Regression with DPI calculation

[Maurice Amsellem <ma...@systar.com>]

Hi Justin,

While testing 4.12 RC3 I just noticed a regression in the way RuntimeDPIProvider computes runtime DPI on iPad.

The intent is to get 160 DPI on iPad and 320 DPI on iPAD retina.

This is done by testing version includes iOS and extent = 1024 ( 2048 for retina).

I didn't test "os" contains iPAD so that DPI calculation are the same on the device (where os contains "iPAD") and when running on ADL (where os would be either WIN or MAC).

Since you have added the test on os = iPAd, now runtime DPI is 160 on the device (correct) but 120 when simulating iPAd on ADL.

IMO, that's confusing.

Can you please set back the test to what it was before, so that iPAD on device or ADL give the same result.

Thank you.

However, 

RE: Regression with DPI calculation

[Maurice Amsellem <ma...@systar.com>]

Well, the test is on iOS only and extent be 1024 or 2048.

AFAIK, iPads are the only iOS devices so far to have the above characteristics.

http://en.wikipedia.org/wiki/List_of_iOS_devices

This could change in the future, but for now, it's true.

I know this is kind of hacky, but I couldn't find a better way that complies with both ADL and devices.

I think that AIR Capabilities should be more consistent when running on ADL, but unfortunately it's not, and we have to do  with it.
I will raise a JIRA at Adobe about this  when I have some time.

WDYT?

Maurice 

-----Message d'origine-----
De : Justin Mclean [mailto:justin@classsoftware.com] 
Envoyé : dimanche 2 mars 2014 22:55
À : dev@flex.apache.org
Objet : Re: Regression with DPI calculation 

Hi,

Anyway you can see to avoid those calculations on  non iPad devices?

Justin

Re: Regression with DPI calculation

[Justin Mclean <ju...@classsoftware.com>]

Hi,

Anyway you can see to avoid those calculations on  non iPad devices?

Justin

RE: Regression with DPI calculation

[Maurice Amsellem <ma...@systar.com>]

I will do.

Maurice 

-----Message d'origine-----
De : Justin Mclean [mailto:justin@classsoftware.com] 
Envoyé : lundi 3 mars 2014 03:57
À : dev@flex.apache.org
Objet : Re: Regression with DPI calculation 

Hi,

Change back to if (iOS) - would be nice if there was another way as it punishes iPhone uses.

Given that the OS doesn't change at runtime you could probably calculate that once and use a cached result for  all future calls.

Thanks,
Justin

Re: Regression with DPI calculation

[Justin Mclean <ju...@classsoftware.com>]

Hi,

Change back to if (iOS) - would be nice if there was another way as it punishes iPhone uses.

Given that the OS doesn't change at runtime you could probably calculate that once and use a cached result for  all future calls.

Thanks,
Justin

RE: Regression with DPI calculation

[Maurice Amsellem <ma...@systar.com>]

>Will changing it back to isIOS (rather than isIPad) have the same issue in the emulator?
No, it will be fine in the emulator, because "IOS" is in version, which is the same on ADL and device (IIRC).

Note: maybe you would want to change Platform.isIPad ASDOC comment to mention that is returns true only on the device, not ADL. 

>And this is not a regression as such as that code in question hadn't been released in 4.11.
Agree this is not a regression.

Thank you.

Maurice 

-----Message d'origine-----
De : Justin Mclean [mailto:justin@classsoftware.com] 
Envoyé : dimanche 2 mars 2014 23:43
À : dev@flex.apache.org
Objet : Re: Regression with DPI calculation 

Hi,

Will changing it back to isIOS (rather than isIPad) have the same issue in the emulator?

And this is not a regression as such as that code in question hadn't been released in 4.11.

Thanks,
Justin

Re: Regression with DPI calculation

[Justin Mclean <ju...@classsoftware.com>]

Hi,

Will changing it back to isIOS (rather than isIPad) have the same issue in the emulator?

And this is not a regression as such as that code in question hadn't been released in 4.11.

Thanks,
Justin