summaryrefslogtreecommitdiff
path: root/serial/VirtualSerial/queue.cpp
diff options
context:
space:
mode:
author5an7y <[email protected]>2026-06-22 17:31:47 -0700
committer5an7y <[email protected]>2026-06-22 17:31:47 -0700
commitbe91bbd2dc4478a57a4007fc32fe320da79f85e1 (patch)
treeef9db4a561de4c3af23d70bbfb1570fdf5ef474b /serial/VirtualSerial/queue.cpp
parent6c6c308f1fe6ec1ce2ce7d0d7b3b455305a4a346 (diff)
Add NT-version-scoped exclusions and a parallel multi-version CI matrix
Exclusions: - Add MinNtTargetVersion/MaxNtTargetVersion columns to exclusions.csv so a row can apply only within an _NT_TARGET_VERSION build-number range (parsed from -NtTargetVersion, e.g. 22000 from 10.0.22000). Import-SampleExclusions filters this range at load time alongside the existing build-number range; blank = unbounded. - Exclude the samples that fail only when linking against older library sets, with the reason taken from the build logs: audio.sysvad (<=22000, KSJACK_DESCRIPTION3); network.netadaptercx.netvadapter and network.wlan.wificx (<=22621, NDIS/DDI version); powerlimit.plclient/plpolicy (<=22621, POWER_LIMIT_ATTRIBUTES); storage classpnp/storahci (<=22621, STOR_ADDRESS_TYPE_NVME, Debug-only) and storage.msdsm (Debug|x64). usb.usbview is intentionally NOT excluded: it fails on every version for a known host reason (missing .NET 4.7.2/4.8.1 targeting packs). CI: - ci.yml and ci-pr.yml: add _NT_TARGET_VERSION as a manual matrix axis (4 newest versions, latest-first) so each version x configuration x platform runs on its own parallel runner. - Build-Samples.ps1 and Join-CsvReports.ps1 now write an easy-to-scan Markdown summary to GITHUB_STEP_SUMMARY: each build job shows counts and a failures table with the first error; the report job shows per-version totals and a consolidated failures list, plus the colour-coded sample x version HTML/CSV overview. Co-authored-by: Copilot <[email protected]>
Diffstat (limited to 'serial/VirtualSerial/queue.cpp')
0 files changed, 0 insertions, 0 deletions