Skip to content

Bump xmscore/xmsgrid and regenerate CI with xmsconan 2.18.0 - #95

Merged
gagelarsen merged 1 commit into
masterfrom
bump/xmsconan-2.18.0
Aug 14, 2026
Merged

Bump xmscore/xmsgrid and regenerate CI with xmsconan 2.18.0#95
gagelarsen merged 1 commit into
masterfrom
bump/xmsconan-2.18.0

Conversation

@gagelarsen

Copy link
Copy Markdown
Member

Third library onto the VS2019 bridge, after xmscore 7.0.12 and xmsgrid 9.0.11. Six lines, two coupled changes.

1. Dependency bumps — required, not routine

dependency from to
xmscore 7.0.11 7.0.12
xmsgrid 9.0.10 9.0.11

xms_dependencies pins exact versions, and neither 7.0.11 nor 9.0.10 was ever built for msvc 192. Only 7.0.12 and 9.0.11 exist on aquaveo-vs2019, so left at the old pins the VS2019 build of xmsinterp fails at graph resolution before compiling a single file.

Both bumps are inert for the VS2022 side. Each of those releases contained only its own dependency-and-CI bump — no C++ source change — and each is published on both remotes at a single matching recipe revision (xmscore c58afa49…, xmsgrid 66ac6991…).

2. xmsconan ==2.16.0>=2.18.0

CI must generate the recipe with the same xmsconan the manual VS2019 track uses. XmsConan2File.export() copies xms_conan2_file.py into the export folder, so its contents feed the recipe revision hash, and 2.18.0 changes that file by 192 lines (the msvc 192 fork). Generating with 2.16.0 in CI and 2.18.0 locally would publish different rrevs for the same version.

Blast radius on the VS2022 side: none

  • CI diff is pins only — verified by filtering the diff for non-xmsconan changes.
  • No [vs2019_dependency_overrides] table, so the generated conanfile.py gains no new attribute.
  • Regenerated recipe verified: xms_dependencies = ["xmscore/7.0.12", "xmsgrid/9.0.11"], testing_framework = "cxxtest" (relevant because gtest/1.17.0 has no msvc 192 build).

Last regeneration commit

2.17.0 floated generated CI onto >= with --upgrade, so future xmsconan releases arrive on the next CI run. The flake job picks up --upgrade, which it lacked under the == pin.

🤖 Generated with Claude Code

Third library onto the VS2019 bridge, after xmscore 7.0.12 and xmsgrid 9.0.11.

**xmscore 7.0.11 -> 7.0.12, xmsgrid 9.0.10 -> 9.0.11.** xms_dependencies pins
exact versions, and neither 7.0.11 nor 9.0.10 was ever built for msvc 192 --
only 7.0.12 and 9.0.11 exist on aquaveo-vs2019. Left at the old pins the
VS2019 build fails at graph resolution before compiling anything.

Both bumps are inert for the msvc 194 / gcc / clang builds. Each of those
releases contained only its own dependency-and-CI bump, no C++ source change,
and each is published on both remotes at a single matching recipe revision
(xmscore c58afa49..., xmsgrid 66ac6991...).

**xmsconan 2.16.0 -> 2.18.0.** CI has to generate the recipe with the same
xmsconan the manual VS2019 track uses. XmsConan2File.export() copies
xms_conan2_file.py into the export folder, so its contents feed the recipe
revision hash, and 2.18.0 changes that file by 192 lines (the msvc 192 fork).
Generating with 2.16.0 in CI and 2.18.0 locally would publish different rrevs
for the same version, leaving aquaveo-stable and aquaveo-vs2019 out of step on
what reads as one release.

Nothing else in the generated output moves -- the CI diff is pins only.
build.toml declares no [vs2019_dependency_overrides] table, so the generated
conanfile.py gains no new attribute, and on any toolchain that is not msvc 192
the dependency graph is unchanged.

2.17.0 floated generated CI onto >= with --upgrade, so this should be the last
regeneration commit this repo needs.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
@gagelarsen
gagelarsen merged commit a620f76 into master Aug 14, 2026
18 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant