fix(semconv): avoid duplicate metric recording without attributes

[jstar0]

Fixes #8780

The retained v1.32.0-v1.34.0 generated metric helpers record through the empty-attributes fast path and then fall through to the general attribute path. This causes histogram helpers to export duplicate count/sum observations and invokes gauge helpers twice.

This updates all 60 affected helpers across the three retained semantic-convention versions and adds a semconvkit AST regression check so generated metric helpers must return after their no-attribute fast path.

Validation:

• go test ./semconv/v1.32.0/...
• go test ./semconv/v1.33.0/...
• go test ./semconv/v1.34.0/...
• go test ./semconvkit from internal/tools
• make precommit completed generation, lint, README, and module checks, then stopped at the existing bridge/opentracing gRPC test with a server preface EOF; the changed semconv scopes pass.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.4%. Comparing base (1ad6638) to head (2a1366e).

Additional details and impacted files

@@          Coverage Diff          @@
##            main   #8849   +/-   ##
=====================================
  Coverage   88.4%   88.4%           
=====================================
  Files        331     331           
  Lines      21001   21001           
=====================================
  Hits       18572   18572           
  Misses      2429    2429           

see 2 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[jstar0]

The remaining check-links failure appears to be baseline-only, not caused by this PR.

• Base: 736a14fcdca28b8cf5237e6b9b166ec6ed832bf7
• Head: 37b816049b598997524f21847cf7e0322772a5c1
• The six failing 403 URLs are unchanged in semconv/v1.32.0, v1.33.0, and v1.34.0 systemconv/metric.go.
• The PR diff for those files only adds the no-attributes early return to Uptime.Record; it does not modify the referenced URLs.
• The base branch's preceding check-links runs also reported unrelated 403 failures, confirming the checker is sensitive to existing external link availability.

All code-relevant checks for this head are passing. No unrelated generated-document link edits are proposed in this PR.

[MrAlias]

The production change is focused, but the regression coverage needs to establish the one-call behavior before this is ready.

[jstar0]

Addressed the requested regression coverage in 5e0020a35.

• Added a counting-instrument regression for semconv/v1.34.0/goconv.ScheduleDuration.Record without attributes; one helper call now must produce exactly one underlying Record call.
• Tightened the generator test to enumerate all 60 affected Record methods across v1.32.0-v1.34.0 and require exactly one empty-attribute fast path for each.

Verification:

• go test ./semconvkit from internal/tools: passed.
• go test -race ./semconv/v1.34.0/goconv: passed.
• make precommit reached the repository test phase, but is currently blocked by bridge/opentracing/TestBridgeTracer_ExtractAndInject_gRPC, which fails with error reading server preface: EOF in both race and non-race runs. No production code is changed by this follow-up.

[jstar0]

Resolved the CHANGELOG.md conflict with current main via signed merge commit 2a1366e6e. The #8849 entry is kept under [Unreleased] > Fixed (the released sections are unchanged) and ./verify_released_changelog.sh main passes. No implementation files changed; the branch is mergeable again.