feat: submit sqlh to awesome-go, add missing tests and badge (#38)

[kirill-scherba]

This PR prepares sqlh for the awesome-go list submission (issue #38) and adds missing test coverage for previously untested helper functions.

Changes

Quality improvements for awesome-go criteria

• Added tests for SetWheresJoinAnd, SetDistinct, SetName, SetNumRows/GetNumRows
• Added test for InsertWithCallback error/rollback path
• Added test for Table[T].InsertId wrapper
• Added test for Table[T].Close no-op behavior
• Added test for getLastInsertID SQLite path (via transaction)
• Added test for detectDialect and dialectFromQuerier with *sql.Tx fallback
• Added test for rebind passthrough for non-PostgreSQL dialects

Documentation

• Added "Mentioned in Awesome Go" badge to README.md
• Updated docs/progress.md with awesome-go submission status
• Updated docs/projectbrief.md with current version and distribution links

awesome-go PR

• Forked avelino/awesome-go → kirill-scherba/awesome-go
• Submitted PR #6401 to avelino/awesome-go (SQL Query Builders section)
• Entry is alphabetically ordered and follows awesome-go formatting rules

Verification

• All tests pass (go test ./...)
• Main package coverage: ~75% (improved from ~73%)
• First commit: 2024-09-21 (well over 5 months)
• SemVer releases: v0.8.0, v0.7.1, v0.7.0, etc.
• CI configured: .github/workflows/test.yml with SQLite, MySQL, PostgreSQL
• BSD-3 license, go.mod, README, pkg.go.dev docs present

Closes #38

[kirill-scherba]

Verdict: REQUEST CHANGES

Architecture

• The PR keeps the implementation architecture intact: no production API or SQL generation path is changed, and the added tests are isolated in the existing SQLite test suite. That part fits the existing design.

Correctness

• Blocking: README.md adds the Mentioned in Awesome Go badge before sqlh is actually accepted/listed in avelino/awesome-go. The upstream PR is still open (avelino/awesome-go#6401), so this badge is currently a false project status signal. awesome-go's own CONTRIBUTING guide presents this badge under the post-acceptance section ("Congrats, your project got accepted - what now"), not as a submission badge. Do not merge this badge until the upstream PR is accepted. Fix by either removing the README badge from this PR, or replacing it with a neutral note elsewhere that says the submission PR is pending.
• Non-blocking cleanup: the TestGetLastInsertID/Unsupported driver subtest name/comment does not match what it tests. It opens SQLite and asserts detectDialect(db) does not panic; it does not cover an unsupported driver. Rename the subtest/comment to reflect the real behavior, or add a true fake-driver test if unsupported-driver coverage is intended.

Tests

• go test ./... passes locally in the worktree.
• The added tests improve coverage for helper constructors, rollback behavior, table wrapper InsertId, SQLite getLastInsertID, and dialect fallback for *sql.Tx. No production behavior regression found.

Docs / Memory Bank consistency

• docs/progress.md and docs/projectbrief.md were updated, which is correct because this PR changes project distribution/marketing status.
• However, README docs are not consistent with reality until awesome-go accepts the upstream PR. Keep Memory Bank entries that say "submitted / awaiting review"; remove the accepted-status badge from README until acceptance.

Required fix before approval

1. Remove the Mentioned in Awesome Go badge from README.md until avelino/awesome-go#6401 is merged.
2. Optionally rename the misleading Unsupported driver test subcase/comment.

No ai/approved label added.

[kirill-scherba]

Verdict: APPROVE

Architecture

• The PR remains correctly scoped: no production API, SQL generation, or transaction semantics are changed.
• Added tests are isolated to the existing SQLite-focused test suite and exercise helper constructors, callback rollback behavior, table wrapper delegation, SQLite getLastInsertID, dialect detection, *sql.Tx fallback handling, and non-PostgreSQL rebind passthrough.
• The prior README badge issue is fixed: the repository no longer claims accepted awesome-go status before upstream acceptance.

Correctness

• The blocking issue from the previous review is resolved: the premature Mentioned in Awesome Go badge was removed from README.md.
• The misleading Unsupported driver subtest was renamed to detectDialect does not panic with real driver, and its comment now matches the actual behavior under test.
• The awesome-go submission status remains correctly represented as pending via avelino/awesome-go#6401.
• No behavior regressions found in the production code because only tests and documentation are changed.

Tests

• Verified locally in the worktree with go test -count=1 ./....
• Result: all packages pass.

Docs / Memory Bank consistency

• docs/progress.md now correctly says the README badge is deferred until upstream acceptance.
• docs/projectbrief.md accurately records that sqlh was submitted to awesome-go PR #6401, without claiming acceptance.
• Documentation state is now consistent with the actual repository state and upstream PR lifecycle.

No blocking issues remain. ai/approved labels may be applied to PR #43 and issue #38.

[kirill-scherba]

Approved after fixes review.

All review findings from the earlier round were resolved:

1. Premature Mentioned in Awesome Go badge removed from README.
2. Misleading Unsupported driver subtest renamed and commented correctly.
3. Memory Bank docs (docs/progress.md, docs/projectbrief.md) updated to reflect pending status.

ai/approved was already set on PR #43 and issue #38. approved label added — orchestrator will merge automatically.