I have an app more or less ready to ship and for many reasons I would prefer to remain on corona public build 971 (e.g. widgets buttons). I do not use:
system.getInfo(“deviceID”)
ads networks (e.g. Rev Mob).
I have retina graphics / iphone 5 compatibility (even on iPad). Is there any reason why I should migrate to newer public build? E.g. does Corona SDK access UDIDs “in the background”?
I’d change, if I were you. You can download the old widget library from GitHub, if you want that.
I’d always advise downloading the newest version of a product, even if you can’t see a difference. There are usually unseen updates that might speed it up a bit, or slightly improve this or that, bugs that have been fixed, issues that have been updated, and suchlike.
Not to mention new things that are added… PNG display captures and nearest-neighbor texture filters were my favorites for this version 
Plus, why stay with the old version when you have a newer one?
C
–[[Ah, you! I have just finished implementing your particle system. Testing on device tomorrow morning. Simulator shows no leaks, however painfully slow frame-rate when in simulator + console debug mode (why is that?)… Wish me luck 
]]–
Back to the topic: the reason is simple, I am not sure what might or might not work besides the widgets (behind the scenes only Corona knows what they have done) and I have just finished painful beta testing (save for particles, which I had decided to integrate as an add-on at the last moment). Extra testing = extra delay.
Having said that, generally speaking, you have a point.
Are you using analytics?
Yes I log events in Flurry, but I am not consciously trying to get UDIDs.
I’d have to go back through the release notes, but I’m pretty sure the UDID was removed after 971 and even though you are not getting the UDID, the analytics library may. You can always submit with 971 and see if it gets approved, if not, you could go to 1076 and use the widget 1.0 library from our github.
Concerning “I am not sure what might or might not work besides the widgets”, as far as I can tell, it’s worked exactly the same except for widgets. I couldn’t tell whether I was using 971 or the newer one, other than the widgets - and they took about 20 mins to fix (search for “default=”, replace with “defaultFile=”, search for “over=”, replace with “overFile=”, if you want to do the least). Of course, I did start work on imageSheet widget integration, but it’s not really required.
C
Sorry for the late reply, we might be in very different time zones.
@Rob – I think that you are right. So the analytics might be a problem here.
@Caleb P – Thanks a lot for the suggestion.
Sounds like I have three possible courses of action here. Thanks a lot guys this was very useful.
NOTE 1: I have managed to find an old note from Joshua Quick in another thread confirming that UDIDs have been taken out of analytics by build #1076. The post is the second to last:
http://forums.coronalabs.com/topic/33702-crash-reports-on-android-devices/
Many thanks guys. Always a pleasure dealing with this community.
I have opted for migrating to 1076 straight away and have an interesting “quirk” to report.
When you use an image sheet image for creating a button and you wanted to rescale such button upon creation (say that it is too big in the image sheet), you cannot actually do it by using different width and height parameters inside a widget button “constructor”. You could with widgets 1.0+corona public build 971. I hope that this behaviour is limited to widgets because i cannot really remember if and where I have used display,newImageRect(…) with different heights and widths from the image sheet.
I’d change, if I were you. You can download the old widget library from GitHub, if you want that.
I’d always advise downloading the newest version of a product, even if you can’t see a difference. There are usually unseen updates that might speed it up a bit, or slightly improve this or that, bugs that have been fixed, issues that have been updated, and suchlike.
Not to mention new things that are added… PNG display captures and nearest-neighbor texture filters were my favorites for this version
Plus, why stay with the old version when you have a newer one?
C
–[[Ah, you! I have just finished implementing your particle system. Testing on device tomorrow morning. Simulator shows no leaks, however painfully slow frame-rate when in simulator + console debug mode (why is that?)… Wish me luck ]]–
Back to the topic: the reason is simple, I am not sure what might or might not work besides the widgets (behind the scenes only Corona knows what they have done) and I have just finished painful beta testing (save for particles, which I had decided to integrate as an add-on at the last moment). Extra testing = extra delay.
Having said that, generally speaking, you have a point.
Are you using analytics?
Yes I log events in Flurry, but I am not consciously trying to get UDIDs.
I’d have to go back through the release notes, but I’m pretty sure the UDID was removed after 971 and even though you are not getting the UDID, the analytics library may. You can always submit with 971 and see if it gets approved, if not, you could go to 1076 and use the widget 1.0 library from our github.
Concerning “I am not sure what might or might not work besides the widgets”, as far as I can tell, it’s worked exactly the same except for widgets. I couldn’t tell whether I was using 971 or the newer one, other than the widgets - and they took about 20 mins to fix (search for “default=”, replace with “defaultFile=”, search for “over=”, replace with “overFile=”, if you want to do the least). Of course, I did start work on imageSheet widget integration, but it’s not really required.
C
Sorry for the late reply, we might be in very different time zones.
@Rob – I think that you are right. So the analytics might be a problem here.
@Caleb P – Thanks a lot for the suggestion.
Sounds like I have three possible courses of action here. Thanks a lot guys this was very useful.
NOTE 1: I have managed to find an old note from Joshua Quick in another thread confirming that UDIDs have been taken out of analytics by build #1076. The post is the second to last:
http://forums.coronalabs.com/topic/33702-crash-reports-on-android-devices/
Many thanks guys. Always a pleasure dealing with this community.
I have opted for migrating to 1076 straight away and have an interesting “quirk” to report.
When you use an image sheet image for creating a button and you wanted to rescale such button upon creation (say that it is too big in the image sheet), you cannot actually do it by using different width and height parameters inside a widget button “constructor”. You could with widgets 1.0+corona public build 971. I hope that this behaviour is limited to widgets because i cannot really remember if and where I have used display,newImageRect(…) with different heights and widths from the image sheet.