fix: compute relative timestamps from actual data#581
Merged
QuantumExplorer merged 3 commits intoFeb 17, 2026
Conversation
Replace hardcoded display strings like "Received: 1 day ago" with real relative timestamps computed from document createdAt/updatedAt fields using chrono::Utc::now(). Fixes dashpay#579
Contributor
📝 WalkthroughWalkthroughAdded a new Changes
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 2❌ Failed checks (2 warnings)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing touches
🧪 Generate unit tests (beta)
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. Comment |
Collaborator
Author
|
@coderabbitai review |
Contributor
✅ Actions performedReview triggered.
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Fix all issues with AI agents
Before applying any fix, first verify the finding against the current code and
decide whether a code change is actually needed. If the finding is not valid or
no change is required, do not modify code for that item and briefly explain why
it was skipped.
In `@src/ui/dashpay/contact_requests.rs`:
- Around line 26-51: The function `format_relative_time` is duplicated (in
contact_requests.rs and send_payment.rs); extract it into a shared utility
(e.g., add it to `src/ui/dashpay/mod.rs` or a new `src/ui/helpers.rs`) as a
single pub function (pub fn format_relative_time(...)) and re-export or import
it from both `contact_requests.rs` and `send_payment.rs`; update both files to
remove their local copies and call the shared `format_relative_time`, ensuring
visibility (pub/use) and tests/build still compile and preserving the existing
behavior (including handling of `LocalResult` and the millis threshold).
In `@src/ui/dashpay/send_payment.rs`:
- Around line 797-806: DashPayPaymentHistory currently lacks a timestamp so
format_relative_time(payment.timestamp) gets zeros; update the backend
tuple/type to include a timestamp field (e.g., extend DashPayPaymentHistory from
a 5-tuple to include timestamp or convert to a struct with a timestamp), change
any constructors that set timestamp: 0 in display_task_result and elsewhere to
populate the real timestamp from the fetched payment record, and adjust all
callsites (including the UI code that builds payment_time_text and the consumer
of DashPayPaymentHistory) to read the new timestamp field so
format_relative_time receives real timestamps instead of zero.
- Around line 38-42: The u64->i64 casts for timestamp used when building
dt_result (via Utc.timestamp_millis_opt and Utc.timestamp_opt) can wrap on
extreme values; replace the direct casts with a checked conversion using
i64::try_from(timestamp).ok()? and use that converted i64 for both
timestamp_millis_opt / timestamp_opt branches (adjusting control flow so the
early ? return is allowed — the enclosing function already returns Option),
ensuring you reference the converted value instead of doing timestamp as i64.
- Around line 34-54: The function `format_relative_time` is duplicated; extract
it into a shared utility module (e.g., create a `helpers` or `dashpay::util`
module) and move the implementation there, keeping the same signature `fn
format_relative_time(timestamp: u64) -> Option<String>`; then replace the local
implementations in `send_payment.rs` and `contact_requests.rs` with
`use`/`pub(crate) use` or an import from the new module and call the single
implementation instead so both files reference the shared function.
🧹 Nitpick comments (2)
🤖 Fix all nitpicks with AI agents
Before applying any fix, first verify the finding against the current code and decide whether a code change is actually needed. If the finding is not valid or no change is required, do not modify code for that item and briefly explain why it was skipped. In `@src/ui/dashpay/send_payment.rs`: - Around line 38-42: The u64->i64 casts for timestamp used when building dt_result (via Utc.timestamp_millis_opt and Utc.timestamp_opt) can wrap on extreme values; replace the direct casts with a checked conversion using i64::try_from(timestamp).ok()? and use that converted i64 for both timestamp_millis_opt / timestamp_opt branches (adjusting control flow so the early ? return is allowed — the enclosing function already returns Option), ensuring you reference the converted value instead of doing timestamp as i64. - Around line 34-54: The function `format_relative_time` is duplicated; extract it into a shared utility module (e.g., create a `helpers` or `dashpay::util` module) and move the implementation there, keeping the same signature `fn format_relative_time(timestamp: u64) -> Option<String>`; then replace the local implementations in `send_payment.rs` and `contact_requests.rs` with `use`/`pub(crate) use` or an import from the new module and call the single implementation instead so both files reference the shared function.src/ui/dashpay/send_payment.rs (2)
38-42:u64 as i64cast can silently wrap on extreme values.The
timestamp as i64casts on lines 39 and 41 will wrap to negative if au64exceedsi64::MAX. While realistic epoch timestamps won't hit this, a defensivei64::try_from(timestamp).ok()?is cheap and prevents surprises from corrupt data.Proposed fix
- Utc.timestamp_millis_opt(timestamp as i64) + Utc.timestamp_millis_opt(i64::try_from(timestamp).ok()?) } else { - Utc.timestamp_opt(timestamp as i64, 0) + Utc.timestamp_opt(i64::try_from(timestamp).ok()?, 0) };Note: this requires a small restructure since the early
?return needs the enclosing function to returnOption, which it already does.🤖 Prompt for AI Agents
Before applying any fix, first verify the finding against the current code and decide whether a code change is actually needed. If the finding is not valid or no change is required, do not modify code for that item and briefly explain why it was skipped. In `@src/ui/dashpay/send_payment.rs` around lines 38 - 42, The u64->i64 casts for timestamp used when building dt_result (via Utc.timestamp_millis_opt and Utc.timestamp_opt) can wrap on extreme values; replace the direct casts with a checked conversion using i64::try_from(timestamp).ok()? and use that converted i64 for both timestamp_millis_opt / timestamp_opt branches (adjusting control flow so the early ? return is allowed — the enclosing function already returns Option), ensuring you reference the converted value instead of doing timestamp as i64.
34-54: Duplicatedformat_relative_time— extract to a shared module.This function is identical to the one in
src/ui/dashpay/contact_requests.rs(lines 28-50). Having the same logic in two places means future changes (e.g., adjusting the "just now" threshold or adding more granularity) must be applied in both locations.Consider moving it to a common utility module (e.g.,
src/ui/helpers.rsor a newsrc/ui/dashpay/util.rs) and importing it from both files.🤖 Prompt for AI Agents
Before applying any fix, first verify the finding against the current code and decide whether a code change is actually needed. If the finding is not valid or no change is required, do not modify code for that item and briefly explain why it was skipped. In `@src/ui/dashpay/send_payment.rs` around lines 34 - 54, The function `format_relative_time` is duplicated; extract it into a shared utility module (e.g., create a `helpers` or `dashpay::util` module) and move the implementation there, keeping the same signature `fn format_relative_time(timestamp: u64) -> Option<String>`; then replace the local implementations in `send_payment.rs` and `contact_requests.rs` with `use`/`pub(crate) use` or an import from the new module and call the single implementation instead so both files reference the shared function.
Comment on lines
+797
to
+806
| let payment_time_text = format_relative_time(payment.timestamp) | ||
| .map(|t| format!("• {}", t)) | ||
| .unwrap_or_default(); | ||
| if !payment_time_text.is_empty() { | ||
| ui.label( | ||
| RichText::new(payment_time_text) | ||
| .small() | ||
| .color(DashColors::text_secondary(dark_mode)), | ||
| ); | ||
| } |
Contributor
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# First, let's check the file exists and examine lines around 856 and 889
wc -l src/ui/dashpay/send_payment.rsRepository: dashpay/dash-evo-tool
Length of output: 99
🏁 Script executed:
#!/bin/bash
# Check lines 856 and 889 to see if they hardcode timestamp: 0
sed -n '850,860p' src/ui/dashpay/send_payment.rs
echo "---"
sed -n '883,893p' src/ui/dashpay/send_payment.rsRepository: dashpay/dash-evo-tool
Length of output: 1172
🏁 Script executed:
#!/bin/bash
# Search for format_relative_time implementation to see how it handles values
rg -n "fn format_relative_time" --type=rust -A 10Repository: dashpay/dash-evo-tool
Length of output: 1655
🏁 Script executed:
#!/bin/bash
# Look for DashPayPaymentHistory structure and construction
rg -n "DashPayPaymentHistory" --type=rust -B 2 -A 8Repository: dashpay/dash-evo-tool
Length of output: 2863
🏁 Script executed:
#!/bin/bash
# Check how payments are created in display_task_result (lines 856 and 889 context)
sed -n '840,870p' src/ui/dashpay/send_payment.rs
echo "---"
sed -n '870,910p' src/ui/dashpay/send_payment.rsRepository: dashpay/dash-evo-tool
Length of output: 2946
Backend payment history lacks timestamp data structure to support relative timestamps.
Both lines 856 and 889 hardcode timestamp: 0 with a TODO comment. Since format_relative_time() returns None for zero timestamps, the relative timestamp label is silently omitted for all payments fetched via display_task_result(). This partially undermines the fix from issue #579—users only see relative timestamps for payments loaded from the local database, not freshly fetched ones.
The root cause is that DashPayPaymentHistory in src/backend_task/mod.rs:161 is defined as a 5-tuple (tx_id, contact_name, amount, is_incoming, memo) without a timestamp field, so there is no mechanism to propagate real timestamps from the backend. Addressing this requires extending the tuple structure to include timestamp and updating the backend to fetch and return actual payment timestamps.
🤖 Prompt for AI Agents
Before applying any fix, first verify the finding against the current code and
decide whether a code change is actually needed. If the finding is not valid or
no change is required, do not modify code for that item and briefly explain why
it was skipped.
In `@src/ui/dashpay/send_payment.rs` around lines 797 - 806, DashPayPaymentHistory
currently lacks a timestamp so format_relative_time(payment.timestamp) gets
zeros; update the backend tuple/type to include a timestamp field (e.g., extend
DashPayPaymentHistory from a 5-tuple to include timestamp or convert to a struct
with a timestamp), change any constructors that set timestamp: 0 in
display_task_result and elsewhere to populate the real timestamp from the
fetched payment record, and adjust all callsites (including the UI code that
builds payment_time_text and the consumer of DashPayPaymentHistory) to read the
new timestamp field so format_relative_time receives real timestamps instead of
zero.
Deduplicate the identical format_relative_time function that existed in both contact_requests.rs and send_payment.rs. Move it to the dashpay mod.rs as a pub(crate) function and import from both files.
Collaborator
Author
|
@coderabbitai review |
Contributor
✅ Actions performedReview triggered.
|
Contributor
There was a problem hiding this comment.
🤖 Fix all issues with AI agents
Verify each finding against the current code and only fix it if needed.
In `@src/ui/dashpay/mod.rs`:
- Around line 32-36: The code casts timestamp: u64 to i64 directly when calling
Utc.timestamp_millis_opt / Utc.timestamp_opt which can wrap for values >
i64::MAX; replace the direct casts with a safe conversion (e.g. use
i64::try_from(timestamp) or timestamp.checked_sub(...) pattern) and handle the
Err/overflow case explicitly before computing dt_result (for example return None
or clamp to i64::MAX), updating the branch that constructs dt_result (the
Utc.timestamp_millis_opt / Utc.timestamp_opt calls) to use the safely converted
i64 value instead of (timestamp as i64).
- Around line 26-48: The function format_relative_time is currently placed
between import blocks; move the entire fn format_relative_time(...)
implementation so all use statements (the import block around lines ~21–22 and
the later import ending at ~49) are contiguous and the function appears after
the imports. Ensure you update the file so only import declarations appear
together and the format_relative_time function (and any duplicate placement
referenced at 49-49) is relocated below the final import to restore clean module
organization.
- Around line 26-48: Add inline unit tests for the pure utility function
format_relative_time: create a #[cfg(test)] mod with multiple #[test] cases that
assert (1) timestamp == 0 returns None, (2) a small seconds-range timestamp
(e.g., now - few seconds) yields Some("just now"), (3) a known past timestamp
returns a non-empty human-readable Some(String) matching expected substring, and
(4) a value just below/above the 1_000_000_000_000 threshold exercises both the
seconds and milliseconds branches; locate tests alongside format_relative_time
in src/ui/dashpay/mod.rs and use deterministic fixed timestamps (or compute from
Utc::now()) so they are reliable.
🧹 Nitpick comments (3)
🤖 Fix all nitpicks with AI agents
Verify each finding against the current code and only fix it if needed. In `@src/ui/dashpay/mod.rs`: - Around line 32-36: The code casts timestamp: u64 to i64 directly when calling Utc.timestamp_millis_opt / Utc.timestamp_opt which can wrap for values > i64::MAX; replace the direct casts with a safe conversion (e.g. use i64::try_from(timestamp) or timestamp.checked_sub(...) pattern) and handle the Err/overflow case explicitly before computing dt_result (for example return None or clamp to i64::MAX), updating the branch that constructs dt_result (the Utc.timestamp_millis_opt / Utc.timestamp_opt calls) to use the safely converted i64 value instead of (timestamp as i64). - Around line 26-48: The function format_relative_time is currently placed between import blocks; move the entire fn format_relative_time(...) implementation so all use statements (the import block around lines ~21–22 and the later import ending at ~49) are contiguous and the function appears after the imports. Ensure you update the file so only import declarations appear together and the format_relative_time function (and any duplicate placement referenced at 49-49) is relocated below the final import to restore clean module organization. - Around line 26-48: Add inline unit tests for the pure utility function format_relative_time: create a #[cfg(test)] mod with multiple #[test] cases that assert (1) timestamp == 0 returns None, (2) a small seconds-range timestamp (e.g., now - few seconds) yields Some("just now"), (3) a known past timestamp returns a non-empty human-readable Some(String) matching expected substring, and (4) a value just below/above the 1_000_000_000_000 threshold exercises both the seconds and milliseconds branches; locate tests alongside format_relative_time in src/ui/dashpay/mod.rs and use deterministic fixed timestamps (or compute from Utc::now()) so they are reliable.src/ui/dashpay/mod.rs (3)
32-36:u64 as i64cast can silently wrap for values >i64::MAX.Realistic timestamps won't hit this, but a defensive guard is cheap and prevents surprising negative-date results from corrupt data.
Proposed fix
- let dt_result = if timestamp > 1_000_000_000_000 { - Utc.timestamp_millis_opt(timestamp as i64) + let Ok(ts) = i64::try_from(timestamp) else { + return None; + }; + let dt_result = if timestamp > 1_000_000_000_000 { + Utc.timestamp_millis_opt(ts) } else { - Utc.timestamp_opt(timestamp as i64, 0) + Utc.timestamp_opt(ts, 0) };🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/ui/dashpay/mod.rs` around lines 32 - 36, The code casts timestamp: u64 to i64 directly when calling Utc.timestamp_millis_opt / Utc.timestamp_opt which can wrap for values > i64::MAX; replace the direct casts with a safe conversion (e.g. use i64::try_from(timestamp) or timestamp.checked_sub(...) pattern) and handle the Err/overflow case explicitly before computing dt_result (for example return None or clamp to i64::MAX), updating the branch that constructs dt_result (the Utc.timestamp_millis_opt / Utc.timestamp_opt calls) to use the safely converted i64 value instead of (timestamp as i64).
26-48: Function body is placed between two import blocks.Lines 26–48 define
format_relative_timesandwiched betweenusestatements (line 21–22 and line 49). Moving the function below all imports (after line 50) improves readability.Also applies to: 49-49
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/ui/dashpay/mod.rs` around lines 26 - 48, The function format_relative_time is currently placed between import blocks; move the entire fn format_relative_time(...) implementation so all use statements (the import block around lines ~21–22 and the later import ending at ~49) are contiguous and the function appears after the imports. Ensure you update the file so only import declarations appear together and the format_relative_time function (and any duplicate placement referenced at 49-49) is relocated below the final import to restore clean module organization.
26-48: Consider adding inline unit tests for this shared utility.
format_relative_timeis a pure function now shared across multiple UI files — it's a good candidate for inline#[cfg(test)]coverage (e.g., zero →None, seconds-range →"just now", known past timestamp → reasonable relative string, millis-threshold boundary).As per coding guidelines: "Unit tests should be inline in source files using
#[test]attribute."🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/ui/dashpay/mod.rs` around lines 26 - 48, Add inline unit tests for the pure utility function format_relative_time: create a #[cfg(test)] mod with multiple #[test] cases that assert (1) timestamp == 0 returns None, (2) a small seconds-range timestamp (e.g., now - few seconds) yields Some("just now"), (3) a known past timestamp returns a non-empty human-readable Some(String) matching expected substring, and (4) a value just below/above the 1_000_000_000_000 threshold exercises both the seconds and milliseconds branches; locate tests alongside format_relative_time in src/ui/dashpay/mod.rs and use deterministic fixed timestamps (or compute from Utc::now()) so they are reliable.
QuantumExplorer
approved these changes
Feb 17, 2026
QuantumExplorer
pushed a commit
that referenced
this pull request
Feb 17, 2026
* fix: compute relative timestamps from actual data (#581) * fix: compute relative timestamps from actual data Replace hardcoded display strings like "Received: 1 day ago" with real relative timestamps computed from document createdAt/updatedAt fields using chrono::Utc::now(). Fixes #579 * style: fix import ordering per cargo fmt * refactor: extract format_relative_time to shared dashpay module Deduplicate the identical format_relative_time function that existed in both contact_requests.rs and send_payment.rs. Move it to the dashpay mod.rs as a pub(crate) function and import from both files. --------- Co-authored-by: PastaClaw <thepastaclaw@users.noreply.github.com> * fix: update platform for DIP-18 HRP and improve SPV sync progress (#575) * fix: update dashpay/platform to d6f4eb9 for DIP-18 HRP fix The previous platform revision used incorrect bech32m HRP prefixes (evo/tevo) for Platform addresses. The updated commit uses the correct DIP-0018 prefixes (dash/tdash). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * refactor(spv): use dash-spv SyncProgress API directly Replace the intermediate SyncProgress/DetailedSyncProgress/SyncStage translation layer with direct use of dash_spv::sync::SyncProgress. This eliminates ~120 lines of bridge code in spawn_progress_watcher() and determine_sync_stage(), and lets the UI query per-manager progress (headers, filter_headers, filters, masternodes, blocks) via the upstream API. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): improve sync progress accuracy and robustness - Log headers() error before discarding for easier debugging - Restore download-window optimization for headers progress on checkpoint-resumed syncs (progress starts near 0% not 83%) - Replace catch-all _ => 0.0 with explicit SyncState variant matches so new variants produce compile errors - Fix filters progress to use current_height/target_height instead of downloaded/target_height (session count vs absolute height mismatch) - Restore peer count in sync status text - Add "Querying peer heights" label and diffs_processed to masternode status for more informative sync messages Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): use height-based blocks progress to prevent bar from jumping The blocks progress bar was bouncing backward because it used processed/requested session counters whose denominator grows as filters discover more matching blocks. Switch to last_processed block height relative to headers target_height, which only increases. Display "current / target" heights instead of percentage on the blocks bar. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): improve progress bar resilience across resume and network switch - Add windowed progress tracking for filters (consistent with headers/filter_headers) - Reset blocks_stage_start on Error state so recovery gets a fresh window - Track spv_progress_network to detect network changes and rebuild progress state from the new network's sync_progress instead of resetting to zero - Eliminate redundant clone in progress watcher (move instead of clone twice) - Consolidate all progress state reset logic into rebuild_spv_progress_state() Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * build: add git and cargo permissions to Claude Code workflow (#565) * build: add git and cargo permissions to Claude Code workflow Allow Claude to run git fetch/merge/checkout/rebase/push and cargo build/test/clippy/fmt commands. Switch model to opus. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: use claude_args for model and allowed tools The `model` and `allowed_tools` inputs are not declared in claude-code-action@v1 and are silently ignored. Move them to `claude_args` with --model and --allowedTools flags. Also fix deprecated colon syntax (`:*`) to space syntax (` *`). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: use single quotes to prevent glob expansion in allowed tools Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * build: sandbox cargo commands and add git permissions for Claude - Add safe-cargo.sh wrapper that strips CI secrets before running cargo - Use --allowedTools for git and safe-cargo, --disallowedTools for raw cargo - Document safe-cargo usage in CLAUDE.md Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: switch safe-cargo.sh from denylist to allowlist approach Use `env -i` (start with empty environment, explicitly pass only what cargo needs) instead of `env -u` (strip known secrets). This is more robust against future secrets being added to the workflow. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * build: deny Claude from editing CI scripts and workflows Prevent Claude from modifying .github/scripts/ and .github/workflows/ to ensure the safe-cargo wrapper cannot be tampered with. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: conditionally pass optional env vars in safe-cargo.sh Empty PROTOC="" caused prost build scripts to fail with "protoc not found". Now optional vars (PROTOC, CC, CXX, etc.) are only passed when set and non-empty. Tested: build, test, fmt all pass through the wrapper. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * rabbit feedback * build: move safe-cargo.sh to scripts/ and allow +nightly fmt Move safe-cargo.sh from .github/scripts/ to top-level scripts/ for better discoverability. Add detailed comment explaining why the wrapper exists (prevent CI secret exfiltration via build scripts). Update all references in claude.yml, CLAUDE.md, and permission settings. Add `+nightly fmt` to allowedTools so Claude can follow CLAUDE.md formatting instructions in CI. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: PastaClaw <thepastaclaw@users.noreply.github.com> Co-authored-by: lklimek <842586+lklimek@users.noreply.github.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
lklimek
added a commit
that referenced
this pull request
Feb 17, 2026
…d UX (#577) * fix: update dashpay/platform to d6f4eb9 for DIP-18 HRP fix The previous platform revision used incorrect bech32m HRP prefixes (evo/tevo) for Platform addresses. The updated commit uses the correct DIP-0018 prefixes (dash/tdash). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * refactor(spv): use dash-spv SyncProgress API directly Replace the intermediate SyncProgress/DetailedSyncProgress/SyncStage translation layer with direct use of dash_spv::sync::SyncProgress. This eliminates ~120 lines of bridge code in spawn_progress_watcher() and determine_sync_stage(), and lets the UI query per-manager progress (headers, filter_headers, filters, masternodes, blocks) via the upstream API. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): improve sync progress accuracy and robustness - Log headers() error before discarding for easier debugging - Restore download-window optimization for headers progress on checkpoint-resumed syncs (progress starts near 0% not 83%) - Replace catch-all _ => 0.0 with explicit SyncState variant matches so new variants produce compile errors - Fix filters progress to use current_height/target_height instead of downloaded/target_height (session count vs absolute height mismatch) - Restore peer count in sync status text - Add "Querying peer heights" label and diffs_processed to masternode status for more informative sync messages Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): use height-based blocks progress to prevent bar from jumping The blocks progress bar was bouncing backward because it used processed/requested session counters whose denominator grows as filters discover more matching blocks. Switch to last_processed block height relative to headers target_height, which only increases. Display "current / target" heights instead of percentage on the blocks bar. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): improve progress bar resilience across resume and network switch - Add windowed progress tracking for filters (consistent with headers/filter_headers) - Reset blocks_stage_start on Error state so recovery gets a fresh window - Track spv_progress_network to detect network changes and rebuild progress state from the new network's sync_progress instead of resetting to zero - Eliminate redundant clone in progress watcher (move instead of clone twice) - Consolidate all progress state reset logic into rebuild_spv_progress_state() Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): update ConnectionStatus immediately on connect/disconnect Propagate SPV status into ConnectionStatus right after start_spv() and stop_spv() so the UI reflects the change on the next frame instead of waiting for the next throttled trigger_refresh() cycle (2-10 seconds). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): check cancellation token in listener tasks to prevent shutdown hang SPV finality and reconcile listeners blocked app shutdown for up to 10s because they never checked the global cancellation token. Add cancel branches to their tokio::select! loops so they exit immediately on shutdown. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * chore: remove log spamming line * fix(spv): make listener handler calls cancellation-aware and add shutdown trace logging Wrap reconcile_spv_wallets(), handle_spv_finality_event(), and the debounce sleep inside tokio::select! with the cancellation token so shutdown can interrupt them even when blocked on locks held by the SPV sync thread. Add trace-level logging to TaskManager::shutdown() for per-task join timing to aid future shutdown diagnostics. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): disable Disconnect button during stopping and poll status faster - Disable the Disconnect button while SPV status is Stopping to prevent double-clicks and provide visual feedback - Poll SPV status every 200ms during Stopping instead of the 10s connected interval so the Stopped transition is reflected within 1s - Reset the throttle timer in stop_spv() so fast polling starts immediately Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * refactor(tasks): add mandatory task names to spawn_sync for shutdown diagnostics Change spawn_sync() signature to require a &'static str name. The JoinSet now yields the task name on completion, letting shutdown() log which tasks finished and which ones timed out. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * chore: log task names * refactor(spv): eliminate separate OS thread and tokio runtime Replace the dedicated SPV thread + 4-worker tokio runtime with a spawned task on the main 12-worker runtime via TaskManager::spawn_sync. This simplifies shutdown (SPV loop now tracked in unified JoinSet), removes cross-runtime complexity, and improves debuggability. Additional changes: - Fix: zeroize xprv_str after wallet import (security H-1) - Fix: sanitize devnet_name in build_spv_data_dir to prevent path traversal - Add 21 integration tests covering lifecycle, concurrency, deadlock detection, and live testnet sync Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): derive stop_token as child of global cancel for clean shutdown The spv_request_handler selected on stop_token, which was independent of the global cancellation token. During window-close shutdown only the global token was cancelled, leaving the request handler running and causing a ~5s hang until the TaskManager timeout aborted it. Fix by creating stop_token as a child_token() of the global cancel. This also simplifies run_spv_loop and run_sync_and_monitor by removing the redundant global_cancel parameter — a single stop_token now covers both explicit SpvManager::stop() and application-wide shutdown. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * CLAUDE: minor changes about docs * fix(spv): bump rust-dashcore to d8bc066 for faster disconnect and add shutdown tracing Pin rust-dashcore patches to commit d8bc066 which includes: - sync manager task loop exits on network errors instead of logging indefinitely - sync coordinator signal_shutdown() cancels tasks before network disconnect - connection tasks race Peer::connect() against shutdown token Add debug tracing to SpvManager::stop() and run_sync_and_monitor() to measure client.stop() duration, and trace-level polling of SPV status in ConnectionStatus. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * chore: update rust-dashcore rev * merge: resolve v1.0-dev conflicts for #577 (#585) * fix: compute relative timestamps from actual data (#581) * fix: compute relative timestamps from actual data Replace hardcoded display strings like "Received: 1 day ago" with real relative timestamps computed from document createdAt/updatedAt fields using chrono::Utc::now(). Fixes #579 * style: fix import ordering per cargo fmt * refactor: extract format_relative_time to shared dashpay module Deduplicate the identical format_relative_time function that existed in both contact_requests.rs and send_payment.rs. Move it to the dashpay mod.rs as a pub(crate) function and import from both files. --------- Co-authored-by: PastaClaw <thepastaclaw@users.noreply.github.com> * fix: update platform for DIP-18 HRP and improve SPV sync progress (#575) * fix: update dashpay/platform to d6f4eb9 for DIP-18 HRP fix The previous platform revision used incorrect bech32m HRP prefixes (evo/tevo) for Platform addresses. The updated commit uses the correct DIP-0018 prefixes (dash/tdash). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * refactor(spv): use dash-spv SyncProgress API directly Replace the intermediate SyncProgress/DetailedSyncProgress/SyncStage translation layer with direct use of dash_spv::sync::SyncProgress. This eliminates ~120 lines of bridge code in spawn_progress_watcher() and determine_sync_stage(), and lets the UI query per-manager progress (headers, filter_headers, filters, masternodes, blocks) via the upstream API. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): improve sync progress accuracy and robustness - Log headers() error before discarding for easier debugging - Restore download-window optimization for headers progress on checkpoint-resumed syncs (progress starts near 0% not 83%) - Replace catch-all _ => 0.0 with explicit SyncState variant matches so new variants produce compile errors - Fix filters progress to use current_height/target_height instead of downloaded/target_height (session count vs absolute height mismatch) - Restore peer count in sync status text - Add "Querying peer heights" label and diffs_processed to masternode status for more informative sync messages Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): use height-based blocks progress to prevent bar from jumping The blocks progress bar was bouncing backward because it used processed/requested session counters whose denominator grows as filters discover more matching blocks. Switch to last_processed block height relative to headers target_height, which only increases. Display "current / target" heights instead of percentage on the blocks bar. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): improve progress bar resilience across resume and network switch - Add windowed progress tracking for filters (consistent with headers/filter_headers) - Reset blocks_stage_start on Error state so recovery gets a fresh window - Track spv_progress_network to detect network changes and rebuild progress state from the new network's sync_progress instead of resetting to zero - Eliminate redundant clone in progress watcher (move instead of clone twice) - Consolidate all progress state reset logic into rebuild_spv_progress_state() Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * build: add git and cargo permissions to Claude Code workflow (#565) * build: add git and cargo permissions to Claude Code workflow Allow Claude to run git fetch/merge/checkout/rebase/push and cargo build/test/clippy/fmt commands. Switch model to opus. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: use claude_args for model and allowed tools The `model` and `allowed_tools` inputs are not declared in claude-code-action@v1 and are silently ignored. Move them to `claude_args` with --model and --allowedTools flags. Also fix deprecated colon syntax (`:*`) to space syntax (` *`). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: use single quotes to prevent glob expansion in allowed tools Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * build: sandbox cargo commands and add git permissions for Claude - Add safe-cargo.sh wrapper that strips CI secrets before running cargo - Use --allowedTools for git and safe-cargo, --disallowedTools for raw cargo - Document safe-cargo usage in CLAUDE.md Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: switch safe-cargo.sh from denylist to allowlist approach Use `env -i` (start with empty environment, explicitly pass only what cargo needs) instead of `env -u` (strip known secrets). This is more robust against future secrets being added to the workflow. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * build: deny Claude from editing CI scripts and workflows Prevent Claude from modifying .github/scripts/ and .github/workflows/ to ensure the safe-cargo wrapper cannot be tampered with. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: conditionally pass optional env vars in safe-cargo.sh Empty PROTOC="" caused prost build scripts to fail with "protoc not found". Now optional vars (PROTOC, CC, CXX, etc.) are only passed when set and non-empty. Tested: build, test, fmt all pass through the wrapper. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * rabbit feedback * build: move safe-cargo.sh to scripts/ and allow +nightly fmt Move safe-cargo.sh from .github/scripts/ to top-level scripts/ for better discoverability. Add detailed comment explaining why the wrapper exists (prevent CI secret exfiltration via build scripts). Update all references in claude.yml, CLAUDE.md, and permission settings. Add `+nightly fmt` to allowedTools so Claude can follow CLAUDE.md formatting instructions in CI. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: PastaClaw <thepastaclaw@users.noreply.github.com> Co-authored-by: lklimek <842586+lklimek@users.noreply.github.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * chore: bump rust-dashcore * chore: imports --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> Co-authored-by: Pasta Lil Claw <pasta+claw@dashboost.org> Co-authored-by: PastaClaw <thepastaclaw@users.noreply.github.com> Co-authored-by: PastaPastaPasta <6443210+PastaPastaPasta@users.noreply.github.com>
PastaPastaPasta
added a commit
that referenced
this pull request
Feb 17, 2026
* fix: skip best chain lock polling in SPV mode (#567) * Initial plan * Skip chain lock refresh in SPV mode Co-authored-by: lklimek <842586+lklimek@users.noreply.github.com> * Document SPV guard intent Co-authored-by: lklimek <842586+lklimek@users.noreply.github.com> --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: lklimek <842586+lklimek@users.noreply.github.com> * fix: compute relative timestamps from actual data (#581) * fix: compute relative timestamps from actual data Replace hardcoded display strings like "Received: 1 day ago" with real relative timestamps computed from document createdAt/updatedAt fields using chrono::Utc::now(). Fixes #579 * style: fix import ordering per cargo fmt * refactor: extract format_relative_time to shared dashpay module Deduplicate the identical format_relative_time function that existed in both contact_requests.rs and send_payment.rs. Move it to the dashpay mod.rs as a pub(crate) function and import from both files. --------- Co-authored-by: PastaClaw <thepastaclaw@users.noreply.github.com> * fix: update platform for DIP-18 HRP and improve SPV sync progress (#575) * fix: update dashpay/platform to d6f4eb9 for DIP-18 HRP fix The previous platform revision used incorrect bech32m HRP prefixes (evo/tevo) for Platform addresses. The updated commit uses the correct DIP-0018 prefixes (dash/tdash). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * refactor(spv): use dash-spv SyncProgress API directly Replace the intermediate SyncProgress/DetailedSyncProgress/SyncStage translation layer with direct use of dash_spv::sync::SyncProgress. This eliminates ~120 lines of bridge code in spawn_progress_watcher() and determine_sync_stage(), and lets the UI query per-manager progress (headers, filter_headers, filters, masternodes, blocks) via the upstream API. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): improve sync progress accuracy and robustness - Log headers() error before discarding for easier debugging - Restore download-window optimization for headers progress on checkpoint-resumed syncs (progress starts near 0% not 83%) - Replace catch-all _ => 0.0 with explicit SyncState variant matches so new variants produce compile errors - Fix filters progress to use current_height/target_height instead of downloaded/target_height (session count vs absolute height mismatch) - Restore peer count in sync status text - Add "Querying peer heights" label and diffs_processed to masternode status for more informative sync messages Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): use height-based blocks progress to prevent bar from jumping The blocks progress bar was bouncing backward because it used processed/requested session counters whose denominator grows as filters discover more matching blocks. Switch to last_processed block height relative to headers target_height, which only increases. Display "current / target" heights instead of percentage on the blocks bar. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): improve progress bar resilience across resume and network switch - Add windowed progress tracking for filters (consistent with headers/filter_headers) - Reset blocks_stage_start on Error state so recovery gets a fresh window - Track spv_progress_network to detect network changes and rebuild progress state from the new network's sync_progress instead of resetting to zero - Eliminate redundant clone in progress watcher (move instead of clone twice) - Consolidate all progress state reset logic into rebuild_spv_progress_state() Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * build: add git and cargo permissions to Claude Code workflow (#565) * build: add git and cargo permissions to Claude Code workflow Allow Claude to run git fetch/merge/checkout/rebase/push and cargo build/test/clippy/fmt commands. Switch model to opus. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: use claude_args for model and allowed tools The `model` and `allowed_tools` inputs are not declared in claude-code-action@v1 and are silently ignored. Move them to `claude_args` with --model and --allowedTools flags. Also fix deprecated colon syntax (`:*`) to space syntax (` *`). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: use single quotes to prevent glob expansion in allowed tools Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * build: sandbox cargo commands and add git permissions for Claude - Add safe-cargo.sh wrapper that strips CI secrets before running cargo - Use --allowedTools for git and safe-cargo, --disallowedTools for raw cargo - Document safe-cargo usage in CLAUDE.md Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: switch safe-cargo.sh from denylist to allowlist approach Use `env -i` (start with empty environment, explicitly pass only what cargo needs) instead of `env -u` (strip known secrets). This is more robust against future secrets being added to the workflow. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * build: deny Claude from editing CI scripts and workflows Prevent Claude from modifying .github/scripts/ and .github/workflows/ to ensure the safe-cargo wrapper cannot be tampered with. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: conditionally pass optional env vars in safe-cargo.sh Empty PROTOC="" caused prost build scripts to fail with "protoc not found". Now optional vars (PROTOC, CC, CXX, etc.) are only passed when set and non-empty. Tested: build, test, fmt all pass through the wrapper. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * rabbit feedback * build: move safe-cargo.sh to scripts/ and allow +nightly fmt Move safe-cargo.sh from .github/scripts/ to top-level scripts/ for better discoverability. Add detailed comment explaining why the wrapper exists (prevent CI secret exfiltration via build scripts). Update all references in claude.yml, CLAUDE.md, and permission settings. Add `+nightly fmt` to allowedTools so Claude can follow CLAUDE.md formatting instructions in CI. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * fix: handle errors instead of panicking on corrupted database blobs (#561) * fix: return error instead of panicking on corrupted database blobs (#560) - Replace `unreachable!()` in `get_scheduled_votes` with warning log and default to false for unexpected `executed` column values - Change `QualifiedIdentity::from_bytes()` to return `Result` instead of panicking via `.expect()` - Propagate deserialization errors as `rusqlite::Error` in all 6 callers so corrupted database is surfaced to the user rather than silently ignored or crashing the app - Add `CorruptedBlobError` newtype in database module to eliminate repeated `FromSqlConversionFailure` boilerplate Closes #560 Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * chore: impl thiserror * chore: fmt * build: add Claude Code GitHub workflow and settings (#552) Cherry-pick from v1.0-dev to enable @claude mentions in PRs. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * chore: fix after merge * fix: skip corrupted identity blobs in get_wallets instead of aborting Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * chore: fail on corrupted identity * doc: document error handling in the db --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * feat: track DAPI endpoint availability in connection status (#533) * fix: connection status not updated * chore: rabbit feedback * Initial plan * Track DAPI connection status and display in tooltips and connection info - Add dapi_total_endpoints and dapi_available fields to ConnectionStatus - Factor DAPI availability into overall_connected status (RED when no endpoints available) - Query SDK AddressList during periodic refresh for endpoint counts and availability - Display DAPI status in connection indicator tooltip - Display DAPI status in network chooser Connection Status card (all modes) - Add dapi_status_label() helper for consistent status text formatting Co-authored-by: lklimek <842586+lklimek@users.noreply.github.com> * Address review feedback: store available endpoint count and extract DRY helper - Changed dapi_available from AtomicBool to AtomicU16 (dapi_available_endpoints) to store the count of available endpoints instead of just a boolean - Display format now shows "Available ({available}/{total} endpoints)" - Extracted repeated DAPI status rendering into add_dapi_status_label() helper function in network_chooser_screen.rs to eliminate code duplication Co-authored-by: lklimek <842586+lklimek@users.noreply.github.com> * Fix rustfmt formatting issues in connection_status.rs - Reorder atomic imports (AtomicU8 before AtomicU16) per rustfmt - Wrap long .store() call to respect line length - Wrap long format!() strings to respect line length Co-authored-by: lklimek <842586+lklimek@users.noreply.github.com> * Fix borrow checker error: extract DAPI status values before mutable self borrow The add_dapi_status_label helper was capturing a &ConnectionStatus reference (derived from self) in closures, which extended the immutable borrow past the self.render_spv_sync_progress() mutable borrow on line 594. Fix: change add_dapi_status_label to accept pre-computed owned values (dapi_total, dapi_available, dapi_label) instead of &ConnectionStatus, and extract those values early alongside other status fields. Co-authored-by: lklimek <842586+lklimek@users.noreply.github.com> * Switch Platform to c2c88e4 and use get_live_addresses() for accurate available count Updated dash-sdk dependency to Platform commit c2c88e4a988ce930 which adds AddressList::get_live_addresses() method. Replaced the workaround that used get_live_address().is_some() (which could only tell if at least one endpoint was live) with get_live_addresses().len() to get the exact count of available non-banned DAPI endpoints. Co-authored-by: lklimek <842586+lklimek@users.noreply.github.com> * chore: typo + network changes * chore: apply feedback * chore: rabbit review * chore: rabbit feedback * chore: rabbitting * Remove overall_connected_with method (deleted upstream in base branch) Co-authored-by: lklimek <842586+lklimek@users.noreply.github.com> * Merge v1.0-dev into copilot/update-dapi-connection-status Resolved modify/delete conflict on src/context.rs: removed the file since v1.0-dev refactored it into src/context/mod.rs and submodules, which already include our Arc<ConnectionStatus> changes. Co-authored-by: lklimek <842586+lklimek@users.noreply.github.com> * chore: fix platfom versioning issues * chore: add todo * chore: fmt --------- Co-authored-by: Lukasz Klimek <842586+lklimek@users.noreply.github.com> Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> * refactor(spv): eliminate separate thread/runtime, improve shutdown and UX (#577) * fix: update dashpay/platform to d6f4eb9 for DIP-18 HRP fix The previous platform revision used incorrect bech32m HRP prefixes (evo/tevo) for Platform addresses. The updated commit uses the correct DIP-0018 prefixes (dash/tdash). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * refactor(spv): use dash-spv SyncProgress API directly Replace the intermediate SyncProgress/DetailedSyncProgress/SyncStage translation layer with direct use of dash_spv::sync::SyncProgress. This eliminates ~120 lines of bridge code in spawn_progress_watcher() and determine_sync_stage(), and lets the UI query per-manager progress (headers, filter_headers, filters, masternodes, blocks) via the upstream API. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): improve sync progress accuracy and robustness - Log headers() error before discarding for easier debugging - Restore download-window optimization for headers progress on checkpoint-resumed syncs (progress starts near 0% not 83%) - Replace catch-all _ => 0.0 with explicit SyncState variant matches so new variants produce compile errors - Fix filters progress to use current_height/target_height instead of downloaded/target_height (session count vs absolute height mismatch) - Restore peer count in sync status text - Add "Querying peer heights" label and diffs_processed to masternode status for more informative sync messages Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): use height-based blocks progress to prevent bar from jumping The blocks progress bar was bouncing backward because it used processed/requested session counters whose denominator grows as filters discover more matching blocks. Switch to last_processed block height relative to headers target_height, which only increases. Display "current / target" heights instead of percentage on the blocks bar. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): improve progress bar resilience across resume and network switch - Add windowed progress tracking for filters (consistent with headers/filter_headers) - Reset blocks_stage_start on Error state so recovery gets a fresh window - Track spv_progress_network to detect network changes and rebuild progress state from the new network's sync_progress instead of resetting to zero - Eliminate redundant clone in progress watcher (move instead of clone twice) - Consolidate all progress state reset logic into rebuild_spv_progress_state() Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): update ConnectionStatus immediately on connect/disconnect Propagate SPV status into ConnectionStatus right after start_spv() and stop_spv() so the UI reflects the change on the next frame instead of waiting for the next throttled trigger_refresh() cycle (2-10 seconds). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): check cancellation token in listener tasks to prevent shutdown hang SPV finality and reconcile listeners blocked app shutdown for up to 10s because they never checked the global cancellation token. Add cancel branches to their tokio::select! loops so they exit immediately on shutdown. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * chore: remove log spamming line * fix(spv): make listener handler calls cancellation-aware and add shutdown trace logging Wrap reconcile_spv_wallets(), handle_spv_finality_event(), and the debounce sleep inside tokio::select! with the cancellation token so shutdown can interrupt them even when blocked on locks held by the SPV sync thread. Add trace-level logging to TaskManager::shutdown() for per-task join timing to aid future shutdown diagnostics. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): disable Disconnect button during stopping and poll status faster - Disable the Disconnect button while SPV status is Stopping to prevent double-clicks and provide visual feedback - Poll SPV status every 200ms during Stopping instead of the 10s connected interval so the Stopped transition is reflected within 1s - Reset the throttle timer in stop_spv() so fast polling starts immediately Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * refactor(tasks): add mandatory task names to spawn_sync for shutdown diagnostics Change spawn_sync() signature to require a &'static str name. The JoinSet now yields the task name on completion, letting shutdown() log which tasks finished and which ones timed out. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * chore: log task names * refactor(spv): eliminate separate OS thread and tokio runtime Replace the dedicated SPV thread + 4-worker tokio runtime with a spawned task on the main 12-worker runtime via TaskManager::spawn_sync. This simplifies shutdown (SPV loop now tracked in unified JoinSet), removes cross-runtime complexity, and improves debuggability. Additional changes: - Fix: zeroize xprv_str after wallet import (security H-1) - Fix: sanitize devnet_name in build_spv_data_dir to prevent path traversal - Add 21 integration tests covering lifecycle, concurrency, deadlock detection, and live testnet sync Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): derive stop_token as child of global cancel for clean shutdown The spv_request_handler selected on stop_token, which was independent of the global cancellation token. During window-close shutdown only the global token was cancelled, leaving the request handler running and causing a ~5s hang until the TaskManager timeout aborted it. Fix by creating stop_token as a child_token() of the global cancel. This also simplifies run_spv_loop and run_sync_and_monitor by removing the redundant global_cancel parameter — a single stop_token now covers both explicit SpvManager::stop() and application-wide shutdown. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * CLAUDE: minor changes about docs * fix(spv): bump rust-dashcore to d8bc066 for faster disconnect and add shutdown tracing Pin rust-dashcore patches to commit d8bc066 which includes: - sync manager task loop exits on network errors instead of logging indefinitely - sync coordinator signal_shutdown() cancels tasks before network disconnect - connection tasks race Peer::connect() against shutdown token Add debug tracing to SpvManager::stop() and run_sync_and_monitor() to measure client.stop() duration, and trace-level polling of SPV status in ConnectionStatus. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * chore: update rust-dashcore rev * merge: resolve v1.0-dev conflicts for #577 (#585) * fix: compute relative timestamps from actual data (#581) * fix: compute relative timestamps from actual data Replace hardcoded display strings like "Received: 1 day ago" with real relative timestamps computed from document createdAt/updatedAt fields using chrono::Utc::now(). Fixes #579 * style: fix import ordering per cargo fmt * refactor: extract format_relative_time to shared dashpay module Deduplicate the identical format_relative_time function that existed in both contact_requests.rs and send_payment.rs. Move it to the dashpay mod.rs as a pub(crate) function and import from both files. --------- Co-authored-by: PastaClaw <thepastaclaw@users.noreply.github.com> * fix: update platform for DIP-18 HRP and improve SPV sync progress (#575) * fix: update dashpay/platform to d6f4eb9 for DIP-18 HRP fix The previous platform revision used incorrect bech32m HRP prefixes (evo/tevo) for Platform addresses. The updated commit uses the correct DIP-0018 prefixes (dash/tdash). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * refactor(spv): use dash-spv SyncProgress API directly Replace the intermediate SyncProgress/DetailedSyncProgress/SyncStage translation layer with direct use of dash_spv::sync::SyncProgress. This eliminates ~120 lines of bridge code in spawn_progress_watcher() and determine_sync_stage(), and lets the UI query per-manager progress (headers, filter_headers, filters, masternodes, blocks) via the upstream API. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): improve sync progress accuracy and robustness - Log headers() error before discarding for easier debugging - Restore download-window optimization for headers progress on checkpoint-resumed syncs (progress starts near 0% not 83%) - Replace catch-all _ => 0.0 with explicit SyncState variant matches so new variants produce compile errors - Fix filters progress to use current_height/target_height instead of downloaded/target_height (session count vs absolute height mismatch) - Restore peer count in sync status text - Add "Querying peer heights" label and diffs_processed to masternode status for more informative sync messages Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): use height-based blocks progress to prevent bar from jumping The blocks progress bar was bouncing backward because it used processed/requested session counters whose denominator grows as filters discover more matching blocks. Switch to last_processed block height relative to headers target_height, which only increases. Display "current / target" heights instead of percentage on the blocks bar. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix(spv): improve progress bar resilience across resume and network switch - Add windowed progress tracking for filters (consistent with headers/filter_headers) - Reset blocks_stage_start on Error state so recovery gets a fresh window - Track spv_progress_network to detect network changes and rebuild progress state from the new network's sync_progress instead of resetting to zero - Eliminate redundant clone in progress watcher (move instead of clone twice) - Consolidate all progress state reset logic into rebuild_spv_progress_state() Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * build: add git and cargo permissions to Claude Code workflow (#565) * build: add git and cargo permissions to Claude Code workflow Allow Claude to run git fetch/merge/checkout/rebase/push and cargo build/test/clippy/fmt commands. Switch model to opus. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: use claude_args for model and allowed tools The `model` and `allowed_tools` inputs are not declared in claude-code-action@v1 and are silently ignored. Move them to `claude_args` with --model and --allowedTools flags. Also fix deprecated colon syntax (`:*`) to space syntax (` *`). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: use single quotes to prevent glob expansion in allowed tools Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * build: sandbox cargo commands and add git permissions for Claude - Add safe-cargo.sh wrapper that strips CI secrets before running cargo - Use --allowedTools for git and safe-cargo, --disallowedTools for raw cargo - Document safe-cargo usage in CLAUDE.md Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: switch safe-cargo.sh from denylist to allowlist approach Use `env -i` (start with empty environment, explicitly pass only what cargo needs) instead of `env -u` (strip known secrets). This is more robust against future secrets being added to the workflow. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * build: deny Claude from editing CI scripts and workflows Prevent Claude from modifying .github/scripts/ and .github/workflows/ to ensure the safe-cargo wrapper cannot be tampered with. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: conditionally pass optional env vars in safe-cargo.sh Empty PROTOC="" caused prost build scripts to fail with "protoc not found". Now optional vars (PROTOC, CC, CXX, etc.) are only passed when set and non-empty. Tested: build, test, fmt all pass through the wrapper. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * rabbit feedback * build: move safe-cargo.sh to scripts/ and allow +nightly fmt Move safe-cargo.sh from .github/scripts/ to top-level scripts/ for better discoverability. Add detailed comment explaining why the wrapper exists (prevent CI secret exfiltration via build scripts). Update all references in claude.yml, CLAUDE.md, and permission settings. Add `+nightly fmt` to allowedTools so Claude can follow CLAUDE.md formatting instructions in CI. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: PastaClaw <thepastaclaw@users.noreply.github.com> Co-authored-by: lklimek <842586+lklimek@users.noreply.github.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * chore: bump rust-dashcore * chore: imports --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> Co-authored-by: Pasta Lil Claw <pasta+claw@dashboost.org> Co-authored-by: PastaClaw <thepastaclaw@users.noreply.github.com> Co-authored-by: PastaPastaPasta <6443210+PastaPastaPasta@users.noreply.github.com> * ci: cancel in-progress workflow runs on new push (#590) * ci: cancel in-progress workflow runs on new push Add concurrency groups to Tests and Clippy workflows so that previous runs are automatically cancelled when a new commit is pushed to the same branch or PR. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * trigger ci --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * fix: use correct DashPay profile field "publicMessage" instead of "bio" (#582) The DashPay contract schema defines the field as "publicMessage", not "bio". This caused profile bios to never load in contact info. Co-authored-by: PastaClaw <thepastaclaw@users.noreply.github.com> --------- Co-authored-by: Copilot <198982749+Copilot@users.noreply.github.com> Co-authored-by: lklimek <842586+lklimek@users.noreply.github.com> Co-authored-by: PastaClaw <thepastaclaw@users.noreply.github.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> Co-authored-by: PastaPastaPasta <6443210+PastaPastaPasta@users.noreply.github.com>
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
2 participants
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.
Replace hardcoded display strings like "Received: 1 day ago" with real relative timestamps computed from document createdAt/updatedAt fields using chrono::Utc::now().
Fixes #579
Split from #580 — timestamps only, no unrelated changes.
Summary by CodeRabbit