join: exit nonzero when the left file cannot be read (#2169) (#2171)

The join verb's left-file ingest silently swallowed read errors: a
nonexistent or malformed left file produced empty (or partial) output
with exit code 0.

Two defects, one per ingest path:

* Unsorted (half-streaming) path, ingestLeftFile in
  pkg/transformers/join.go: errors from the record-reader's error
  channel were dropped ("TODO: propagate error to caller"), as were
  record-reader construction errors. All left-file read failures were
  silent.

* Sorted (-s) path, readRecord in
  pkg/transformers/utils/join_bucket_keeper.go: the error channel was
  consulted, but the record-reader goroutine sends the error and the
  end-of-stream marker on separate channels, so the select could see
  end-of-stream first and drop the error. A nonexistent left file with
  -s exited 0 about half the time (racy).

Both paths now print "mlr: <error>" to stderr and exit 1, matching how
read errors on regular (right) input files are reported. On receipt of
the end-of-stream marker, both consumers now do a final non-blocking
drain of the error channel, which is deterministic since the reader
sends any error before the marker.

Discovered while verifying #377.

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

pkg/transformers/join.go

@@ -491,8 +491,8 @@ func (tr *TransformerJoin) ingestLeftFile() {
 	// TODO: perhaps increase recordsPerBatch, and/or refactor
 	recordReader, err := input.Create(readerOpts, 1)
 	if err != nil {
-		// TODO: propagate error to caller
+		fmt.Fprintf(os.Stderr, "mlr: %v\n", err)
-		return
+		os.Exit(1)
 	}
 
 	// Set the initial context for the left-file.
@@ -521,9 +521,8 @@ func (tr *TransformerJoin) ingestLeftFile() {
 		select {
 
 		case err := <-errorChannel:
-			// TODO: propagate error to caller
+			fmt.Fprintf(os.Stderr, "mlr: %v\n", err)
-			_ = err
+			os.Exit(1)
-			return
 
 		case leftrecsAndContexts := <-readerChannel:
 			// TODO: temp for batch-reader refactor
@@ -532,6 +531,17 @@ func (tr *TransformerJoin) ingestLeftFile() {
 			leftrecAndContext.Record = utils.KeepLeftFieldNames(leftrecAndContext.Record, tr.leftKeepFieldNameSet)
 
 			if leftrecAndContext.EndOfStream {
+				// The record-reader may have sent an error (e.g. the left file
+				// is missing or unreadable) immediately before its end-of-stream
+				// marker. Since those are separate channels, the select can see
+				// the end-of-stream marker first -- so, check the error channel
+				// before declaring the ingest complete.
+				select {
+				case err := <-errorChannel:
+					fmt.Fprintf(os.Stderr, "mlr: %v\n", err)
+					os.Exit(1)
+				default:
+				}
 				done = true
 				break // breaks the switch, not the for, in Golang
 			}

pkg/transformers/utils/join_bucket_keeper.go

@@ -553,6 +553,17 @@ func (keeper *JoinBucketKeeper) readRecord() *types.RecordAndContext {
 		leftrecAndContext := leftrecsAndContexts[0]
 		leftrecAndContext.Record = KeepLeftFieldNames(leftrecAndContext.Record, keeper.leftKeepFieldNameSet)
 		if leftrecAndContext.EndOfStream { // end-of-stream marker
+			// The record-reader may have sent an error (e.g. the left file is
+			// missing or unreadable) immediately before its end-of-stream
+			// marker. Since those are separate channels, the select can see
+			// the end-of-stream marker first -- so, check the error channel
+			// before declaring the left-file read complete.
+			select {
+			case err := <-keeper.errorChannel:
+				fmt.Fprintf(os.Stderr, "mlr: %v\n", err)
+				os.Exit(1)
+			default:
+			}
 			keeper.recordReaderDone = true
 			return nil
 		}

test/cases/verb-join/left-file-malformed-sorted/cmd

@@ -0,0 +1 @@
+mlr --icsv --opprint join -s -j a -f test/input/join-left-ragged.csv test/input/abixy.csv

test/cases/verb-join/left-file-malformed-sorted/experr

@@ -0,0 +1 @@
+mlr: CSV header/data length mismatch 2 != 4 at filename test/input/join-left-ragged.csv row 3

test/cases/verb-join/left-file-malformed-unsorted/cmd

@@ -0,0 +1 @@
+mlr --icsv --opprint join -j a -f test/input/join-left-ragged.csv test/input/abixy.csv

test/cases/verb-join/left-file-malformed-unsorted/experr

@@ -0,0 +1 @@
+mlr: CSV header/data length mismatch 2 != 4 at filename test/input/join-left-ragged.csv row 3

test/cases/verb-join/non-windows-left-file-not-found-sorted/cmd

@@ -0,0 +1 @@
+mlr --icsv --opprint join -s -j a -f /nonesuch/nope/never test/input/abixy.csv

test/cases/verb-join/non-windows-left-file-not-found-sorted/experr

@@ -0,0 +1 @@
+mlr: open /nonesuch/nope/never: no such file or directory

test/cases/verb-join/non-windows-left-file-not-found-unsorted/cmd

@@ -0,0 +1 @@
+mlr --icsv --opprint join -j a -f /nonesuch/nope/never test/input/abixy.csv

test/cases/verb-join/non-windows-left-file-not-found-unsorted/experr

@@ -0,0 +1 @@
+mlr: open /nonesuch/nope/never: no such file or directory

test/input/join-left-ragged.csv

@@ -0,0 +1,3 @@
+a,l
+pan,one
+eks,two,three,four