Skip to content
Closed
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
68 changes: 53 additions & 15 deletions .github/workflows/ci.yml
Original file line number Diff line number Diff line change
Expand Up @@ -125,7 +125,8 @@ jobs:
with:
bun-version: 1.3.14
- name: Install ZIP tools
run: sudo apt-get update && sudo apt-get install -y zip unzip
# Hosted images already carry both tools; install only if one is missing.
run: command -v zip >/dev/null && command -v unzip >/dev/null || (sudo apt-get update && sudo apt-get install -y zip unzip)
- name: Install project dependencies
run: bun install --frozen-lockfile --ignore-scripts
- name: Run Linux tests, builds and package checks
Expand Down Expand Up @@ -172,7 +173,8 @@ jobs:
with:
bun-version: 1.3.14
- name: Install ZIP tools
run: sudo apt-get update && sudo apt-get install -y zip unzip
# Hosted images already carry both tools; install only if one is missing.
run: command -v zip >/dev/null && command -v unzip >/dev/null || (sudo apt-get update && sudo apt-get install -y zip unzip)
- name: Install project dependencies
run: bun install --frozen-lockfile --ignore-scripts
- name: Run Linux tests, builds and package checks
Expand Down Expand Up @@ -216,9 +218,23 @@ jobs:
# ownership semantics without relaxing the product's SID policy.
$name = 'progress' + [Guid]::NewGuid().ToString('N').Substring(0, 10)
$password = ConvertTo-SecureString ('Aa1!' + [Guid]::NewGuid().ToString('N')) -AsPlainText -Force
$output = Join-Path $env:RUNNER_TEMP ($name + '.stdout')
$errors = Join-Path $env:RUNNER_TEMP ($name + '.stderr')
$trace = Join-Path $env:RUNNER_TEMP 'windows-migration-trace.jsonl'
# The two native files run side by side, each in its own serial test
# runner. Only the service file starts a server and request workers,
# which are what take Setup coordination ports; the platform file
# drives the setup controller directly. Each runs under Start-Process
# -Wait inside a thread job, so it is waited for with everything it
# started, exactly as when they ran one after the other. Starting the
# two at the same instant can be refused once, so the start is retried.
$runs = @('windows-platform', 'windows-service') | ForEach-Object {
[pscustomobject]@{
File = $_
Output = Join-Path $env:RUNNER_TEMP ($name + '.' + $_ + '.stdout')
Errors = Join-Path $env:RUNNER_TEMP ($name + '.' + $_ + '.stderr')
Job = $null
ExitCode = $null
}
}
# Leave a fixed diagnostic even if principal/process creation fails.
'{"request":"runner-start","completed":false}' | Set-Content -LiteralPath $trace -Encoding utf8
$created = $false
Expand All @@ -227,23 +243,45 @@ jobs:
$created = $true
Add-LocalGroupMember -SID 'S-1-5-32-545' -Member $user
$credential = [PSCredential]::new(($env:COMPUTERNAME + '\' + $name), $password)
$process = Start-Process -FilePath (Get-Command node).Source -ArgumentList '--experimental-strip-types --test --test-concurrency=1 tests/backend/windows-platform.test.ts tests/backend/windows-service.test.ts' -WorkingDirectory $env:GITHUB_WORKSPACE -Credential $credential -LoadUserProfile -WindowStyle Hidden -Wait -PassThru -RedirectStandardOutput $output -RedirectStandardError $errors
if (Test-Path $output) {
Get-Content -LiteralPath $output
foreach ($run in $runs) {
$run.Job = Start-ThreadJob -ArgumentList $run, $credential, $env:GITHUB_WORKSPACE, (Get-Command node).Source -ScriptBlock {
param($run, $credential, $workspace, $node)
# The secondary logon service creates one process at a time and
# refuses a second request while it is busy; retry only that refusal.
$process = $null
for ($attempt = 1; -not $process; $attempt++) {
try {
$process = Start-Process -FilePath $node -ArgumentList "--experimental-strip-types --test --test-concurrency=1 tests/backend/$($run.File).test.ts" -WorkingDirectory $workspace -Credential $credential -LoadUserProfile -WindowStyle Hidden -Wait -PassThru -RedirectStandardOutput $run.Output -RedirectStandardError $run.Errors -ErrorAction Stop
} catch {
if ($attempt -ge 30 -or $_.Exception.Message -notmatch 'already running') { throw }
Start-Sleep -Seconds 1
}
}
$process.ExitCode
}
}
$null = $runs.Job | Wait-Job
foreach ($run in $runs) { $run.ExitCode = @(Receive-Job -Job $run.Job -ErrorAction Stop)[-1] }
foreach ($run in $runs) {
if (Test-Path $run.Output) {
Get-Content -LiteralPath $run.Output
}
if (Test-Path $run.Errors) { Get-Content -LiteralPath $run.Errors }
}
if (Test-Path $errors) { Get-Content -LiteralPath $errors }
if ($process.ExitCode -ne 0) { throw 'Native Windows tests failed.' }
if (@($runs | Where-Object { $_.ExitCode -ne 0 }).Count -gt 0) { throw 'Native Windows tests failed.' }
} finally {
# Preserve per-request evidence even on an assertion/process failure.
# Never upload raw stdout/stderr, local principal names or paths.
if (Test-Path $output) {
$traces = @(Get-Content -LiteralPath $output | ForEach-Object {
if ($_ -match '^\s*(?:#\s*)?WINDOWS_MIGRATION_TRACE (.+)$') { $Matches[1] }
})
if ($traces.Count -gt 0) { $traces | Add-Content -LiteralPath $trace -Encoding utf8 }
foreach ($run in $runs) {
if (Test-Path $run.Output) {
$traces = @(Get-Content -LiteralPath $run.Output | ForEach-Object {
if ($_ -match '^\s*(?:#\s*)?WINDOWS_MIGRATION_TRACE (.+)$') { $Matches[1] }
})
if ($traces.Count -gt 0) { $traces | Add-Content -LiteralPath $trace -Encoding utf8 }
}
}
if ($created) { Remove-LocalUser -Name $name }
Remove-Item -LiteralPath $output, $errors -Force -ErrorAction SilentlyContinue
Remove-Item -LiteralPath @($runs | ForEach-Object { $_.Output; $_.Errors }) -Force -ErrorAction SilentlyContinue
}
$requests = @(Get-Content -LiteralPath $trace | ForEach-Object { $_ | ConvertFrom-Json } | Where-Object { $_.request -and $_.request -notin @('runner-start','summary') })
if ($requests.Count -eq 0) { throw 'Native Windows request trace is missing.' }
Expand Down
24 changes: 23 additions & 1 deletion CHANGELOG.md
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,27 @@ All notable changes to this project are documented in this file. The format is b

## [Unreleased]

## [1.5.2] - 2026-11-19

Cleanup follow-ups from the 1.5.1 review and a companion system prompt section that stays identical across requests, so it no longer defeats a provider's prompt cache. The minimum host remains OpenChamber 2.7.1. Protocol and schema stay at 2.

### Changed

- The companion's system prompt section holds only the fixed activation guidance and is byte-identical on every request where it is present. The one-request hint after a dispatch or plan write, the sub-agent resume capsule and the delegation refusal note keep their text and eligibility but are delivered after the cacheable prefix: as a text part after a user message, or at the end of the last tool result of a tool message, never on the Code Mode `execute` wrapper's result or a non-text result, and otherwise on the next request with a valid carrier. A delivered note stays in the same place on later requests built from the same history, keyed per session and bounded to 128 sessions, 256 carriers and 128 KiB of note text per session; notes due together are joined in a fixed order, a capsule or refusal is delivered again when its text changed or its last carrier is gone, and one fixed line clears a refusal that no longer applies. Every path that removes the section removes the added notes. Providers that drop plugin system sections now also receive these notes. Cache hit rates were not measured; tests assert the byte-for-byte prefix (#165, #167; reported in #161).
- Companion inline and packaged guidance tells an agent to clear a task's dependency block against the plan's current prerequisites, read before a handoff and when a prerequisite finishes: a prerequisite the plan no longer lists no longer holds the task, and a skipped prerequisite that is still listed counts as unfinished. Decision, review and external blockers are left alone, and the guidance stays advisory (#162).
- Coordinator refusal guidance validates a pending or reserved delegation receipt against its dispatch and declared tasks before inferring that cleanup retired it, and reports incomplete guidance when the check fails. Cancellation receipts, bound receipts and the explicit-link recovery are read as before (#162).
- `docs/ui.md` and `docs/contracts.md` state when chosen plans are checked by `/v1/search-state` point lookup, when loaded rows or a list scan are used instead, and which revision a retry compares against (#162).
- The large-library gate bounds each timed window by request and reply offsets read in the browser evaluation that starts and stops it, counts cadence, list starts and both inbox forms by the requests sent inside it, and shows run reads per window as an informational row. It records how an artifact attributed its traffic, compares the list-carried inbox form on its own, and reports a budget it cannot compare as unverified with exit status 2 instead of passing it. QA tooling only; no shipped behaviour changes (#166).
- The `release-1-5-2` OpenSpec change records the proposal, design and tasks (#160). The 1.5.0, 1.5.1 and 1.5.2 changes are archived in that order with their task lists reconciled, the 1.5.1 deltas corrected to keep the scenario names and requirement lengths the specs hold, and a Purpose written for each of the ten capabilities the first two archives created; the third creates none.
- Portable tests run faster, locally and in CI, with no test removed or relaxed. The serial backend group that takes Setup coordination ports runs beside a second lane (both typechecks, the parallel backend group, and the browser-model and automation tests) instead of before them, in one `--test-concurrency=1` invocation that nothing else joins; the second lane has its worker count capped at cores minus two and runs under `nice`. With `CP_BUILT_SERVICE=1`, the one file that takes ports moves into the serial group. New `bun run test:fast` (everything except the serial port-bound group) and `bun run test:all` (everything portable CI tests except packaging and the rebuild comparison) share one plan with CI through a POSIX-first runner whose stop path on Windows is best effort and was not run there. Node's compile cache, emptied at the start of each run, serves every step. The Linux jobs install the ZIP tools only when the image lacks them, and the Windows job starts its two native test files as two processes at once. The Linux test step fell from about 305 s to about 150 to 230 s in CI (#170).

### Fixed

- Tidy and Select all result lines separate plans that were skipped (refused as changed, closed, reused or not found, store busy, left out of the view or never sent after Stop) from plans whose request got no valid answer and may have been applied, in English, Russian and Simplified Chinese with plural rules. Undo still restores only validated receipts; Try again recounts skipped and unconfirmed plans together, or sends the restore again for the remainder after a partial Undo (#163).
- A bulk run that ends after an account, store or binding change the frame observed, including a switch away and back, no longer resets caches, changes the selection, closes the opened plan or shows a result on the board then shown; a validated Delete receipt still removes the review marker of the account and store the run started in. The busy state belongs to the request that set it, so an observed switch releases it and a late answer cannot complete, release or reset a newer run. Cleanup controls follow the latest health reply the controller accepted, whichever remembered list is shown, and a failed list read applies its health facts in both directions (#164).
- A count that finishes in the background, a run's progress and the end of a run with its result line repaint as background data and wait while a control is held, so a press on Stop or a click on another control is not lost; a refresh no longer expires a result that has not been painted yet. A deferred cleanup focus move is decided when it is applied and dropped if focus has moved to another control (#168).
- Tidy and Select all counts taken from a plan list read in more than one page say that plans may have been missed, on the count line and on the deletion step; a single-page list and rows chosen one by one stay exact, and Try again stays exact only when it found every plan it names. A search freshness reply certifies only the result page it was asked about by number, and only when it covered every plan that page shows; a page opened by Load more starts unchecked; a cleanup in the frame retires freshness replies in flight; and a plan a reply shows gone leaves the selection whichever page is shown. Under Select all, a plan taken out by a receipt or a reply carries the cleanup generation that took it out, so a plan restored elsewhere is counted again while a deselected plan stays out until its row is clicked or Select all is pressed again (#169).

## [1.5.1] - 2026-11-18

Full-page responsiveness with large libraries, visible Tidy and multi-select cleanup, and map and count fixes. The minimum host remains OpenChamber 2.7.1. Protocol and schema stay at 2.
Expand Down Expand Up @@ -346,7 +367,8 @@ The first major release: a calmer, at-a-glance progress experience with an inter

Baseline before automated releases; see the repository history.

[Unreleased]: https://github.com/Phototonic/openchamber-progress/compare/v1.5.1...HEAD
[Unreleased]: https://github.com/Phototonic/openchamber-progress/compare/v1.5.2...HEAD
[1.5.2]: https://github.com/Phototonic/openchamber-progress/compare/v1.5.1...v1.5.2
[1.5.1]: https://github.com/Phototonic/openchamber-progress/compare/v1.5.0...v1.5.1
[1.5.0]: https://github.com/Phototonic/openchamber-progress/compare/v1.4.1...v1.5.0
[1.4.1]: https://github.com/Phototonic/openchamber-progress/compare/v1.4.0...v1.4.1
Expand Down
Loading