joe Bowser
Which callbacks in particular are missing? There are some callbacks we
don't touch because they can break the WebView.  These can include
anything that handles gestures (which when done wrong, can break touch
events), and other view related callbacks.  I'd highly recommend
writing tests for the existing behaviour before adding the callbacks,
since we've had very bad luck with this in the past w.r.t. the back
button.

On Tue, Feb 4, 2014 at 11:20 AM, Naik, Archana na...@lab126.com wrote:
 Hi, Devs

 There are 2 classes - CordovaWebViewClient and CordovaChromeClient in 
 amazon-fireos repo. They basically handle WebView callbacks. Recently, while 
 using Cordova internally for our projects we needed to handle few more 
 callbacks and noticed they were missing. I was wondering if there is any 
 reason why some callbacks were not handled? If not, I am cool.
 I would like to add the ones we need for now so want to make sure I am not 
 missing anything. :)

 Also, one more thing, we would like to pass on these callbacks to plugins 
 incase they want to handle it. If no plugin wants handle it then callback 
 handling will fallback to the default cordova one.

 Please let me know your thoughts.
 Thanks
 Archana


Re: Plugins Release!

2014-02-04 Thread Andrew Grieve
Sounds good!


On Tue, Feb 4, 2014 at 2:19 PM, Steven Gill stevengil...@gmail.com wrote:

 I am going to take the silence as lazy consensus. I will make sure to
 include the new file plugin as well.

 I will make sure to have a blog post of changes to review before I publish.

 -Steve



 On Mon, Feb 3, 2014 at 10:33 AM, Steven Gill stevengil...@gmail.com
 wrote:

  Hey All,
 
  What is the general feeling on me moving forward with a plugins release
  today? I could start the process this afternoon if there aren't any
  objections or concerns.
 
  Are there any plugins that shouldn't be released?
 



Re: Android/amazon-fireos: web view callbacks

2014-02-04 Thread Naik, Archana
So for example, onCreateWindow() or onProgressChanged(). We might need to
add onCreateWindow() but onProgressChanged() not that urgent.

More importantly, we want to let plugins handle there callbacks first and
incase they don't we want to fallback on default handlers in these classes.

Archana

On 2/4/14 11:29 AM, Joe Bowser bows...@gmail.com wrote:

Which callbacks in particular are missing? There are some callbacks we
don't touch because they can break the WebView.  These can include
anything that handles gestures (which when done wrong, can break touch
events), and other view related callbacks.  I'd highly recommend
writing tests for the existing behaviour before adding the callbacks,
since we've had very bad luck with this in the past w.r.t. the back
button.

On Tue, Feb 4, 2014 at 11:20 AM, Naik, Archana na...@lab126.com wrote:
 Hi, Devs

 There are 2 classes - CordovaWebViewClient and CordovaChromeClient in
amazon-fireos repo. They basically handle WebView callbacks. Recently,
while using Cordova internally for our projects we needed to handle few
more callbacks and noticed they were missing. I was wondering if there
is any reason why some callbacks were not handled? If not, I am cool.
 I would like to add the ones we need for now so want to make sure I am
not missing anything. :)

 Also, one more thing, we would like to pass on these callbacks to
plugins incase they want to handle it. If no plugin wants handle it then
callback handling will fallback to the default cordova one.

 Please let me know your thoughts.
 Thanks
 Archana



Introducing myself

2014-02-04 Thread Martin Gonzalez Glez
Hi everyone! My name is Martin Gonzalez, I work in the IBM Mobile
Foundation, and my job role it's help and contribute with this amazing
project that Cordova is. I really want to contribute to the success and
development in any way possible.
You may already know some folks, that I'm working with (Marcel Kinard,
Michael Billau, James Jong, Lisa Deluca).
I have good command using Javascript, Java, C/C++, HTML and another
languages, however, I don't have much experience with Cordova yet, but I'm
really focused on increase my knowledge and expertise.
If you have any questions, please feel free to contact me.
Thanks.
Martin Gonzalez.


Re: Introducing myself

2014-02-04 Thread Brian LeRoux
Welcome Martin! Glad to have you here and feel free to bug the list for
help anytime. =)


On Tue, Feb 4, 2014 at 12:42 PM, Martin Gonzalez Glez 
martin.c.glez.g...@gmail.com wrote:

 Hi everyone! My name is Martin Gonzalez, I work in the IBM Mobile
 Foundation, and my job role it's help and contribute with this amazing
 project that Cordova is. I really want to contribute to the success and
 development in any way possible.
 You may already know some folks, that I'm working with (Marcel Kinard,
 Michael Billau, James Jong, Lisa Deluca).
 I have good command using Javascript, Java, C/C++, HTML and another
 languages, however, I don't have much experience with Cordova yet, but I'm
 really focused on increase my knowledge and expertise.
 If you have any questions, please feel free to contact me.
 Thanks.
 Martin Gonzalez.



Odd tags

2014-02-04 Thread Josh Soref
The tag name appears to be 3.3.4-rc.1, but the release notes seem to be
3.4.0-rc.1, is the tag errant?

cordova-cli $ git show 3.3.4-rc.1
commit 81ac7f90af279f54cea38d8927fa389504e062ba
Author: Steven Gill stevengil...@gmail.com
Date:   Thu Jan 30 17:36:28 2014 -0800

updated release notes

diff --git a/RELEASENOTES.md b/RELEASENOTES.md
index 8400962..a28ffe1 100644
--- a/RELEASENOTES.md
+++ b/RELEASENOTES.md
@@ -20,6 +20,11 @@
 --
 # Cordova-cli Release Notes
 
+## 3.4.0-rc.1
+* updated to 3.4.0-rc.1
+* added firefoxos to readme
+* updated release notes
+
 ## 3.3.1-0.3.0 (Jan 30, 2015)
 * Setting version to 3.3.1-0.3.0; updated plugman reference to 0.19.0
 * CB-5913 Fail more gracefully on Windows when symlinks fail.





Re: Introducing myself

2014-02-04 Thread Lorin Beer
Hi Martin, welcome!


On Tue, Feb 4, 2014 at 12:52 PM, Brian LeRoux b...@brian.io wrote:

 Welcome Martin! Glad to have you here and feel free to bug the list for
 help anytime. =)


 On Tue, Feb 4, 2014 at 12:42 PM, Martin Gonzalez Glez 
 martin.c.glez.g...@gmail.com wrote:

  Hi everyone! My name is Martin Gonzalez, I work in the IBM Mobile
  Foundation, and my job role it's help and contribute with this amazing
  project that Cordova is. I really want to contribute to the success and
  development in any way possible.
  You may already know some folks, that I'm working with (Marcel Kinard,
  Michael Billau, James Jong, Lisa Deluca).
  I have good command using Javascript, Java, C/C++, HTML and another
  languages, however, I don't have much experience with Cordova yet, but
 I'm
  really focused on increase my knowledge and expertise.
  If you have any questions, please feel free to contact me.
  Thanks.
  Martin Gonzalez.
 



Re: Odd tags

2014-02-04 Thread Steven Gill
ahh yes, it is supposed to be 3.4.0-rc.1 I will fix that now


On Tue, Feb 4, 2014 at 1:25 PM, Josh Soref jso...@blackberry.com wrote:

 The tag name appears to be 3.3.4-rc.1, but the release notes seem to be
 3.4.0-rc.1, is the tag errant?

 cordova-cli $ git show 3.3.4-rc.1
 commit 81ac7f90af279f54cea38d8927fa389504e062ba
 Author: Steven Gill stevengil...@gmail.com
 Date:   Thu Jan 30 17:36:28 2014 -0800

 updated release notes

 diff --git a/RELEASENOTES.md b/RELEASENOTES.md
 index 8400962..a28ffe1 100644
 --- a/RELEASENOTES.md
 +++ b/RELEASENOTES.md
 @@ -20,6 +20,11 @@
  --
  # Cordova-cli Release Notes

 +## 3.4.0-rc.1
 +* updated to 3.4.0-rc.1
 +* added firefoxos to readme
 +* updated release notes
 +
  ## 3.3.1-0.3.0 (Jan 30, 2015)
  * Setting version to 3.3.1-0.3.0; updated plugman reference to 0.19.0
  * CB-5913 Fail more gracefully on Windows when symlinks fail.






RE: Plugins Release!

2014-02-04 Thread Herm Wong


Sounds good to me!
 From: agri...@chromium.org
 Date: Tue, 4 Feb 2014 14:35:01 -0500
 Subject: Re: Plugins Release!
 To: dev@cordova.apache.org
 
 Sounds good!
 
 
 On Tue, Feb 4, 2014 at 2:19 PM, Steven Gill stevengil...@gmail.com wrote:
 
  I am going to take the silence as lazy consensus. I will make sure to
  include the new file plugin as well.
 
  I will make sure to have a blog post of changes to review before I publish.
 
  -Steve
 
 
 
  On Mon, Feb 3, 2014 at 10:33 AM, Steven Gill stevengil...@gmail.com
  wrote:
 
   Hey All,
  
   What is the general feeling on me moving forward with a plugins release
   today? I could start the process this afternoon if there aren't any
   objections or concerns.
  
   Are there any plugins that shouldn't be released?
  
 
  

Re: Plugins Release!

2014-02-04 Thread Joe Bowser
Don't do it!  I think File still needs some work:

https://issues.apache.org/jira/browse/CB-5974

On Tue, Feb 4, 2014 at 3:18 PM, Herm Wong kingoftheo...@hotmail.com wrote:


 Sounds good to me!
 From: agri...@chromium.org
 Date: Tue, 4 Feb 2014 14:35:01 -0500
 Subject: Re: Plugins Release!
 To: dev@cordova.apache.org

 Sounds good!


 On Tue, Feb 4, 2014 at 2:19 PM, Steven Gill stevengil...@gmail.com wrote:

  I am going to take the silence as lazy consensus. I will make sure to
  include the new file plugin as well.
 
  I will make sure to have a blog post of changes to review before I publish.
 
  -Steve
 
 
 
  On Mon, Feb 3, 2014 at 10:33 AM, Steven Gill stevengil...@gmail.com
  wrote:
 
   Hey All,
  
   What is the general feeling on me moving forward with a plugins release
   today? I could start the process this afternoon if there aren't any
   objections or concerns.
  
   Are there any plugins that shouldn't be released?
  
 



Re: Android/amazon-fireos: web view callbacks

2014-02-04 Thread Andrew Grieve
I think both would be fine to add. onCreateWindow is a bit tougher though
because you don't know the URL upfront. It'd be hard to multiplex it... we
might just say that plugins can register to receive the next call to
window.open() though. Did you have any ideas about how this would look?


On Tue, Feb 4, 2014 at 3:12 PM, Naik, Archana na...@lab126.com wrote:

 So for example, onCreateWindow() or onProgressChanged(). We might need to
 add onCreateWindow() but onProgressChanged() not that urgent.

 More importantly, we want to let plugins handle there callbacks first and
 incase they don't we want to fallback on default handlers in these classes.

 Archana

 On 2/4/14 11:29 AM, Joe Bowser bows...@gmail.com wrote:

 Which callbacks in particular are missing? There are some callbacks we
 don't touch because they can break the WebView.  These can include
 anything that handles gestures (which when done wrong, can break touch
 events), and other view related callbacks.  I'd highly recommend
 writing tests for the existing behaviour before adding the callbacks,
 since we've had very bad luck with this in the past w.r.t. the back
 button.
 
 On Tue, Feb 4, 2014 at 11:20 AM, Naik, Archana na...@lab126.com wrote:
  Hi, Devs
 
  There are 2 classes - CordovaWebViewClient and CordovaChromeClient in
 amazon-fireos repo. They basically handle WebView callbacks. Recently,
 while using Cordova internally for our projects we needed to handle few
 more callbacks and noticed they were missing. I was wondering if there
 is any reason why some callbacks were not handled? If not, I am cool.
  I would like to add the ones we need for now so want to make sure I am
 not missing anything. :)
 
  Also, one more thing, we would like to pass on these callbacks to
 plugins incase they want to handle it. If no plugin wants handle it then
 callback handling will fallback to the default cordova one.
 
  Please let me know your thoughts.
  Thanks
  Archana




Re: Introducing myself

2014-02-04 Thread Andrew Grieve
Welcome Martin!


On Tue, Feb 4, 2014 at 4:31 PM, Lorin Beer lorin.beer@gmail.com wrote:

 Hi Martin, welcome!


 On Tue, Feb 4, 2014 at 12:52 PM, Brian LeRoux b...@brian.io wrote:

  Welcome Martin! Glad to have you here and feel free to bug the list for
  help anytime. =)
 
 
  On Tue, Feb 4, 2014 at 12:42 PM, Martin Gonzalez Glez 
  martin.c.glez.g...@gmail.com wrote:
 
   Hi everyone! My name is Martin Gonzalez, I work in the IBM Mobile
   Foundation, and my job role it's help and contribute with this amazing
   project that Cordova is. I really want to contribute to the success and
   development in any way possible.
   You may already know some folks, that I'm working with (Marcel Kinard,
   Michael Billau, James Jong, Lisa Deluca).
   I have good command using Javascript, Java, C/C++, HTML and another
   languages, however, I don't have much experience with Cordova yet, but
  I'm
   really focused on increase my knowledge and expertise.
   If you have any questions, please feel free to contact me.
   Thanks.
   Martin Gonzalez.
  
 



Re: Android/amazon-fireos: web view callbacks

2014-02-04 Thread Joe Bowser
I think that both would be fine to add.  It's probably a good idea to
add these hooks, to be honest, since we don't have any default
behaviour associated with them AFAIK.

On Tue, Feb 4, 2014 at 3:22 PM, Andrew Grieve agri...@chromium.org wrote:
 I think both would be fine to add. onCreateWindow is a bit tougher though
 because you don't know the URL upfront. It'd be hard to multiplex it... we
 might just say that plugins can register to receive the next call to
 window.open() though. Did you have any ideas about how this would look?


 On Tue, Feb 4, 2014 at 3:12 PM, Naik, Archana na...@lab126.com wrote:

 So for example, onCreateWindow() or onProgressChanged(). We might need to
 add onCreateWindow() but onProgressChanged() not that urgent.

 More importantly, we want to let plugins handle there callbacks first and
 incase they don't we want to fallback on default handlers in these classes.

 Archana

 On 2/4/14 11:29 AM, Joe Bowser bows...@gmail.com wrote:

 Which callbacks in particular are missing? There are some callbacks we
 don't touch because they can break the WebView.  These can include
 anything that handles gestures (which when done wrong, can break touch
 events), and other view related callbacks.  I'd highly recommend
 writing tests for the existing behaviour before adding the callbacks,
 since we've had very bad luck with this in the past w.r.t. the back
 button.
 
 On Tue, Feb 4, 2014 at 11:20 AM, Naik, Archana na...@lab126.com wrote:
  Hi, Devs
 
  There are 2 classes - CordovaWebViewClient and CordovaChromeClient in
 amazon-fireos repo. They basically handle WebView callbacks. Recently,
 while using Cordova internally for our projects we needed to handle few
 more callbacks and noticed they were missing. I was wondering if there
 is any reason why some callbacks were not handled? If not, I am cool.
  I would like to add the ones we need for now so want to make sure I am
 not missing anything. :)
 
  Also, one more thing, we would like to pass on these callbacks to
 plugins incase they want to handle it. If no plugin wants handle it then
 callback handling will fallback to the default cordova one.
 
  Please let me know your thoughts.
  Thanks
  Archana




CB-5974, or, Re-thinking broken-by-default (Was: Re: Plugins Release!)

2014-02-04 Thread Ian Clelland
On Tue, Feb 4, 2014 at 6:20 PM, Joe Bowser bows...@gmail.com wrote:

 Don't do it!  I think File still needs some work:

 https://issues.apache.org/jira/browse/CB-5974


It's too early yet to tell whether this has become a problem, but obviously
this is something that people are going to run into (and are already
running into [1])

Background:

Because of CB-5915 [2] (and it's android sister, CB-5916 [3]), developers
who upgrade File will find that their applications crash with a log message
the first time they use any of the File API. (I'd crash sooner, but File is
not currently a load-on-startup sort of plugin)

There was a lot of discussion about this on the list [4], and I thought
that we had reached consensus, but maybe it needs one more hard look.

The Problem:

The log message hopefully tells them what they need to do to fix the
problem, but many (most?) devs aren't going to see it; they're only going
to see that their app is broken now, and come to us (hopefully) or
StackOverflow (more likely) to figure out why and what to do about it. They
will probably be loud, we will probably be blamed for their apps breaking
(no matter how simple the fix is,) and it will probably be a bad time for
everybody.

It's unfortunate that we feel we have to do this; there just doesn't seem
to be any other way to encourage developers to use anything but the old
locations for persistent files. On Android this is the
root-of-the-sdcard-even-if-its-just-a-partition-shared-by-all-apps
location. On iOS, its the Documents directory, although Library makes far
more sense for these sort of files.

If we added a default value, the only possible default we could use is
Compatibility, which means that approximately 100% of new apps will ship
with that default. We *can't* use the new location as the default, because
that will cause real pain for the users, not the devs. Real users will lose
access to real data, and that worries me much more than a mob of angry devs
with pitchforks.


Options:

1. Go big or go home: Make it crash harder. I was already going to add a
paragraph to the plugin.xml file to be shown on install, but we *could*
make the app break on launch, rather than on the first use of File.
We could force File to load on startup, or we could make Javascript alert()
the dev on startup (something like please fix config.xml and then delete
this line), or we could break the plugin encapsulation and make
ConfigParser check for this special case.
Maybe we could go even further and find a way to make it break on build. At
least then apps wouldn't make it to the store broken.

2. Leave it the way it is. Brace for the angry mob with blog posts, release
notes, install guides, READMEs, vigilance on StackOverfow, and hope that
it's enough.

3. Put in the safe default and live with it. Accept that every single
Cordova app is going to use the default and that their file storage roots
will suck. By the time you've educated a developer, chances are their apps
have already been released and have users with stored data. Work hard on a
migration utility/plugin and make sure that it can never possibly lose data.

With my Google hat[5] on, I don't mind #3 -- we've ensured that the apps
created with our framework use the new defaults, since they're all new apps
by definition. Wearing my Apache hat, though, I don't think it's best for
Cordova in the long term. At some point, I think we should rip off the
bandaid. If we don't do it now, with a major rev bump of File, then we're
just postponing the hurt.

There may be other options; lets try to get consensus on this before
pulling the trigger.


[1] https://issues.apache.org/jira/browse/CB-5899
[2] https://issues.apache.org/jira/browse/CB-5915
[3] https://issues.apache.org/jira/browse/CB-5916
[4] http://markmail.org/message/tzcljj3xgycbkx3g
[5] http://www.flickr.com/photos/nooks/6858535568/



 On Tue, Feb 4, 2014 at 3:18 PM, Herm Wong kingoftheo...@hotmail.com
 wrote:
 
 
  Sounds good to me!
  From: agri...@chromium.org
  Date: Tue, 4 Feb 2014 14:35:01 -0500
  Subject: Re: Plugins Release!
  To: dev@cordova.apache.org
 
  Sounds good!
 
 
  On Tue, Feb 4, 2014 at 2:19 PM, Steven Gill stevengil...@gmail.com
 wrote:
 
   I am going to take the silence as lazy consensus. I will make sure to
   include the new file plugin as well.
  
   I will make sure to have a blog post of changes to review before I
 publish.
  
   -Steve
  
  
  
   On Mon, Feb 3, 2014 at 10:33 AM, Steven Gill stevengil...@gmail.com
   wrote:
  
Hey All,
   
What is the general feeling on me moving forward with a plugins
 release
today? I could start the process this afternoon if there aren't any
objections or concerns.
   
Are there any plugins that shouldn't be released?
   
  
 



Re: CB-5974, or, Re-thinking broken-by-default (Was: Re: Plugins Release!)

2014-02-04 Thread Josh Soref
Is it impossible to have reads merged from both locations, but writes go to the 
new location, and when a write completes in the new location, delete the old 
one?