Carried over from a deploy.py run -- incrementPackageID bumps VersionBuild but
deploy.py only auto-commits HiAPI-docsite-2025.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The 3.2 line is retracted -- the major version stays fixed for commercial
positioning, so 3.1 remains the development line. Version stamp only; no code
changed on the way out or back.
VersionBuild resumes at 158, where it stood before the detour, instead of
restarting from the 143 the feed serves. Everything between was built locally and
stamped into real assemblies even though deploy.py deliberately withholds the
upload until the API has settled commercially, and re-minting a spent number with
different code is the one failure that yields two Hi.WpfPlus.dll claiming the same
version. Next upload is 3.1.159; the 143 -> 159 gap on the feed is intended.
3.2.1 was packed locally and never uploaded, so nothing outside this machine saw
the 3.2 line at all.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The 3.1 line ends at Hi.WpfPlus 3.1.143 -- tag v3.1.143, part of the aligned
hinc-suite/3.1.175 set. master now produces 3.2 and VersionBuild restarts at 0,
so the first published 3.2 package is 3.2.1: IncrementPackageID.targets bumps
before the build, not after it.
The 3.1.158 stamp this replaces was a local increment from a deploy run whose
upload step is commented out, so it never reached the feed and nothing depends on
it. Fixes for customers pinned to 3.1 come off the support/3.1 branch.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
WpfDispUtil.cs carried a stale `using Google.Protobuf.WellKnownTypes;`
with zero matching symbol usage in the file body — it resolved only
because HiGeom used to re-export Google.Protobuf transitively via its
now-removed Grpc.Tools / Google.Protobuf PackageRefs. CS0246 on
'Google' once that transitive bring-in went away. VersionBuild bump
140→141 rides along for the NuGet republish.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>