Try to use the official box sdk instead of having our own fork or switch to the crossplatform sdk that we have

I just got an email about them wanting us to upgrade to TLS 1.1 or Higher before June 15. This will probably be addressed in updated SDKs not sure if it is already in our version. In general, we miss out on their updates by having our own forks which should be avoided if possible.

Indeed, IIRC the big reason was that we wanted to support earlier iOS versions(<iOS 9).

That's correct and the reason why we still use the SDK you patched years ago.

However, I don't see where the currently used SDK mandates the use of TLS versions older than 1.1 unless this is done implicitly on the server-side when BoxAPIVersion 2.0 is requested. The current code mandates https everywhere, but does neither check nor set TLS versions. Thus, I'm a bit puzzled what they are asking for exactly.

changed milestone to %3.1.1

Turns out that box doesn't support cocoapoads anymore and encourages Users to switch to Carthage. Furthermore the last available sdk over cocoapods (box-ios-sdk 1.0.14) has no tvOS support.

So the current options are :

1. Adding Carthage support just for this SDK which would also allow tvOS then

It might make sense to consider step by step to migrate everything to Carthage in the long run see:

https://medium.com/xcblog/carthage-or-cocoapods-that-is-the-question-1074edaafbcb,

https://medium.com/ios-os-x-development/cocoapods-vs-carthage-675633e89c3e

2. Wait for the Crossplatform sdk to be merged into VLC and switch all Cloud services as part of bigger feature work.

seeing this I'd recommend adding carthage support to stay up to date with box latest changes.

How hard would it be to roll our own cocoapods support for box until libcloudstorage is merged?

No clue to be honest but do you really want to do and support that for every release ?

Unless they change a lot of their file hierarchy in every release, I don’t see to much work here (unless they do weekly releases).

We could also do a more radical approach and remove support for box.com until libcloudstorage is merged. We should see if we can get usage numbers compared to google drive and Dropbox, which appear to be reasonably popular.

I was thinking about the same thing earlier box has 44 million users vs. 500 million users on dropbox & 800 million for GDrive see : https://expandedramblings.com/index.php/box-statistics/ & https://expandedramblings.com/index.php/dropbox-statistics/ & https://en.wikipedia.org/wiki/Google_Drive In addition we see this here : #246 (closed)

Let's do it and remove it then. If the protest is noticeable, we can still revert and modernize the existing implementation.

Either way, moving to libcloudstorage will be needed for all the Clouds at some point in the not so distant future.

I'm still a bit hesitant though because box is actively growing and given that we had 80000 users on GDrive I expect 4400 to be effected by that cut.

I think leaving it as is and fix the other issues until we integrate libcloudstorage is a less radical approach.

to be resolved as part of #255

changed milestone to %3.2.0

changed milestone to %Future

added Cloud label

mentioned in issue #1849

marked this issue as related to #1849