fix(pubky): new bindings that end the Android "You're offline" failures - #125
Merged
Conversation
Picks up pubky-core-ffi-fork@93e3245, which is two fixes: the upstream sync to
pubky 0.10, and a relay-client TLS config the fork adds on top. Both are needed
— measured on-device, neither alone stopped the errors.
Background. rustls-platform-verifier hard-fails revocation checking on Android:
Let's Encrypt no longer publishes an OCSP responder, Android's
PKIXRevocationChecker only tries OCSP, and the crate's Kotlin half turns the
resulting CertPathValidatorException into `StatusCode.Revoked`:
certificate was revoked: java.security.cert.CertPathValidatorException:
Certificate does not specify OCSP responder
That reaches Kotlin as `HTTP transport error`, which `isNetworkFailure()`
matches, so users were told they were offline on a working connection.
pubky 0.10 fixes this for the SDK's ICANN client. It does not cover pkarr's
relay client, which is built from reqwest's defaults — after the bump alone the
device still logged 8 `Revoked` failures, every one straight after "Resolving
homeserver … via PKARR", and Home failed to load. The fork now hands pkarr's
relay client the same webpki-roots-without-revocation config through
`ClientBuilder::reqwest_client`.
This is the third fix to the same user-visible symptom and the last of them.
#122 shipped the verifier's missing Kotlin half; without it nothing verified at
all. This makes what it verifies actually pass.
API change to absorb: 0.10's `signin` takes a `ClientId`, so `signUp`, `signIn`,
`startAuthFlow`, `put`, `putBytes` and `deleteFile` each gained a `clientId`
argument. Added as `LOOPKY_CLIENT_ID` / `loopkyClientId` — a domain string, as
the type wants, and deliberately not the application id or bundle identifier.
Byte-for-byte the same on both platforms and across build types: it travels to
the homeserver, so it names Loopky, not this build of it.
Verified on-device (Redmi Note 11, Android 13, release build), same account and
navigation each time:
after #122 alone 0 ClassNotFound, 8 "Revoked", Home failed to load
after 0.10 alone still 8 "Revoked" — the relay client
with these bindings 0 ClassNotFound, 0 TLS failures, 0 transport errors,
0 Loopky errors; session resolves in 1.8s, not 12
Regenerated libs are `llvm-strip`ped to match upstream, taking the release APK
from 121 MB to 108 MB. Still over Play's 100 MB ceiling — separate packaging
work, not this.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
This was referenced Aug 24, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Third and final fix to the same user-visible symptom. #122 shipped the verifier's missing Kotlin half — without it nothing verified at all. This makes what it verifies actually pass.
Picks up pubky-core-ffi-fork#2, which is two fixes: the upstream sync to pubky 0.10, and a relay-client TLS config the fork adds on top. Both are needed — measured on-device, neither alone stopped the errors.
Why
rustls-platform-verifierhard-fails revocation checking on Android: Let's Encrypt no longer publishes an OCSP responder, Android'sPKIXRevocationCheckeronly tries OCSP, and the crate's Kotlin half turns the resulting exception intoStatusCode.Revoked:That reaches Kotlin as
HTTP transport error, whichisNetworkFailure()matches — so users were told they were offline on a working connection. There is no version to upgrade to: the crate's own escape hatch isBuildConfig.TEST-gated and 0.7.0 is the latest release.pubky 0.10 fixes this for the SDK's ICANN client. It does not cover pkarr's relay client — after the bump alone the device still logged 8
Revokedfailures, every one straight afterResolving homeserver … via PKARR, and Home failed to load. The fork now hands that client the same webpki-roots-without-revocation config.API change to absorb
0.10's
signintakes aClientId, sosignUp,signIn,startAuthFlow,put,putBytesanddeleteFileeach gained aclientIdargument.Added as
LOOPKY_CLIENT_ID/loopkyClientId="loopky.app"— a domain string, as the type wants (its own example isfranky.pubky.app), and deliberately not the application id or bundle identifier. Byte-for-byte identical on both platforms and across build types: it travels to the homeserver, so it names Loopky, not this build of it.Verified on-device
Redmi Note 11, Android 13, release build, same account and navigation each time:
Session resolves in 1.8 s instead of failing for 12.
Size
Regenerated libs are
llvm-stripped to match upstream, taking the release APK from 121 MB to 108 MB. Still over Play's 100 MB APK ceiling — that is the separate packaging work, not this.🤖 Generated with Claude Code