[Abhishek-Mohan]

Which issue does this PR close?

Closes #7101.

Rationale for this change

Four Scala comments still name the user-guide iceberg-writes.md page removed by #7090. Two give a nonexistent path; two give a bare filename that can be confused with the contributor-guide page.

What changes are included in this PR?

Point all four comments to docs/source/user-guide/latest/iceberg.md and name "Accepted divergences" under Writes. The references cover compression, fanout ordering, and file rolling in the two Iceberg benchmarks.

The patch changes comments only. #5724 and its follow-up #7226 also update the write-benchmark reference on their open branches; all four references remain stale on main at submission.

How are these changes tested?

• Verified that the target page, Writes and Accepted divergences headings, and the compression, fanout-ordering, and file-rolling entries exist.
• Confirmed no stale iceberg-writes.md references remain in Scala sources.
• Compared non-comment source lines against the base: unchanged in all four files.
• git diff --check passed.
• Offline Spotless check passed for all four changed files:

./mvnw -o spotless:check "-DspotlessFiles=.*(IcebergWriteProtoTranslation|CometIcebergWriteActionSuite|CometIcebergEndToEndBenchmark|CometIcebergWriteBenchmark)\.scala"

Local validation at 2d59d4c67

No new tests were added because only comments changed. Additional local validation was run on macOS arm64 at 2d59d4c67. The head has since only merged main; none of the four files changed, and the diff against main is the same:

• Native debug build passed.
• CI lint/compile command (package -DskipTests scalafix:scalafix -Dscalafix.mode=CHECK -Psemanticdb) passed for Spark 3.5.9 / Scala 2.12 / JDK 17; Spark 4.0.4 / JDK 17 and JDK 21; Spark 4.1.3 / JDK 17; and Spark 4.2.0 / JDK 17.
• Strict-warning main/test compilation passed with Spark 3.5 + Delta.
• IcebergWriteProtoTranslationSuite and CometIcebergWriteActionSuite completed with the results below. The fanout-ordering test passed on Spark 3.5, both 4.0 runs, and 4.1.

Spark	Scala	JDK	Passed	Failed	Canceled
3.5.9	2.13	17	135	0	5
4.0.4	2.13	17	137	0	3
4.0.4	2.13	21	137	0	3
4.1.3	2.13	17	140	0	0
4.2.0	2.13	17	26	0	114

Spark 4.2 has no Iceberg runtime in this profile: its Iceberg-dependent cases were canceled, so that row does not establish native Iceberg write coverage. Other cancellations are existing version guards. Spark 3.4 was excluded from this requested local matrix because it is deprecated.

Local Iceberg Spark and Spark SQL suite runs were stopped as unnecessary for a comment-only change. Local results are macOS coverage, not a claim that Linux CI has passed.

CI

The pull request tier runs on the fork (CI (fork)). It passed on earlier heads 9443c765d and d562fe16a. One later run failed only because GitHub lost the runner in the unrelated native-executors job. The merge queue runs Spark SQL on Spark 4.1 and Iceberg 1.11.

Review skills

Reviewed against main with the repository skills review-comet-pr, review-comet-docs, and review-comet-iceberg-write-pr. review-comet-expression-pr is routed by the serde/ path but has nothing to check, because no expression or executable code changes. No findings. The docs-review mechanical checks found no "native" or "vectorized" wording, config keys, operator names, version claims, or markers, and no file under docs/source changes. Compared with branch-1.2, the PR changes no behavior.

Council review

Reviewed using the t3dotgg/tstack council workflow. In both rounds every delegate was satisfied with no required code changes. Round 2 asked only for this description update.

Rounds: 2

Delegate	Round 1 (2d59d4c67)	Round 2 (2dfdc72a2)
Claude Opus 5.5	9.5/10	9.5/10
Gemini 4 Argon	10/10	10/10
GPT-6 Astra	10/10	10/10

Created with GPT-6 Astra and Claude Opus 5.5 in OpenCode through T3 Code.