feat: implement -j raw-errors flag for ODBC sqlcmd compatibility by dlevy-msft-sql · Pull Request #759 · microsoft/go-sqlcmd

dlevy-msft-sql · 2026-05-19T21:25:57Z

Problem

go-sqlcmd did not implement the -j switch from the ODBC-based sqlcmd, which is required to preserve raw driver-supplied error text. This was tracked in discussion #292 as a gap blocking ODBC sqlcmd retirement.

Root Cause

The default formatter (pkg/sqlcmd/format.go) unconditionally stripped the "mssql: " prefix that go-mssqldb prepends to every mssql.Error.Error() string. There was no way for a caller to opt out, and no CLI flag exposed the behavior.

Solution

Implement -j / --raw-errors with semantics that match the ODBC reference (Sql/utils/sqlcmd/console/src/Formatter.cpp, gated on m_PrintRawErrorMessages):

When -j is set, the driver-supplied prefix is preserved verbatim.
The Msg N, Level N, State N, Server S, Line N header is always emitted (matches ODBC).
Screen-width wrapping via fitToScreen is always applied (matches ODBC).

The formatter is extended via a non-breaking variadic functional-options pattern, so existing callers (including internal/sql/mssql.go) compile unchanged.

Changes

File	Change
`pkg/sqlcmd/format.go`	Add `rawErrors` field, `FormatterOption` type, `WithRawErrors`; gate `strings.TrimPrefix(msg, "mssql: ")` in `AddError`.
`pkg/sqlcmd/format_test.go`	Add `TestAddErrorStripsMssqlPrefixByDefault` and `TestAddErrorWithRawErrorsKeepsMssqlPrefix` — two parallel tests, each readable end-to-end, covering both gate states.
`cmd/sqlcmd/sqlcmd.go`	Add `RawErrors` to `SQLCmdArguments`; register `-j` / `--raw-errors`; pass `WithRawErrors` into the formatter.
`cmd/sqlcmd/sqlcmd_test.go`	Add `-j` case to `TestValidCommandLineToArgsConversion`.

Testing

go build ./... clean.
go vet ./... clean (only pre-existing credman warnings on windows/v0.21.0).
go test ./pkg/sqlcmd/ ./cmd/sqlcmd/ -count=1 against a live SQL Server: PASS.
The two TestAddError… tests assert the header and message body appear in both modes and that the "mssql: " prefix is stripped exactly when -j is unset (would fail in either direction if the gate logic regressed).

Related Issues

Closes the -j gap in discussion #292.

Adds the -j / --raw-errors flag to the legacy sqlcmd CLI, mirroring the ODBC sqlcmd's m_PrintRawErrorMessages behavior (Sql/utils/sqlcmd/console/src/Formatter.cpp). When set, the driver-supplied 'mssql: ' prefix is preserved in error output; the Msg/Level/State/Server header and screen-width wrapping are unaffected, matching ODBC sqlcmd which also only gates the prefix strip. Closes the -j gap tracked in discussion microsoft#292.

Copilot

Pull request overview

Note

Copilot was unable to run its full agentic suite in this review.

Adds ODBC-compatible -j/--raw-errors support so go-sqlcmd can optionally preserve the go-mssqldb "mssql: " error prefix, improving parity with ODBC sqlcmd behavior.

Changes:

Extend the default formatter with a functional-options pattern and a WithRawErrors option.
Gate stripping of the "mssql: " prefix in AddError based on the new option.
Wire -j/--raw-errors through CLI args and add targeted test coverage.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

File	Description
pkg/sqlcmd/format.go	Adds `rawErrors` option plumbing and gates `mssql:` prefix trimming in `AddError`.
pkg/sqlcmd/format_test.go	Adds coverage validating prefix stripping behavior in default vs raw modes.
cmd/sqlcmd/sqlcmd.go	Introduces `RawErrors` CLI flag and passes formatter option through.
cmd/sqlcmd/sqlcmd_test.go	Adds CLI parsing test for `-j`.

Comments suppressed due to low confidence (1)

pkg/sqlcmd/format.go:1

The docstring is internally inconsistent: it describes the prefix as "driver-supplied" while also stating it is prepended by go-mssqldb. Consider rewording to clarify that -j preserves the go-mssqldb-added prefix (or more generally, preserves the error string exactly as returned by the Go driver) to avoid confusing API consumers.

// Copyright (c) Microsoft Corporation.

dlevy-msft-sql · 2026-05-19T21:43:49Z

Took the docstring feedback as well — reworded both the WithRawErrors docstring and the -j flag help text in 878e6d4 to explicitly name the go-mssqldb client driver's mssql: prefix instead of the ambiguous "driver-supplied" phrasing.

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 4 comments.

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 3 comments.

dlevy-msft-sql · 2026-05-19T22:33:41Z

 	rootCmd.Flags().IntVar(&args.DriverLoggingLevel, "driver-logging-level", 0, localizer.Sprintf("Level of mssql driver messages to print"))
 	rootCmd.Flags().BoolVarP(&args.ExitOnError, "exit-on-error", "b", false, localizer.Sprintf("Specifies that sqlcmd exits and returns a %s value when an error occurs", localizer.DosErrorLevel))
 	rootCmd.Flags().IntVarP(&args.ErrorLevel, "error-level", "m", 0, localizer.Sprintf("Controls which error messages are sent to %s. Messages that have severity level greater than or equal to this level are sent", localizer.StdoutName))
+	rootCmd.Flags().BoolVarP(&args.RawErrors, "raw-errors", "j", false, localizer.Sprintf("Do not strip the \"mssql: \" prefix from error messages"))


Going to push back on this one, similar to the test-coupling thread above.

The whole point of CLI flag help is to let a user predict the concrete diff they'll see in their output. With -j set, every error line will literally contain mssql: where it wouldn't otherwise; that's user-observable text, not implementation. Replacing the concrete string with abstract phrasing like "preserve raw driver error text" makes the help less actionable — a user reading it can't tell what -j will actually change in their terminal.

format.go hardcodes strings.TrimPrefix(msg, "mssql: ") against that exact literal. If go-mssqldb ever changes its prefix format, the gate, the help text, and the tests all need to update together — that coupling is intentional and consistent with how the rest of the flag help in this file describes concrete behavior (see -b, -h, etc.).

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 3 comments.

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 5 comments.

… long-form case

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

dlevy-msft-sql requested a review from Copilot May 19, 2026 21:27

dlevy-msft-sql added this to the v1.12.0 - ODBC SQLCMD Interchangeability milestone May 19, 2026

Copilot AI reviewed May 19, 2026

View reviewed changes

Comment thread pkg/sqlcmd/format.go Outdated

Comment thread pkg/sqlcmd/format_test.go

Copilot started reviewing on behalf of dlevy-msft-sql May 19, 2026 21:39 View session

dlevy-msft-sql added 2 commits May 19, 2026 16:41

refactor: skip nil FormatterOption to harden variadic API

881b158

docs: clarify -j wording to name the go-mssqldb prefix explicitly

878e6d4

dlevy-msft-sql requested a review from Copilot May 19, 2026 21:47

Copilot AI reviewed May 19, 2026

View reviewed changes

Comment thread cmd/sqlcmd/sqlcmd.go Outdated

Comment thread pkg/sqlcmd/format.go Outdated

Comment thread pkg/sqlcmd/format_test.go Outdated

Comment thread pkg/sqlcmd/format.go

refactor: trim -j docs and split test for DAMP readability

456773a

dlevy-msft-sql requested a review from Copilot May 19, 2026 22:02

Copilot AI reviewed May 19, 2026

View reviewed changes

Comment thread pkg/sqlcmd/format_test.go

Comment thread cmd/sqlcmd/sqlcmd.go Outdated

Copilot started reviewing on behalf of dlevy-msft-sql May 19, 2026 22:14 View session

docs: drop driver name from -j help text

7b23593

dlevy-msft-sql requested a review from Copilot May 19, 2026 22:21

Copilot started reviewing on behalf of dlevy-msft-sql May 19, 2026 22:25 View session

Copilot AI reviewed May 19, 2026

View reviewed changes

Copilot started reviewing on behalf of dlevy-msft-sql May 19, 2026 22:29 View session

refactor: unexport formatterOption for closed-set options

d0b37a4

dlevy-msft-sql requested a review from Copilot May 19, 2026 22:37

Copilot AI reviewed May 19, 2026

View reviewed changes

Comment thread pkg/sqlcmd/format.go Outdated

Comment thread pkg/sqlcmd/format_test.go Outdated

refactor: re-export FormatterOption for public SDK consumers

932fb6d

dlevy-msft-sql requested a review from Copilot May 19, 2026 22:50

Copilot AI reviewed May 19, 2026

View reviewed changes

Comment thread pkg/sqlcmd/format_test.go

Comment thread cmd/sqlcmd/sqlcmd.go

test: tighten default-mode assertion to forbid any mssql: occurrence

fc8c7d4

dlevy-msft-sql requested a review from Copilot May 19, 2026 23:05

Copilot started reviewing on behalf of dlevy-msft-sql May 19, 2026 23:07 View session

Copilot AI reviewed May 19, 2026

View reviewed changes

Comment thread cmd/sqlcmd/sqlcmd.go

Comment thread pkg/sqlcmd/format.go Outdated

Comment thread pkg/sqlcmd/format_test.go

docs: decouple WithRawErrors doc from CLI flag name

a65bd09

dlevy-msft-sql requested a review from Copilot May 19, 2026 23:17

Copilot started reviewing on behalf of dlevy-msft-sql May 19, 2026 23:23 View session

Copilot AI reviewed May 19, 2026

View reviewed changes

Comment thread pkg/sqlcmd/format.go Outdated

Comment thread pkg/sqlcmd/format.go Outdated

Comment thread cmd/sqlcmd/sqlcmd.go

Comment thread cmd/sqlcmd/sqlcmd_test.go

Comment thread pkg/sqlcmd/format_test.go

docs: document FormatterOption and opts param; test: add --raw-errors…

7270f01

… long-form case

Copilot started reviewing on behalf of dlevy-msft-sql May 19, 2026 23:27 View session

dlevy-msft-sql requested a review from Copilot May 20, 2026 03:41

Copilot started reviewing on behalf of dlevy-msft-sql May 20, 2026 03:43 View session

Copilot AI reviewed May 20, 2026

View reviewed changes

Comment thread pkg/sqlcmd/format.go

dlevy-msft-sql marked this pull request as ready for review May 20, 2026 03:57

Conversation

dlevy-msft-sql commented May 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Problem

Root Cause

Solution

Changes

Testing

Related Issues

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

dlevy-msft-sql commented May 19, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

dlevy-msft-sql May 19, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

dlevy-msft-sql commented May 19, 2026 •

edited

Loading