[drampelt]

I've been having similar issues with one of our apps, the reviewer can't seem to log in and keeps getting network error messages. No customers have had any issues, I've tried logging in to multiple accounts including the apple reviewer account on multiple networks and I have no issues whatsoever. It's really frustrating.

[kevinob11]

I had the same issue once and was convinced that I must have pointed the reviewer account to the wrong server. I spent 2 days debugging before I finally asked the reviewer to check their network connection, I was approved the next day. They did send a nice response admitting the issue was on their end, lesson learned.

[_qulr]

> lesson learned

Did they learn a lesson, or just you?

[kevinob11]

Just me, spend ~15 minutes sanity checking then just ask them to test again.

[ehsankia]

Yep, I've been hearing a lot of issues like this for the past week. It's definitely an issue with their internal test servers, and even though hundreds of different apps are having the exact same issue, they keep insisting it's the devs fault...

[heavyset_go]

Apple takes 30% of all sales and yet can't even hire skilled quality assurance workers to run the approval process for their App Store?

There needs to be real competition in the mobile app distribution space because this is absurd.

[ehsankia]

I don't think the problem here is with the QA workers. This is clearly a bug in the underlying internal test framework, which the QA workers have very little info about. And the likelihood of a given worker seeing enough of these errors for them to realize it's an issue with the system and not the app is fairly low.

The real issue is that the engineering team who maintains the internal app checking system 1. needs to have infrastructure to detect abnormal amount of a given error and 2. need to notify the QA team so the QA team can communicate it with the devs, rather than just blaming the apps.

[ethbro]

You'd think people would look at the initial healthcare.gov mess and make some conclusions.

One of which should maybe be "Don't strictly isolate teams, with unowned space between their output and the next team's input, and no method by which post-delivery failure reflects back on them."

[_qulr]

Apple said they review 100,000 apps per week.

I don't think anyone can do that sanely at scale.

Which is why nobody should try.

[heavyset_go]

1,000 - 2,000 dedicated quality assurances workers could review between 1 to 2.5 app per hour each week. Adding more personnel cuts this number down drastically.

It certainly can be done. Besides, this is a problem Apple has decided it can handle, since it decided that its customers can't benefit from competition between app stores with different approval processes.

[sjwright]

In my opinion, legitimate competition between App Stores would make the iPhone strictly worse for myself and for everyone I know that owns an iOS device. For me, the mandatory app review process is the big differentiator between the iPhone and Android and is the main reason why I have not switched.

[_qulr]

> 1,000 - 2,000 dedicated quality assurances workers could review between 1 to 2.5 app per hour each week.

Are those people even out there and available?

I mean, maybe there are right now because of Great Depression II, but were they available from 2008 to 2019?

As a company, Apple is generally against remote work, with only grudging exceptions, so that's another issue in hiring.

[com2kid]

Microsoft used to have far more manual testers than that.

2000 testers isn't unreasonable.

[heavyset_go]

> Are those people even out there and available?

There are over 4 million of software engineers in the United States alone, and I'd wager that many of them are capable of doing QA. Apple is a company that is able to pay competitive wages for their talent.

[sjwright]

As a consumer, I'd rather Apple sometimes do a bad job and sometimes piss off developers than do nothing. I'm certainly not saying that Apple is doing a good job, but for most end users, the alternative is strictly worse.

[scblzn]

If it can help you, you should know that they test everything over a VPN that doesn’t support UDP. So if your app makes use of UDP you’ll need a fallback method.

[jchb]

UDP is fine. We have apps on the App Store that use UDP. This is likely a NAT related issue. That is not going to be an issue just for App Store review, but also for tons of other networks out there. As of lately, carrier-grade NAT deployments are very common.

It's usually impossible to establish a two-way UDP "connection" between two peers that are both behind NAT. This varies depending on the type of NAT. If just one peer is behind NAT you can typically establish the "connection" using NAT hole punching. If both peers are behind NAT you may need to proxy the UDP packets using an intermediary server.

There are protocols for hole punching, such as ICE (https://tools.ietf.org/html/rfc8445), but you can also roll your own - if you control all endpoints.

"Connection" in quotation marks because UDP is a connection-less protocol, but the NAT port mappings that are established are sometimes referred to as a connection.

[ifmpx]

What, so iOS can't run UDP-dependant apps at all?!

[packetlost]

What kind VPN doesn't support UDP!?

[causality0]

For example, I've never managed to establish any sort of UDP connection over an AT&T line, neither home nor cellular.

[jchb]

UDP on AT&T and other carriers is just fine, but you need to do NAT hole punching or proxy the UDP packets. See my comment to the grandparent.

[foobarbazetc]

Is this for macOS?

We have the same issue and have never been able to get our macOS app approved (not a big deal but... annoying).

[drampelt]

iOS, we don't have macOS apps so haven't run into it there but could be a similar problem.

[yxhuvud]

Latency issues? Perhaps the reviewer is on the other side of the globe?

[ubertoop]

Firebase auth?