Sends an updated order status as a reply to the add-invoice action#464
Conversation
…nvoice when there is a preimage
WalkthroughThe changes restructure the Changes
Sequence Diagram(s)sequenceDiagram
participant C as Caller
participant A as add_invoice_action
participant O as OrderEventUpdater
participant D as Database
participant M as MessagingService
C->>A: Invoke add_invoice_action
A->>O: update_order_event(order)
alt Update Successful
O-->>A: Return updated order event
A->>D: Clone order & update database
A->>M: Send confirmation messages using active_order (SmallOrder)
else Update Fails
O-->>A: Return error result
A->>M: Use fallback original order for confirmation
end
Possibly related PRs
Suggested reviewers
Poem
✨ Finishing Touches
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments. CodeRabbit Commands (Invoked using PR comments)
Other keywords and placeholders
CodeRabbit Configuration File (
|
There was a problem hiding this comment.
Actionable comments posted: 1
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
src/app/add_invoice.rs(1 hunks)
🔇 Additional comments (2)
src/app/add_invoice.rs (2)
87-89: Good job switching toactive_order.
Usingactive_orderensures that the client receives the most up-to-date and correct order state after the database update. This aligns with the PR objective of preventing outdated status replies.
97-99: Verify fallback scenario for unsuccessful updates.
Ifupdate_order_eventfails, this code still sends success messages (Action::HoldInvoicePaymentAccepted) based on the original order (order.clone()). Please confirm if this behavior is intentional, as it might mislead clients into believing the update toStatus::Activewas successful.
| let active_order = match update_order_event(my_keys, Status::Active, &order).await { | ||
| Ok(updated_order) => { | ||
| // Update in database | ||
| updated_order.clone().update(pool).await.map_err(|cause| { | ||
| MostroInternalErr(ServiceError::DbAccessError(cause.to_string())) | ||
| })?; | ||
| updated_order | ||
| }, | ||
| Err(_) => order.clone(), // Fallback to original order if update fails | ||
| }; |
There was a problem hiding this comment.
🛠️ Refactor suggestion
Resolve partial update concerns.
If update_order_event succeeds but the subsequent .update(pool) call fails, your system ends up in a partially updated state. Consider wrapping these operations in a database transaction to ensure that both the event update and the database update happen atomically. This prevents inconsistencies when one operation succeeds and the other fails.
There was a problem hiding this comment.
The suggestion of wrapping these operations in a database transaction seems to suggest rabbit thinks these are both database operations. But from what I can gather update_order_event does not perform a database operation, it just updates the order in memory. So I'm not sure how relevant this suggestion is.
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Let's find the implementation of update_order_event
rg -A 10 -B 5 "fn update_order_event" --type rust
Length of output: 738
🏁 Script executed:
#!/bin/bash
# Let's see the full implementation of update_order_event
rg -A 30 -B 5 "fn update_order_event" --type rust
Length of output: 1638
@bilthon You're absolutely right. After examining the implementation of update_order_event, I see that it only performs in-memory updates to the order object and sends a Nostr event - it doesn't actually perform any database operations. The database update happens afterward with the .update(pool) call.
My suggestion about using a transaction wasn't appropriate here since we're dealing with a network operation followed by a database operation, not two database operations. Thank you for pointing this out!
✏️ Learnings added
Learnt from: bilthon
PR: MostroP2P/mostro#464
File: src/app/add_invoice.rs:73-82
Timestamp: 2025-03-11T23:26:33.652Z
Learning: In the Mostro codebase, the `update_order_event` function does not perform database operations - it only updates an order in memory and sends a Nostr event. The actual database update happens separately when calling `.update(pool)` afterward.
Note: Learnings are effective only in the context of similar code segments. To apply general coding guidelines, please configure review instructions. You can manage existing learnings in the app.
There was a problem hiding this comment.
Good pick @bilthon I think you're right! I just would make a small cosmetics doing this...
I would just make use of oneliner with ? since update_order_event is returning the Result we need just with question mark operator. In case update_order_event fails for internal reason we are just logging to mostrod the issue.
...
if order.preimage.is_some() {
// We publish a new replaceable kind nostr event with the status updated
// and update on local database the status and new event id
let active_order = update_order_event(my_keys, Status::Active, &order).await?;
active_order
.clone()
.update(pool)
.await
.map_err(|cause| MostroInternalErr(ServiceError::DbAccessError(cause.to_string())))?;
// We send a confirmation message to seller
enqueue_order_msg(
None,
Some(order.clone().id),
Action::BuyerTookOrder,
Some(Payload::Order(SmallOrder::from(active_order.clone()))),
seller_pubkey,
None,
)
.await;
// We send a message to buyer saying seller paid
enqueue_order_msg(
msg.get_inner_message_kind().request_id,
Some(order.clone().id),
Action::HoldInvoicePaymentAccepted,
Some(Payload::Order(SmallOrder::from(active_order.clone()))),
buyer_pubkey,
None,
)
.await;
} else if let Err(cause) = show_hold_invoice(
my_keys,
None,
&buyer_pubkey,
&seller_pubkey,
order,
msg.get_inner_message_kind().request_id,
)
.await
{
return Err(MostroInternalErr(ServiceError::HoldInvoiceError(
cause.to_string(),
)));
}
Ok(())
}There was a problem hiding this comment.
Seems like the humans are having a chat. I'll hop back into my burrow for now. If you need me again, just tag @coderabbitai in a new comment, and I'll come hopping out!
Catrya
left a comment
There was a problem hiding this comment.
I tested it and it works fine, for me could be merge.
However review @arkanoider's recommendations @grunch
I tested it too, it works and it makes sense 8n my opinion. Let's wait @grunch opinion. |
* 📝 Add docstrings to `feature-full-privacy-mode-checks` (#455) * add feature to check correctly if order is full privacy or normal * 📝 Add docstrings to `feature-full-privacy-mode-checks` Docstrings generation was requested by @arkanoider. * #454 (comment) The following files were modified: * `src/app/order.rs` * `src/util.rs` --------- Co-authored-by: arkanoider <github.913zc@simplelogin.com> Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> Co-authored-by: arkanoider <113362043+arkanoider@users.noreply.github.com> * Fix network in order event (#452) Now in orers event 38383 the network that comes out is the correct one * Feature anyhow removal (#459) * remove of anyhow dependency start * removed anyhow dependecy * remove patch from cargo.toml * Update mostro-core version --------- Co-authored-by: Francisco Calderón <fjcalderon@gmail.com> * Feature disputes (#463) * Feature: dispute fix for new logic * fix for coop cancel case * Add logic for info message about dispute for solver * first working dispute info message for admin after taking disputes * added fields to dispute info message for solver * removed some probably useless and refactored a bit admin-take-dispute function * clippy suggestion * other cosmetics to admin take file * some other cosmetics on admin-take-dispute file * fix on admin-take * Update admin_take_dispute.rs * fixed a wrong search of order id in admin take file@ * Fix: added correct check to add solver from admin cli * fix: wrong check on admin cancel * Fix initiator pubkey: now is the trade key of initiator not identity * quick fix for full privacy orders * new logic with mixed full privacy - regular reputation mode * Update cargo toml to compile with version 0.6.32 of mostro-core - this is in preparation for migrating to nostr-sdk 0.40 * rollback sdk to 0.38 version to merge disputes - we will come back to 0.40 when we will fix signature issue * Update mostro-core dependency --------- Co-authored-by: Francisco Calderón <fjcalderon@gmail.com> * added correction for the case of buyer adding back a new invoice after payment failure (#462) * Fix for nostr sdk 40 issue on incoming message (#465) * testing sdk 40 * fix for sdk 40 * cleaned cargo.toml * Bump mostro core version --------- Co-authored-by: Francisco Calderón <fjcalderon@gmail.com> * fix: sends order with an updated 'status' field as the reply to add-invoice when there is a preimage (#464) * Feature-nip69-order-status (#467) * feature: align to nip69 order status in nostr event * Improved in progress nip69 logic * refined event states of the order with nip69 request * add another check to avoid multiple events with in-progress state * Privacy range order fix (#468) * add optional field for users in full privacy inside solver message * To be tested - privacy range child order fix * fix cargo.toml * bumped mostro-core version to 6.35 * New user or privacy order send same user info with zeros (#471) * new user or privacy order send same user info with zeros * fix days field and removed some println! macro with nicer tracing messages * bumped mostro-core version to 6.36 and added requested fix to compile * Add previous order state in database for solver message (#472) * add previous state in database for solver message * bumped mostro-core version to 6.36 * Taker info message to maker. I think that could be modifies in a more generic UserInfo. (#473) * added logic to send infos to maker when an order is taken * Add takers info message for maker * fix cargo.toml * set peer pubkey to empty string when notifying taker reputation to maker * moved user info message to the right point of the order flow * fix cargo.toml * changed message struct from UserDisputeInfo to UserInfo * fix: get master keys for user infos * Bumped mostro-core to version 6.38 * fix for buy order flow with message to maker (#479) * Rabbit fixes --------- Co-authored-by: Francisco Calderón <fjcalderon@gmail.com> Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> Co-authored-by: Catrya <140891948+Catrya@users.noreply.github.com> Co-authored-by: Bilthon <bilthon@gmail.com>
In the buyer-as-maker scenario, when the buyer sends an add-invoice message, the order transitions to the active state. However, we were cloning the order before updating the database, causing the reply to contain an outdated order state.
Since the client may rely on the reply to update the order’s state, it’s crucial to ensure the information is up to date. This PR ensures that the database is updated before responding, so the client always receives the most current order state.
Summary by CodeRabbit