## Panic: nil pointer dereference causes haven-relay to crash

[girino]

Description

haven-relay crashes with a panic due to a nil pointer dereference while handling websocket requests.

Error output

2026/01/28 23:30:27 🔗 listening at 0.0.0.0:3355
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x18 pc=0x11768c8]

goroutine 185 [running]:
main.initRelays.func6(
    {0x196e998?, 0x4001787ec0?},
    {{0x0, 0x0, 0x0}, {0x40013bcf00, 0x4, 0x8}, {0x40001642c8, 0x1, ...}, ...}
)
/app/init.go:227 +0x88

github.com/fiatjaf/khatru.(*Relay).handleRequest(
    0x4000347008,
    {0x196e998, 0x4001787ec0},
    {0x4000c66290, 0x6},
    0x400170ffc0,
    0x40004038f0,
    {{0x0, 0x0, 0x0}, ...}
)
/go/pkg/mod/github.com/fiatjaf/khatru@v0.19.1/responding.go:31 +0x1e4

github.com/fiatjaf/khatru.(*Relay).HandleWebsocket.func2.2(
    {0x4000276800?, 0x40012aaa50?}
)
/go/pkg/mod/github.com/fiatjaf/khatru@v0.19.1/handlers.go:305 +0xf88

created by github.com/fiatjaf/khatru.(*Relay).HandleWebsocket.func2
/go/pkg/mod/github.com/fiatjaf/khatru@v0.19.1/handlers.go:152 +0x2c4

Location

• /app/init.go:227

Impact

The relay process terminates due to a SIGSEGV.

[aaccioly]

Thanks for the report @girino, is this always happening? If so, what OS are you using?

Can you also try to revert this commit over here? 9761e45

[aaccioly]

Humm... init.go:227 uses the WoT, maybe you are receiving DMs before it is fully initialized, I'll make sure that this can't happen (tomorrow :))

[girino]

Humm... init.go:227 uses the WoT, maybe you are receiving DMs before it is fully initialized, I'll make sure that this can't happen (tomorrow :))

yes, that's exactly the case, This happened when i updated my "production" haven. I could see in nginx that it was receiving several requests just before the panic.

[girino]

Thanks for the report @girino, is this always happening? If so, what OS are you using?

It happened several times in a row. Since it was configured to restart automatically, it eventually recovered.

[aaccioly]

@girino , how are you? @sergey3bv made a change that should initialize the WoT before initiliasing the relay. Can you double check if everything is working for you now? (I appreciate this isn't easy to reproduce but you likely have one of the more active relays out there).

[girino]

seems to be working now. thanks.