iPad Air Test - Imagesheet Debacle?

Solar2D iOS
#1

This is my first test on the iPad Air, but have successfully tested on all others for the past four months.  The failure looks like all images [frame references] have been mixed up and image locations have been swapped…  I have not made any changes to the config.lua or build.settings…

My process for building the app (for iOS) is via XCode (latest SDK 8.4) and have my iPad Mini connected (run/build).  However, for the other ipads/iphones that I’ve tested, I take the filename.app, zip it up and put it on DropBox/Google Cloud drive so all the others who I have testing for me (iPad’s/Mini’s/iPhones), can download and install via iTunes.  I repeated this process for my iPad Mini (1st gen), to load via iTunes, and it worked flawlessly.  Has anyone else seen this type of behavior?

App built using imagesheets created by TexturePacker

  • no sheet larger than 2048x2048;

  • 1 scaling variant - not using @2x variant, just reducing all images by 50% and creating one sheet for all devices

  • Texture format = PNG

  • Png Opt Level = 0

  • pixel format RGBA8888

  • Size constraints = Any size

  • Algorithm = MaxRects

  • Multipack (for some sheets)

  • Trim Mode = Trim and Reduce border artifacts

For the past four months, I have successfully tested on all simulator models AND built and tested the app on the following devices:

iOS:

iPad Mini 1st gen. 2nd gen.

iPad 1st gen., 3rd gen.

iPad Air (Retina/ME898X/A) - failed

iPhone 5/6

Android:

Samsung Galaxy S5

Samsung Galaxy Note 3

Lindsay pad (old technology/software)

first gen Kindle 7"/latest gen Kindle 7"

various phones (hi-res)

OSX: 10.10.4, XCode: 6.4, Corona Enterprise build: 2015.2688, TexturePacker: 3.9.2

#2

I have not heard of this problem before nor has anyone else reported it. Can you zip up your test project and file a bug report so we can look at it?

The iPad Air is a higher resolution device and would try to use higher resolution imagesheets if available but it doesn’t sound like the case here. Are you sure you don’t have any alternate scaling .lua files in the project that could be confusing things?

-Tom

#3

No external scaling files whatsoever.  I’ll submit the bug report.  Shouldn’t this have shown up in the simulator?

#4

Any update on this?  Getting close to production release.

#5

Hi @douglascavanaugh,

Do you have a case # for the bug report so I can inspect? I have an iPad Air which I can do a bit of testing on…

Thanks,

Brent

#6

Well I did, but it seems to be evading me at the moment.  Just searched through my email and couldn’t find anything from Corona on a case number…  I submitted the bug on August 4th if that helps along with a the details and a zip file of a scene (which I still have)…  Should I do another?  Arghhh…  :wink:

#7

Sure, file it again. I don’t see any bug reports from August 4th…

#8

Okay, just submitted again and this was the message:

Corona Issue/Bug Submission

Thank you for your submission. We will do our best to evaluate your case and respond as quickly as possible.

Can you confirm?

#9

You should receive an e-mail with a case # shortly. When you do, please post that number here for reference.

#10

Just so that you are aware Brent, I have yet to receive the case #…

#11

Hi @douglascavanaugh,

Please carefully check that it wasn’t flagged by spam e-mail filters…

Brent

#12

Brent, this is important to me, so all folders are/have been/are being monitored…  thank you.

#13

Hi @douglascavanaugh,

I can’t seem to find your case anywhere in our system. Is it “public” in such a way that you could share it here via DropBox or similar? If not, you can send it to me privately and I’ll inspect.

Brent

#14

Just tried to send you a private message, got the following error:

The following errors were found

The member Brent Sorrentino can not use the messaging system

This personal message has not been sent

No attachments, just links…

#15

Please send it to me at brent (at) coronalabs (dot) com. Thanks.

#16

Hi Douglas,

I’m not seeing all elements in your demo compared to the screenshots. For example, the blue lined background is not present in the demo (background is just black). Also, when I reach the “monitor” screen, I see no text labels on the R G or B mini-screens. Is this the issue you’re facing, or something else?

If that is not the issue, then everything looks pretty good to me, and matches your screenshots.

Brent

P.S. - I did notice that the spinning satellites gradually gain speed over time, eventually spinning madly fast. You should check your code as to why this is occurring.

#17

Brent, you are right, that’s not the issue (missing images - I did that because the file size was too large for the bug reporting submission). The issue was that the images were all in the wrong places.  Can you tell me what build you used?  And can you also tell me what model iPad Air Retina you are using?  Thank you Brent!!!

I am “playing” with the applyAngularImpulse so that’s not an issue either.

#18

Also, I glanced at your code and you have the entire project set up for dynamic image selection (various sheets of various resolution for optimal device display). However, you aren’t actually including those different sheets. Now, it should still fall back to the core “@1x” sheets, but I suggest that you don’t tell TP to use dynamic image selection, and throughout your code where you use “display.newImageRect()”, change it to “display.newImage()”.

Brent

#19

Don’t apply angular impulse repeatedly (likely that is happening, although you’ll have to determine why). Use “object.angularVelocity=av” instead, where “av” is a number for the speed you prefer.

#20

My iPad Air is the 1st-gen model. It’s running iOS 8.4. I built using Corona #2646 (latest official build).