Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Apple revoked longtime Mac developer's code signing certificate with no warning

twitter.com

Post reply on HN
Print view View original

141–150 of 180 posts

Re: Apple revoked longtime Mac developer's code signing certificate with no warning

#141

Post by blibble »

Reason077 wrote: post #90
mkathuri wrote: post #38

From the page: https://developer.apple.com/documentation/xcode/notarizing_m... > Notarization is not App Review. The Apple notary service is an automated system that scans your software for malicious content, checks for code-signing issues, and returns the results to you quickly. If there are no issues, the notary service generates a ticket for you to staple to your software; the notary service also publishes that ti…

The developer's website (software.charliemonroe.net) is blocked by my ISP (Vodafone UK)'s adult content filter. This is strange as it does not appear to contain any adult content. I don't think it's related to it being a "YouTube downloader" app either. There are many apps with this functionality and, so far as I can tell, none of the others are blocked.

most of the internet is blocked by those mobile carrier content filters

Re: Apple revoked longtime Mac developer's code signing certificate with no warning

#142

Post by raxxorrax »

ashtonkem wrote: post #108
raxxorrax wrote: post #87

Earlier quoted context omitted.

Eh, yes, we have much more locked down systems with reduced capabilities. Consoles used it maybe for years and they are notorious for being useless for any third party content. You cannot simply write your own drivers for several devices anymore, so you basically booted a lot of engineers. It infected whole industries. We have locked down agricultural devices where users pay for expensive hacks. It is a complete nigh…

Consoles haven’t been open for 30 years, but I can still run unsigned code on my Mac. They’re not a good argument for us actually sliding down a slippery slope. And writing drivers for hardware has often been impossible or effectively so due to poor documentation and/or signing. Do you not remember the era when most WiFi cards didn’t work with Linux, or when graphics cards required closed source binary blobs to even…

I remember the time and I could just fix the driver in some cases at least. And people did and provided solutions.

Windows also introduced software signing for drivers and user software, so yes, prevalence does increase while no benefits are provided. What is the advantage of an nvidia binary blob for Linux if it is signed or not. Was downloading from a trusted source a problem? We had hashes if you really wanted them.

With signing you make yourself dependent on the manufacturer of the OS.

The console argument was a hint that the application gained traction as a DRM utility, not for user security. Because signing here included properties of the medium the software was deployed on, so you couldn't just copy it to another CD/DVD.

Re: Apple revoked longtime Mac developer's code signing certificate with no warning

#143

Post by raxxorrax »

planb wrote: post #114
swiley wrote: post #93

Earlier quoted context omitted.

> Has the slippery slope caused any sliding yet? Yes it has, that’s the point of this whole discussion. Desktop apps now need to be signed and I’d be willing to bet you couldn’t find a signed torrent client.

Why would you be willing to bet if you just can google that the transmission BitTorrent client is in fact gatekeeper-signed?

Great example, they had problems with malware infection...

Re: Apple revoked longtime Mac developer's code signing certificate with no warning

#144

Post by ben509 »

raxxorrax wrote: post #15
pearjuice wrote: post #8

Sadly, this is what a walled garden results in. Please don't be surprised, shocked or even remotely discontent because by signing the ToS you have waived away any and all of your rights regarding the use and publishing of software in this walled garden. The only reason an issue like this will get "fixed" is when this (post/tweet) goes viral and the PR department will work extra hard to correct this.

Don't know why you are downvoted since you are completely correct. It is unfair to the developer but we wouldn't even have this discussion if people rejected app stores. I like that more developers just reject software certification processes. There is zero benefit aside from lock in.

The app store concept is fine, we need a separation of concerns. These are all things I'd rather pay someone else to do:

1. A standard format and tools for code signing.

1b. Actively validating or rejecting code on the end user machine.

2. Purchasing and delivery of software.

3. Trust of code blobs.

They just don't need to be all handled by your OS vendor.

Re: Apple revoked longtime Mac developer's code signing certificate with no warning

#145

Post by ben509 »

AlexandrB wrote: post #94
veidr wrote: post #70

Earlier quoted context omitted.

You are overstating the case pretty wildly. They have the ability to cause the OS to automatically delete binaries based on checksum, not "remotely login". And every time they have used this awesome power, it has been for good and everybody here would be OK with it. They have never used this power to like fuck with some developer because he violated YouTube TOS, or used the Taiwanese flag in his app, etc. As a user,…

> As a user, I want my OS vendor to have this power and use it appropriately. As a user I want my OS vendor to have this power with my explicit permission and use it appropriately. I also want to be able to disable this completely if I choose. There's going to come a time where Apple is going to do something shitty with this ability. Giving up the freedom to run whatever code you want on your computer is dangerous lo…

> There's going to come a time where Apple is going to do something shitty with this ability.

Or even "someone who has the credentials Apple owns" does it. I presume they require more than one employee to be able to do that, but that only mitigates insider threat, it doesn't remove it. They could also potentially be fooled into doing some damage.

Trust should never be evaluated on whether you think people are trustworthy; bad actors can impersonate, coerce or confuse good actors.

Re: Apple revoked longtime Mac developer's code signing certificate with no warning

#146

Post by echelon »

_qulr wrote: post #67
CubsFan1060 wrote: post #14

Perhaps we should wait to get the whole story to discuss. Didn't we just go through this with the "Apple doesn't return 30% on refund" fiasco last week?

Except that it's been verified by many people, and you can verify on your own Mac, that the developer's certificate has indeed been revoked, and the apps don't work. That part of the story is indisputable.

Honestly, no company should have this much control over people's businesses and livelihoods.

iPhone is a generic pocket computer, and it has outgrown Apple's desire to maintain a fiefdom.

Congress and the EU should force Apple to allow 3rd party marketplaces and installs. Apple is free to charge 30% for the App Store, but they can't be the only way to get code onto an iPhone. Nor should they be the only first class way of doing it.

Edit: I frequently get downvotes because of this. Who really wants everything going through Apple 100% of the time? I don't understand this perspective at all. Do you like not having control of your devices?

Re: Apple revoked longtime Mac developer's code signing certificate with no warning

#147

Post by Reason077 »

elliekelly wrote: post #131
ttepasse wrote: post #107

Earlier quoted context omitted.

Just wondered: If the adult content filter is ISP-level, can you deactivate it or like in this case report false positives? It sound's as orwellian as Apples certificate shenanigans.

I’m so curious how this even works? Do all ISPs in the UK have an adult content filter on by default?

I believe the majority of UK ISPs now have content filtering on by default for new connections, and all mobile providers have it.

This is a voluntary industry code of practice, not a legal requirement. For some time there was talk of making it law, enabled by a clause in the Digital Economy Act 2017, but this was abandoned in 2019.

https://en.wikipedia.org/wiki/Web_blocking_in_the_United_Kin...

Re: Apple revoked longtime Mac developer's code signing certificate with no warning

#148

Post by michaelt »

gruez wrote: post #18
CubsFan1060 wrote: post #14

Perhaps we should wait to get the whole story to discuss. Didn't we just go through this with the "Apple doesn't return 30% on refund" fiasco last week?

Isn't silently revoking a longtime mac developer's certificate bad practice, regardless of the justification?

If their certificate had been stolen it would be expected. Stolen code signing certificates have been used to distribute malware in the past; for example Stuxnet.

Of course, you'd expect a notification from Apple saying they'd done that and why; and plenty of safety measures to prevent mistaken revocations.

That's not to say I support mandatory code signing - merely that if you're going to have code signing, you also need to revoke stolen certs.

Re: Apple revoked longtime Mac developer's code signing certificate with no warning

#149

Post by _qulr »

ben509 wrote: post #138
planb wrote: post #121

Another possible explanation: The developer‘s certificate leaked and was really used to sign malware. Or his github repo was hacked and something evil was added to his code without him noticing. Maybe I’m just rationalizing, because if Apple is really going down the road that most commenters here suspect, then there will be no arm macbook for me unfortunately... :(

So, that's plausible, and I'm trying to wrap my head around the operational model of signed software. Apple has plenty of docs about it [1] [2] [3] but never really get into "how does this work in the long run when things fail" beyond blithely noting that users won't be able to run your stuff and a host of services will start to fail. A CRL / OSCP makes sense, more or less, for websites as they can simply abandon a c…

> I'd definitely want to revoke it, but if there's a set of valid releases, it seems like you'd want to do a partial revocation, e.g. "valid until YYMMDD." Or have a blacklist / whitelist and mark known good releases.

Yes, this is definitely possible, and why Developer ID signing has a secure timestamp, as specified by the --timestamp flag to the /usr/bin/codesign tool.

When Panic's code signing cert was stolen, they revoked it after a certain date, but old versions of their apps continued to be valid and pass Gatekeeper.

Re: Apple revoked longtime Mac developer's code signing certificate with no warning

#150

Post by pearjuice »

zepto wrote: post #132
raxxorrax wrote: post #15

Earlier quoted context omitted.

Don't know why you are downvoted since you are completely correct. It is unfair to the developer but we wouldn't even have this discussion if people rejected app stores. I like that more developers just reject software certification processes. There is zero benefit aside from lock in.

The downvotes are because we actually don’t known the full story. And of course there are benefits beyond lock-in. Do you know about the attestation service Apple has introduced? How would you build such a thing on your own?

Regardless of the full story, fact is that Apple (and many other owners of walled gardens) can revoke your access to it at a moment's notice without any explanation whatsoever. The reason for them doing this might be there but they owe you no explanation whatsoever. Some public outrage will put them in a position to make a statement or reconsider their decision. But you have no right to it.
Post reply on HN