fix: use canonical structure() attribute names

[krlmlr]

LLM disclosure: this PR was prepared with Claude Code (Claude Opus 4.8).

• By submitting this pull request, I assign the copyright of my contribution to The igraph development team.

What

R-devel deprecates the special structure() arguments .Names, .Dim and .Dimnames; they emit warnings that the test suite escalates (the ~27 warnings noted on #2749).

Replace them with the canonical names, dim and dimnames. structure() maps both spellings to the same attributes, so the constructed objects are identical — no behavior change.

• 42 occurrences: 9 tests/testthat/*.R fixtures (mostly dput() output) and one R/plot.common.R fixture.
• Each replacement verified to be a structure() attribute argument — the list(...) data and its inner element names are untouched.
• All edited files parse cleanly.

Scope

Code changes only, branched off main. The CI/workflow side of the R-devel work lives separately:

• ci: let pak ignore Bioconductor graph on R < 4.4 #2753 — disable Bioconductor on old R (via pak's PKG_USE_BIOCONDUCTOR) so pak resolution stops failing on R 4.2/4.3 where the archived graph releases now 404.
• ci: run R-CMD-check on claude/** branches #2754 — run R-CMD-check on claude/** branches.
• fix: declare R_getRegisteredNamespace() so the C23 sanitizer build compiles #2749 — the R_getRegisteredNamespace C23 sanitizer error, which is a stale CSAN Docker image, not a package bug.

🤖 Generated with Claude Code

---

Generated by Claude Code

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if 8ced78f is merged into main:

• ✔️as_adjacency_matrix: 788ms -> 785ms [-2.13%, +1.19%]
• ✔️as_biadjacency_matrix: 755ms -> 752ms [-2.22%, +1.63%]
• ✔️as_data_frame_both: 1.43ms -> 1.47ms [-2.23%, +7.94%]
• ✔️as_long_data_frame: 3.52ms -> 3.52ms [-4%, +4%]
• ✔️es_attr_filter: 2.68ms -> 2.71ms [-3.36%, +6.12%]
• ✔️graph_from_adjacency_matrix: 148ms -> 147ms [-2.42%, +0.87%]
• ❗🐌graph_from_data_frame: 3.06ms -> 3.13ms [+0.03%, +4.78%]
• ✔️vs_attr_filter: 1.49ms -> 1.5ms [-3.81%, +4.24%]
• ✔️vs_by_name: 997µs -> 1ms [-4.01%, +4.72%]
  Further explanation regarding interpretation and methodology can be found in the documentation.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if ac4a9e5 is merged into main:

• ✔️as_adjacency_matrix: 743ms -> 743ms [-0.33%, +0.24%]
• ✔️as_biadjacency_matrix: 724ms -> 723ms [-0.56%, +0.31%]
• ✔️as_data_frame_both: 1.43ms -> 1.43ms [-1.1%, +0.83%]
• ✔️as_long_data_frame: 3.82ms -> 3.82ms [-0.65%, +0.63%]
• ✔️es_attr_filter: 2.61ms -> 2.61ms [-0.36%, +0.65%]
• ✔️graph_from_adjacency_matrix: 116ms -> 116ms [-0.84%, +1.93%]
• ✔️graph_from_data_frame: 3.26ms -> 3.25ms [-0.89%, +0.51%]
• ✔️vs_attr_filter: 1.46ms -> 1.47ms [-0.14%, +1.86%]
• ✔️vs_by_name: 936µs -> 941µs [-0.37%, +1.38%]
  Further explanation regarding interpretation and methodology can be found in the documentation.

[krlmlr]

Reran checks to measure performance. Even if it was a problem here, we can't do very much about it.