> Somewhere in GitHub's codebase, there's an if-statement checking when a repository was created to decide which ID format to return.
I doubt it. That's the beauty of GraphQL — each object can store its ID however it wants, and the GraphQL layer encodes it in base64. Then when someone sends a request with a base64-encoded ID, there _might_ be an if-statement (or maybe it just does a lookup on the ID). If anything, the if-statement happens _after_ decoding the ID, not before encoding it.
There was never any if-statement that checked the time — before the migration, IDs were created only in the old format. After the migration, they were created in the new format.
> That repository ID (010:Repository2325298) had a clear structure: 010 is some type enum, followed by a colon, the word Repository, and then the database ID 2325298.
It's a classic length prefix. Repository has 10 chars, Tree has 4.
> GitHub's migration guide tells developers to treat the new IDs as opaque strings and treat them as references. However it was clear that there was some underlying structure to these IDs as we just saw with the bitmasking
Great, so now GitHub can't change the structure of their IDs without breaking this person's code. The lesson is that if you're designing an API and want an ID to be opaque you have to literally encrypt it. I find it really demoralizing as an API designer that I have to treat my API's consumers as adversaries who will knowingly and intentionally ignore guidance in the documentation like this.
You could also say, if I tell you something is an opaque identifier, and you introspect it, it's your problem if your code breaks. I told you not to do that.
You could but you would lose the performance benefits you were seeking by encoding information into the ID. But you could also use a randomized, proprietary base64 alphabet rather than properly encrypting the ID.
Encoding a type name into an ID is never really something I've viewed as being about performance. Think of it more like an area code, it's an essential part of the identifier that tells you how to interpret the rest of it.
XOR encryption is cheap and effective. Make the key the static string "IfYouCanReadThisYourCodeWillBreak" or something akin to that. That way, the key itself will serve as a final warning when (not if) the key gets cracked.
Symmetric encryption is computationally ~free, but most of them are conceptually complex. The purpose of encryption here isn't security, it's obfuscation in the service of dissuading people from depending on something they shouldn't, so using the absolutely simplest thing that could possibly work is a positive.
XOR with fixed key is trivially figure-out-able, defeating the purpose. Speck is simple enough that a working implementation is included within the wikipedia article, and most LLMs can oneshot it.
Yes, XOR is a real and fundamental primitive in cryptography, but a cryptographer may view the scheme you described as violating Kerckhoffs's second principle of "secrecy in key only" (sometimes phrased, "if you don't pass in a key, it is encoding and not encryption"). You could view your obscure phrase as a key, or you could view it as a constant in a proprietary, obscure algorithm (which would make it an encoding). There's room for interpretation there.
Note that this is not a one-time pad because we are using the same key material many times.
But this is somewhat pedantic on my part, it's a distinction without a difference in this specific case where we don't actually need secrecy. (In most other cases there would be an important difference.)
> Great, so now GitHub can't change the structure of their IDs without breaking this person's code.
And that is all the fault of the person who treated a documented opaque value as if it has some specific structure.
> The lesson is that if you're designing an API and want an ID to be opaque you have to literally encrypt it.
The lesson is that you should stop caring about breaking people’s code who go against the documentation this way. When it breaks you shrug. Their code was always buggy and it just happened to be working for them until then. You are not their dad. You are not responsible for their misfortune.
> I find it really demoralizing as an API designer that I have to treat my API's consumers as adversaries who will knowingly and intentionally ignore guidance in the documentation like this.
Sounds like you’ve maybe never actually run a service or API library at scale. There’s so many factors that go into a decision like that at a company that it’s never so simple. Is the person impacted influential? You’ve got a reputation hit if they negatively blog about how you screwed them after something was working for years. Is a customer who’s worth 10% of your annual revenue impacted? Bet your ass your management chain won’t let you do a breaking change / revert any you made by declaring an incident.
Even in OSS land, you risk alienating the community you’ve built if they’re meaningfully impact. You only do this if the impact is minimal or you don’t care about alienating anyone using your software.
> The lesson is that you should stop caring about breaking people’s code who go against the documentation this way. When it breaks you shrug. Their code was always buggy and it just happened to be working for them until then. You are not their dad. You are not responsible for their misfortune.
Sure, but good luck running a business with that mindset.
I think more important than worrying about people treating an opaque value as structured data, is wondering _why_ they're doing so. In the case of this blog post, all they wanted to do was construct a URL, which required the integer database ID. Just make sure you expose what people need, so they don't need to go digging.
Other than that, I agree with what others are saying. If people rely on some undocumented aspect of your IDs, it's on them if that breaks.
Can GitHub change their API response rate? Can they increase it? If they do, they’ll break my code ‘cause it expects to receive responses at least after 1200ms. Any faster than that and I get race conditions. I selected the 1200ms number by measuring response rates.
No, you would call me a moron and tell me to go pound sand.
> Great, so now GitHub can't change the structure of their IDs without breaking this person's code
OP can put the decoded IDs into a new column and ignore the structure in the future. The problem was presumably mass querying the Github API to get those numbers needed for functional URLs.
Literally how I designed all the public facing R2 tokens like multipart uploads. It’s also a security barrier because forging and stealing of said tokens is harder and any vulnerability has to be done with cooperation of your servers and can be quickly shut down if needed.
In database design typically it recommends giving out opaque natural keys, and keeping your monotonically increasing integer IDs secret and used internally.
Maybe. Until your natural key changes. Which happens. A lot.
Exposing a surrogate / generated key that is effectively meaningless seems to be wise. Maybe internally Youtube has an index number for all their videos, but they expose a reasonably meaningless coded value to their consumers.
1. You don't want third parties to know how many objects you have
2. You don't want folks to be able to iterate each object by incrementing the id
But if you have composite IDs like this, that doesn't matter. All objects that belong to a repository have the repository id inside them. Incrementing the id gives you more objects from the same repo. Incrementing the repo id gives you...a random object or nothing at all. And if your IDs include a little entropy or a timestamp, you've effectively kneecapped anyone who's trying to abuse this.
This makes no sense. I am developing a product in this space (https://codeinput.com) and GitHub API and GraphQl is a badly entangled mess but you don’t trick your way around ids.
Same for urls, you are supposed to get them directly from GitHub not construct them yourself as format can change and then you find yourself playing a refactor cat-and-mouse game.
Best you can do is an hourly/daily cache for the values.
I wouldn't decode them like this, it's fragile, and global node IDs are supposed to be opaque in GraphQL.
I see that GitHub exposes a `databaseId` field on many of their types (like PullRequest) - is that what you're looking for? [1]
Most GraphQL APIs that serve objects that implement the Node interface just base-64-encode the type name and the database ID, but I definitely wouldn't rely on that always being the case. You can read more about global IDs in GraphQL in the spec in [2].
Also, as pointed out below, Github's GraphQL types also include fields like `permalink` and `url` (and interfaces like `UniformResourceLocatable`) that probably save you from needing to construct it yourself.
The newer global node IDs (which can be forced via the 'X-Github-Next-Global-ID' header [1]) have a prefix indicating the "type" of object delimited by an underscore, then a base64 encoded msgpack payload. For most objects it contains just a version (starting at 0) followed by the numeric "databaseId" field, but some are more complex.
For example, my GitHub user [2] has the node ID "U_kgDOAAhEkg". Users are "U_" and then the remaining data decodes to: [0, 541842] which matches the numeric ID for my user accepted by the REST API [3].
You shouldn't rely on any of this implementation of course, instead just directly query the "databaseId" field from the GraphQL API where you need interoperability. And in the other direction the REST API returns the "node_id" field for the GraphQL API.
For folks who finds this interesting, you might also like [4] which details GitHub's ETag implementation for the REST API.
I had seen GitHub node IDs, although I had not used them or tried to decode them (although I could see they seem to be base64), since I only used the REST API, which reports node IDs but does not use them as input.
It looks like a good explanation of the node IDs, though. However, like another comment says, you should not rely on the format of node IDs.
1. The list of "scopes" are the object hierarchy that owns the resource. That lets you figure out which shard a resource should be in. You want all the resources for the same repository on the same shard, otherwise if you simply hash the id, one shard going down takes down much of your service since everything is spread more or less uniformly across shards.
2. The object identifier is at the end. That should be strictly increasing, so all the resources for the same scope are ordered in the DB. This is one of the benefits of uuid7.
3. The first element is almost certainly a version. If you do a migration like this, you don't want to rule out doing it again. If you're packing bits, it's nearly impossible to know what's in the data without an identifier, so without the version you might not be able know whether the id is new or old.
Another commenter mentioned that you should encrypt this data. Hard pass! Decrypting each id is decidedly slower than b64 decode. Moreover, if you're picking apart IDs, you're relying on an interface that was never made for you. There's nothing sensitive in there: you're just setting yourself up for a possible (probable?) world of pain in the future. GitHub doesn't have to stop you from shooting your foot off.
Moreover, encrypting the contents of the ID makes them sort randomly. This is to be avoided: it means similar/related objects are not stored near each other, and you can't do simple range scans over your data.
You could decrypt the ids on the way in and store both the unencrypted and encrypted versions in the DB, but why? That's a lot of complexity, effort, and resources to stop randos on the Internet from relying on an internal, non-sensitive data format.
As for the old IDs that are still appearing, they are almost certainly:
1. Sharded by their own id (i.e., users are sharded by user id, not repo id), so you don't need additional information. Use something like rendezvous hashing to choose the shard.
2. Got sharded before the new id format was developed, and it's just not worth the trouble to change
chatmasta | 3 hours ago
I doubt it. That's the beauty of GraphQL — each object can store its ID however it wants, and the GraphQL layer encodes it in base64. Then when someone sends a request with a base64-encoded ID, there _might_ be an if-statement (or maybe it just does a lookup on the ID). If anything, the if-statement happens _after_ decoding the ID, not before encoding it.
There was never any if-statement that checked the time — before the migration, IDs were created only in the old format. After the migration, they were created in the new format.
haileys | 2 hours ago
It's a classic length prefix. Repository has 10 chars, Tree has 4.
esafak | an hour ago
agwa | 2 hours ago
Great, so now GitHub can't change the structure of their IDs without breaking this person's code. The lesson is that if you're designing an API and want an ID to be opaque you have to literally encrypt it. I find it really demoralizing as an API designer that I have to treat my API's consumers as adversaries who will knowingly and intentionally ignore guidance in the documentation like this.
maxbond | 2 hours ago
lelandfe | 2 hours ago
vlovich123 | an hour ago
haileys | 2 hours ago
You don't need encryption, a global_id database column with a randomly generated ID will do.
maxbond | 2 hours ago
haileys | 2 hours ago
maxbond | 2 hours ago
pdpi | 2 hours ago
Retr0id | an hour ago
In this particular instance, Speck would be ideal since it supports a 96-bit block size https://en.wikipedia.org/wiki/Speck_(cipher)
pdpi | an hour ago
Retr0id | 28 minutes ago
maxbond | an hour ago
pdpi | an hour ago
maxbond | an hour ago
Note that this is not a one-time pad because we are using the same key material many times.
But this is somewhat pedantic on my part, it's a distinction without a difference in this specific case where we don't actually need secrecy. (In most other cases there would be an important difference.)
nwallin | 2 hours ago
krisoft | 2 hours ago
And that is all the fault of the person who treated a documented opaque value as if it has some specific structure.
> The lesson is that if you're designing an API and want an ID to be opaque you have to literally encrypt it.
The lesson is that you should stop caring about breaking people’s code who go against the documentation this way. When it breaks you shrug. Their code was always buggy and it just happened to be working for them until then. You are not their dad. You are not responsible for their misfortune.
> I find it really demoralizing as an API designer that I have to treat my API's consumers as adversaries who will knowingly and intentionally ignore guidance in the documentation like this.
You don’t have to.
vlovich123 | an hour ago
Even in OSS land, you risk alienating the community you’ve built if they’re meaningfully impact. You only do this if the impact is minimal or you don’t care about alienating anyone using your software.
irjustin | 39 minutes ago
What was the saying? When your scale is big enough, even your bugs have users.
raincole | 35 minutes ago
VScode once broke a very popular extension that used a private API. Microsoft (righteously) didn't bother to ask if the private API had users.
halestock | 59 minutes ago
Sure, but good luck running a business with that mindset.
perfmode | 2 hours ago
bigblind | 2 hours ago
Other than that, I agree with what others are saying. If people rely on some undocumented aspect of your IDs, it's on them if that breaks.
lijok | an hour ago
No, you would call me a moron and tell me to go pound sand.
Weird systems were never supported to begin with.
kevin_thibedeau | an hour ago
OP can put the decoded IDs into a new column and ignore the structure in the future. The problem was presumably mass querying the Github API to get those numbers needed for functional URLs.
cush | an hour ago
vlovich123 | an hour ago
whateveracct | an hour ago
ezyang | 2 hours ago
phibz | 2 hours ago
taftster | 2 hours ago
Exposing a surrogate / generated key that is effectively meaningless seems to be wise. Maybe internally Youtube has an index number for all their videos, but they expose a reasonably meaningless coded value to their consumers.
bastawhiz | 29 minutes ago
1. You don't want third parties to know how many objects you have
2. You don't want folks to be able to iterate each object by incrementing the id
But if you have composite IDs like this, that doesn't matter. All objects that belong to a repository have the repository id inside them. Incrementing the id gives you more objects from the same repo. Incrementing the repo id gives you...a random object or nothing at all. And if your IDs include a little entropy or a timestamp, you've effectively kneecapped anyone who's trying to abuse this.
csomar | an hour ago
There is actually a documented way to do it: https://docs.github.com/en/graphql/guides/using-global-node-...
Same for urls, you are supposed to get them directly from GitHub not construct them yourself as format can change and then you find yourself playing a refactor cat-and-mouse game.
Best you can do is an hourly/daily cache for the values.
siralonso | an hour ago
I see that GitHub exposes a `databaseId` field on many of their types (like PullRequest) - is that what you're looking for? [1]
Most GraphQL APIs that serve objects that implement the Node interface just base-64-encode the type name and the database ID, but I definitely wouldn't rely on that always being the case. You can read more about global IDs in GraphQL in the spec in [2].
[1] https://docs.github.com/en/graphql/reference/objects#pullreq... [2] https://graphql.org/learn/global-object-identification/
siralonso | an hour ago
innoying | an hour ago
For example, my GitHub user [2] has the node ID "U_kgDOAAhEkg". Users are "U_" and then the remaining data decodes to: [0, 541842] which matches the numeric ID for my user accepted by the REST API [3].
You shouldn't rely on any of this implementation of course, instead just directly query the "databaseId" field from the GraphQL API where you need interoperability. And in the other direction the REST API returns the "node_id" field for the GraphQL API.
For folks who finds this interesting, you might also like [4] which details GitHub's ETag implementation for the REST API.
[1] https://docs.github.com/en/graphql/guides/migrating-graphql-... [2] https://api.github.com/user/541842 [3] https://gchq.github.io/CyberChef/#recipe=Find_/_Replace(%7B'... [4] https://github.com/bored-engineer/github-conditional-http-tr...
zzo38computer | an hour ago
It looks like a good explanation of the node IDs, though. However, like another comment says, you should not rely on the format of node IDs.
bastawhiz | 37 minutes ago
2. The object identifier is at the end. That should be strictly increasing, so all the resources for the same scope are ordered in the DB. This is one of the benefits of uuid7.
3. The first element is almost certainly a version. If you do a migration like this, you don't want to rule out doing it again. If you're packing bits, it's nearly impossible to know what's in the data without an identifier, so without the version you might not be able know whether the id is new or old.
Another commenter mentioned that you should encrypt this data. Hard pass! Decrypting each id is decidedly slower than b64 decode. Moreover, if you're picking apart IDs, you're relying on an interface that was never made for you. There's nothing sensitive in there: you're just setting yourself up for a possible (probable?) world of pain in the future. GitHub doesn't have to stop you from shooting your foot off.
Moreover, encrypting the contents of the ID makes them sort randomly. This is to be avoided: it means similar/related objects are not stored near each other, and you can't do simple range scans over your data.
You could decrypt the ids on the way in and store both the unencrypted and encrypted versions in the DB, but why? That's a lot of complexity, effort, and resources to stop randos on the Internet from relying on an internal, non-sensitive data format.
As for the old IDs that are still appearing, they are almost certainly:
1. Sharded by their own id (i.e., users are sharded by user id, not repo id), so you don't need additional information. Use something like rendezvous hashing to choose the shard.
2. Got sharded before the new id format was developed, and it's just not worth the trouble to change