fix #522: run misspell in text mode

Treat Go source files as plain text files by misspell: it allows detecting
issues in strings, variable names, etc. Also, it's the default mode of
a standalone misspell tool.

Also, implement richer and more stable auto-fix of misspell issues:
now it can fix multiple issues in one line.

Author: jirfag

pkg/commands/executor.go

@@ -96,9 +96,9 @@ func NewExecutor(version, commit, date string) *Executor {
 	e.EnabledLintersSet = lintersdb.NewEnabledSet(e.DBManager,
 		lintersdb.NewValidator(e.DBManager), e.log.Child("lintersdb"), e.cfg)
 	e.goenv = goutil.NewEnv(e.log.Child("goenv"))
-	e.contextLoader = lint.NewContextLoader(e.cfg, e.log.Child("loader"), e.goenv)
 	e.fileCache = fsutils.NewFileCache()
 	e.lineCache = fsutils.NewLineCache(e.fileCache)
+	e.contextLoader = lint.NewContextLoader(e.cfg, e.log.Child("loader"), e.goenv, e.lineCache, e.fileCache)
 
 	return e
 }

pkg/golinters/gocritic.go

@@ -109,7 +109,7 @@ func (lint Gocritic) Run(ctx context.Context, lintCtx *linter.Context) ([]result
 	go func() {
 		defer func() {
 			if err := recover(); err != nil {
-				panicErr = fmt.Errorf("panic occured: %s", err)
+				panicErr = fmt.Errorf("panic occurred: %s", err)
 				lintCtx.Log.Warnf("Panic: %s", debug.Stack())
 			}
 		}()

pkg/golinters/misspell.go

@@ -4,7 +4,6 @@ import (
 	"context"
 	"fmt"
 	"go/token"
-	"io/ioutil"
 	"strings"
 
 	"github.com/golangci/misspell"
@@ -15,6 +14,10 @@ import (
 
 type Misspell struct{}
 
+func NewMisspell() *Misspell {
+	return &Misspell{}
+}
+
 func (Misspell) Name() string {
 	return "misspell"
 }
@@ -50,25 +53,52 @@ func (lint Misspell) Run(ctx context.Context, lintCtx *linter.Context) ([]result
 
 	var res []result.Issue
 	for _, f := range getAllFileNames(lintCtx) {
-		fileContent, err := ioutil.ReadFile(f)
+		issues, err := lint.runOnFile(f, &r, lintCtx)
 		if err != nil {
-			return nil, fmt.Errorf("can't read file %s: %s", f, err)
+			return nil, err
 		}
+		res = append(res, issues...)
+	}
+
+	return res, nil
+}
 
-		_, diffs := r.ReplaceGo(string(fileContent))
-		for _, diff := range diffs {
-			text := fmt.Sprintf("`%s` is a misspelling of `%s`", diff.Original, diff.Corrected)
-			pos := token.Position{
-				Filename: f,
-				Line:     diff.Line,
-				Column:   diff.Column + 1,
+func (lint Misspell) runOnFile(fileName string, r *misspell.Replacer, lintCtx *linter.Context) ([]result.Issue, error) {
+	var res []result.Issue
+	fileContent, err := lintCtx.FileCache.GetFileBytes(fileName)
+	if err != nil {
+		return nil, fmt.Errorf("can't get file %s contents: %s", fileName, err)
+	}
+
+	// use r.Replace, not r.ReplaceGo because r.ReplaceGo doesn't find
+	// issues inside strings: it searches only inside comments. r.Replace
+	// searches all words: it treats input as a plain text. A standalone misspell
+	// tool uses r.Replace by default.
+	_, diffs := r.Replace(string(fileContent))
+	for _, diff := range diffs {
+		text := fmt.Sprintf("`%s` is a misspelling of `%s`", diff.Original, diff.Corrected)
+		pos := token.Position{
+			Filename: fileName,
+			Line:     diff.Line,
+			Column:   diff.Column + 1,
+		}
+		var replacement *result.Replacement
+		if lintCtx.Cfg.Issues.NeedFix {
+			replacement = &result.Replacement{
+				Inline: &result.InlineFix{
+					StartCol:  diff.Column,
+					Length:    len(diff.Original),
+					NewString: diff.Corrected,
+				},
 			}
-			res = append(res, result.Issue{
-				Pos:        pos,
-				Text:       text,
-				FromLinter: lint.Name(),
-			})
 		}
+
+		res = append(res, result.Issue{
+			Pos:         pos,
+			Text:        text,
+			FromLinter:  lint.Name(),
+			Replacement: replacement,
+		})
 	}
 
 	return res, nil

pkg/lint/linter/context.go

@@ -5,6 +5,8 @@ import (
 	"golang.org/x/tools/go/packages"
 	"golang.org/x/tools/go/ssa"
 
+	"github.com/golangci/golangci-lint/pkg/fsutils"
+
 	"github.com/golangci/golangci-lint/pkg/config"
 	"github.com/golangci/golangci-lint/pkg/lint/astcache"
 	"github.com/golangci/golangci-lint/pkg/logutils"
@@ -19,9 +21,11 @@ type Context struct {
 
 	SSAProgram *ssa.Program // for unparam and interfacer but not for megacheck (it change it)
 
-	Cfg      *config.Config
-	ASTCache *astcache.Cache
-	Log      logutils.Log
+	Cfg       *config.Config
+	ASTCache  *astcache.Cache
+	FileCache *fsutils.FileCache
+	LineCache *fsutils.LineCache
+	Log       logutils.Log
 }
 
 func (c *Context) Settings() *config.LintersSettings {

pkg/lint/load.go

@@ -11,6 +11,8 @@ import (
 	"strings"
 	"time"
 
+	"github.com/golangci/golangci-lint/pkg/fsutils"
+
 	"github.com/pkg/errors"
 	"golang.org/x/tools/go/loader"
 	"golang.org/x/tools/go/packages"
@@ -31,15 +33,21 @@ type ContextLoader struct {
 	debugf      logutils.DebugFunc
 	goenv       *goutil.Env
 	pkgTestIDRe *regexp.Regexp
+	lineCache   *fsutils.LineCache
+	fileCache   *fsutils.FileCache
 }
 
-func NewContextLoader(cfg *config.Config, log logutils.Log, goenv *goutil.Env) *ContextLoader {
+func NewContextLoader(cfg *config.Config, log logutils.Log, goenv *goutil.Env,
+	lineCache *fsutils.LineCache, fileCache *fsutils.FileCache) *ContextLoader {
+
 	return &ContextLoader{
 		cfg:         cfg,
 		log:         log,
 		debugf:      logutils.Debug("loader"),
 		goenv:       goenv,
 		pkgTestIDRe: regexp.MustCompile(`^(.*) \[(.*)\.test\]`),
+		lineCache:   lineCache,
+		fileCache:   fileCache,
 	}
 }
 
@@ -356,9 +364,11 @@ func (cl ContextLoader) Load(ctx context.Context, linters []*linter.Config) (*li
 			Cwd:   "",  // used by depguard and fallbacked to os.Getcwd
 			Build: nil, // used by depguard and megacheck and fallbacked to build.Default
 		},
-		Cfg:      cl.cfg,
-		ASTCache: astCache,
-		Log:      cl.log,
+		Cfg:       cl.cfg,
+		ASTCache:  astCache,
+		Log:       cl.log,
+		FileCache: cl.fileCache,
+		LineCache: cl.lineCache,
 	}
 
 	separateNotCompilingPackages(ret)

pkg/lint/runner.go

@@ -79,13 +79,12 @@ func NewRunner(astCache *astcache.Cache, cfg *config.Config, log logutils.Log, g
 			processors.NewExcludeRules(excludeRules, lineCache, log.Child("exclude_rules")),
 			processors.NewNolint(astCache, log.Child("nolint"), dbManager),
 
-			processors.NewUniqByLine(),
+			processors.NewUniqByLine(cfg),
 			processors.NewDiff(icfg.Diff, icfg.DiffFromRevision, icfg.DiffPatchFilePath),
 			processors.NewMaxPerFileFromLinter(cfg),
 			processors.NewMaxSameIssues(icfg.MaxSameIssues, log.Child("max_same_issues"), cfg),
 			processors.NewMaxFromLinter(icfg.MaxIssuesPerLinter, log.Child("max_from_linter"), cfg),
 			processors.NewSourceCode(lineCache, log.Child("source_code")),
-			processors.NewReplacementBuilder(log.Child("replacement_builder")), // must be after source code
 			processors.NewPathShortener(),
 		},
 		Log: log,

pkg/result/issue.go

@@ -9,6 +9,13 @@ type Range struct {
 type Replacement struct {
 	NeedOnlyDelete bool     // need to delete all lines of the issue without replacement with new lines
 	NewLines       []string // is NeedDelete is false it's the replacement lines
+	Inline         *InlineFix
+}
+
+type InlineFix struct {
+	StartCol  int // zero-based
+	Length    int // length of chunk to be replaced
+	NewString string
 }
 
 type Issue struct {

pkg/result/processors/fixer.go

@@ -77,6 +77,19 @@ func (f Fixer) fixIssuesInFile(filePath string, issues []result.Issue) error {
 		return errors.Wrapf(err, "failed to make file %s", tmpFileName)
 	}
 
+	// merge multiple issues per line into one issue
+	issuesPerLine := map[int][]result.Issue{}
+	for _, i := range issues {
+		issuesPerLine[i.Line()] = append(issuesPerLine[i.Line()], i)
+	}
+
+	issues = issues[:0] // reuse the same memory
+	for line, lineIssues := range issuesPerLine {
+		if mergedIssue := f.mergeLineIssues(line, lineIssues, origFileLines); mergedIssue != nil {
+			issues = append(issues, *mergedIssue)
+		}
+	}
+
 	issues = f.findNotIntersectingIssues(issues)
 
 	if err = f.writeFixedFile(origFileLines, issues, tmpOutFile); err != nil {
@@ -94,9 +107,76 @@ func (f Fixer) fixIssuesInFile(filePath string, issues []result.Issue) error {
 	return nil
 }
 
+//nolint:gocyclo
+func (f Fixer) mergeLineIssues(lineNum int, lineIssues []result.Issue, origFileLines [][]byte) *result.Issue {
+	origLine := origFileLines[lineNum-1] // lineNum is 1-based
+
+	if len(lineIssues) == 1 && lineIssues[0].Replacement.Inline == nil {
+		return &lineIssues[0]
+	}
+
+	// check issues first
+	for _, i := range lineIssues {
+		if i.LineRange != nil {
+			f.log.Infof("Line %d has multiple issues but at least one of them is ranged: %#v", lineNum, lineIssues)
+			return &lineIssues[0]
+		}
+
+		r := i.Replacement
+		if r.Inline == nil || len(r.NewLines) != 0 || r.NeedOnlyDelete {
+			f.log.Infof("Line %d has multiple issues but at least one of them isn't inline: %#v", lineNum, lineIssues)
+			return &lineIssues[0]
+		}
+
+		if r.Inline.StartCol < 0 || r.Inline.Length <= 0 || r.Inline.StartCol+r.Inline.Length > len(origLine) {
+			f.log.Warnf("Line %d (%q) has invalid inline fix: %#v, %#v", lineNum, origLine, i, r.Inline)
+			return nil
+		}
+	}
+
+	return f.applyInlineFixes(lineIssues, origLine, lineNum)
+}
+
+func (f Fixer) applyInlineFixes(lineIssues []result.Issue, origLine []byte, lineNum int) *result.Issue {
+	sort.Slice(lineIssues, func(i, j int) bool {
+		return lineIssues[i].Replacement.Inline.StartCol < lineIssues[j].Replacement.Inline.StartCol
+	})
+
+	var newLineBuf bytes.Buffer
+	newLineBuf.Grow(len(origLine))
+
+	//nolint:misspell
+	// example: origLine="it's becouse of them", StartCol=5, Length=7, NewString="because"
+
+	curOrigLinePos := 0
+	for _, i := range lineIssues {
+		fix := i.Replacement.Inline
+		if fix.StartCol < curOrigLinePos {
+			f.log.Warnf("Line %d has multiple intersecting issues: %#v", lineNum, lineIssues)
+			return nil
+		}
+
+		if curOrigLinePos != fix.StartCol {
+			newLineBuf.Write(origLine[curOrigLinePos:fix.StartCol])
+		}
+		newLineBuf.WriteString(fix.NewString)
+		curOrigLinePos = fix.StartCol + fix.Length
+	}
+	if curOrigLinePos != len(origLine) {
+		newLineBuf.Write(origLine[curOrigLinePos:])
+	}
+
+	mergedIssue := lineIssues[0] // use text from the first issue (it's not really used)
+	mergedIssue.Replacement = &result.Replacement{
+		NewLines: []string{newLineBuf.String()},
+	}
+	return &mergedIssue
+}
+
 func (f Fixer) findNotIntersectingIssues(issues []result.Issue) []result.Issue {
 	sort.SliceStable(issues, func(i, j int) bool {
-		return issues[i].Line() < issues[j].Line() //nolint:scopelint
+		a, b := issues[i], issues[j] //nolint:scopelint
+		return a.Line() < b.Line()
 	})
 
 	var ret []result.Issue

pkg/result/processors/max_from_linter.go

@@ -33,12 +33,12 @@ func (p *MaxFromLinter) Process(issues []result.Issue) ([]result.Issue, error) {
 		return issues, nil
 	}
 
-	if p.cfg.Issues.NeedFix {
-		// we need to fix all issues at once => we need to return all of them
-		return issues, nil
-	}
-
 	return filterIssues(issues, func(i *result.Issue) bool {
+		if i.Replacement != nil && p.cfg.Issues.NeedFix {
+			// we need to fix all issues at once => we need to return all of them
+			return true
+		}
+
 		p.lc[i.FromLinter]++ // always inc for stat
 		return p.lc[i.FromLinter] <= p.limit
 	}), nil

pkg/result/processors/max_same_issues.go

@@ -38,12 +38,12 @@ func (p *MaxSameIssues) Process(issues []result.Issue) ([]result.Issue, error) {
 		return issues, nil
 	}
 
-	if p.cfg.Issues.NeedFix {
-		// we need to fix all issues at once => we need to return all of them
-		return issues, nil
-	}
-
 	return filterIssues(issues, func(i *result.Issue) bool {
+		if i.Replacement != nil && p.cfg.Issues.NeedFix {
+			// we need to fix all issues at once => we need to return all of them
+			return true
+		}
+
 		p.tc[i.Text]++ // always inc for stat
 		return p.tc[i.Text] <= p.limit
 	}), nil