About Test the App on IOS Devices with multiple IOS Versions

[cthesky]

Hi all,

Since there are many IOS versions and the new version always evolve eventually, can I know how you all handle the app testing on multiple IOS versions on IOS devices? For example, if I want my app support IOS versions 4.3 until 6.0. Should I need to test the app on various devices which run IOS 4.3 until 6.0? or normally We support the latest IOS version? And so far I only own a IPhone so how can I do testing on various version on devices? I am thinking upgrade the version once done a testing on current version but it seems not a good option because once upgrade I can't downgrade it to test my next app. How can I ensure my app can run well in a device with IOS 4.3 to 6.0? Anyone has any idea or suggestions?

Any comments are welcome. Thanks a lot.

[jnoxx]

I just bought all devices, to test on all. I have all iPads, just currently lacking the iPad 1 gen again, but i'm watching out for one. For the rest, we have also all kind of devices at work to double test. You have the simulator for great tests to, just to check if the basics work. (like async vs sync downloading from 5.0 calls etc).

[cthesky]

Originally Posted by jnoxx

I just bought all devices, to test on all. I have all iPads, just currently lacking the iPad 1 gen again, but i'm watching out for one. For the rest, we have also all kind of devices at work to double test.

Thanks for your reply.

I see. So far, I don't have enough $$$ to buy all devices. Hope one day I have enough budget to buy all. haha.

You have the simulator for great tests to, just to check if the basics work. (like async vs sync downloading from 5.0 calls etc).

yup, will use simulator to test and check all basics work. For me, it seems a only way to test my app on various IOS versions.

[jnoxx]

You will get there, good luck.
You can download older simulators to test too normally

[PhoneyDeveloper]

Figure that you need to test only on the last dot release of a particular OS version. So 4.3.5 also covers 4.3. 6.0.1 also covers 6.0. Otherwise it becomes impossible time-wise to do all the testing.

[SteppingStone]

This is really tricky. Apple doesn't make it easy to downgrade software on a device, so the most convenient way to test is to keep multiple devices. A few observations:

1. From a performance tuning point of view, pick the most underpowered devices available. We mostly do iPad development, and iPad 1 and iPad 3 are relatively underpowered given the number of pixels. If something runs smoothly on iPad 1, for example, it will almost certainly run even better on iPad 2.

2. From a general bugginess point of view, there's no substitute for testing on all the major software versions you can. We've seen crashers which only occurred on a particular version; I recall a retain/release issue which caused a crash only on iOS 5.1, but not 5.0 or 6.0. Unfortunately Apple doesn't make this easy.

3. Design with memory warnings in mind if you're going to support < iOS 6.0. When you test, always make sure your app is robust to memory warnings.

4. The simulator can be your friend for testing on various iOS versions. It is far from ideal, but it's definitely helped us to run on different simulator versions. This can catch a lot of UIKit differences, unsupported API sloppiness, even differences in memory management, etc. early on. Dumb mistakes happen and the simulator can help find them.

[cthesky]

Originally Posted by jnoxx

You will get there, good luck.
You can download older simulators to test too normally

ok. Thanks.

----------

Originally Posted by PhoneyDeveloper

Figure that you need to test only on the last dot release of a particular OS version. So 4.3.5 also covers 4.3. 6.0.1 also covers 6.0. Otherwise it becomes impossible time-wise to do all the testing.

I see. So I can more focus my testing on last dot release of a particular OS version. Thanks for figure out.

----------

Originally Posted by SteppingStone

This is really tricky. Apple doesn't make it easy to downgrade software on a device, so the most convenient way to test is to keep multiple devices. A few observations:

1. From a performance tuning point of view, pick the most underpowered devices available. We mostly do iPad development, and iPad 1 and iPad 3 are relatively underpowered given the number of pixels. If something runs smoothly on iPad 1, for example, it will almost certainly run even better on iPad 2.

2. From a general bugginess point of view, there's no substitute for testing on all the major software versions you can. We've seen crashers which only occurred on a particular version; I recall a retain/release issue which caused a crash only on iOS 5.1, but not 5.0 or 6.0. Unfortunately Apple doesn't make this easy.

3. Design with memory warnings in mind if you're going to support < iOS 6.0. When you test, always make sure your app is robust to memory warnings.

4. The simulator can be your friend for testing on various iOS versions. It is far from ideal, but it's definitely helped us to run on different simulator versions. This can catch a lot of UIKit differences, unsupported API sloppiness, even differences in memory management, etc. early on. Dumb mistakes happen and the simulator can help find them.

I see. Got it. Thanks for your reply.