blob: 8724291c6c5714ea8397d9578a42bc80f2397ba1 [file] [log] [blame] [edit]
// Copyright 2013 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package godoc
import (
"bytes"
"encoding/json"
"errors"
"fmt"
"go/ast"
"go/build"
"go/doc"
"go/token"
htmlpkg "html"
htmltemplate "html/template"
"io"
"io/ioutil"
"log"
"net/http"
"os"
pathpkg "path"
"path/filepath"
"sort"
"strings"
"text/template"
"time"
"golang.org/x/tools/godoc/analysis"
"golang.org/x/tools/godoc/util"
"golang.org/x/tools/godoc/vfs"
)
// handlerServer is a migration from an old godoc http Handler type.
// This should probably merge into something else.
type handlerServer struct {
p *Presentation
c *Corpus // copy of p.Corpus
pattern string // url pattern; e.g. "/pkg/"
stripPrefix string // prefix to strip from import path; e.g. "pkg/"
fsRoot string // file system root to which the pattern is mapped; e.g. "/src"
exclude []string // file system paths to exclude; e.g. "/src/cmd"
}
func (s *handlerServer) registerWithMux(mux *http.ServeMux) {
mux.Handle(s.pattern, s)
}
// GetPageInfo returns the PageInfo for a package directory abspath. If the
// parameter genAST is set, an AST containing only the package exports is
// computed (PageInfo.PAst), otherwise package documentation (PageInfo.Doc)
// is extracted from the AST. If there is no corresponding package in the
// directory, PageInfo.PAst and PageInfo.PDoc are nil. If there are no sub-
// directories, PageInfo.Dirs is nil. If an error occurred, PageInfo.Err is
// set to the respective error but the error is not logged.
//
func (h *handlerServer) GetPageInfo(abspath, relpath string, mode PageInfoMode, goos, goarch string) *PageInfo {
info := &PageInfo{Dirname: abspath, Mode: mode}
// Restrict to the package files that would be used when building
// the package on this system. This makes sure that if there are
// separate implementations for, say, Windows vs Unix, we don't
// jumble them all together.
// Note: If goos/goarch aren't set, the current binary's GOOS/GOARCH
// are used.
ctxt := build.Default
ctxt.IsAbsPath = pathpkg.IsAbs
ctxt.IsDir = func(path string) bool {
fi, err := h.c.fs.Stat(filepath.ToSlash(path))
return err == nil && fi.IsDir()
}
ctxt.ReadDir = func(dir string) ([]os.FileInfo, error) {
f, err := h.c.fs.ReadDir(filepath.ToSlash(dir))
filtered := make([]os.FileInfo, 0, len(f))
for _, i := range f {
if mode&NoFiltering != 0 || i.Name() != "internal" {
filtered = append(filtered, i)
}
}
return filtered, err
}
ctxt.OpenFile = func(name string) (r io.ReadCloser, err error) {
data, err := vfs.ReadFile(h.c.fs, filepath.ToSlash(name))
if err != nil {
return nil, err
}
return ioutil.NopCloser(bytes.NewReader(data)), nil
}
// Make the syscall/js package always visible by default.
// It defaults to the host's GOOS/GOARCH, and golang.org's
// linux/amd64 means the wasm syscall/js package was blank.
// And you can't run godoc on js/wasm anyway, so host defaults
// don't make sense here.
if goos == "" && goarch == "" && relpath == "syscall/js" {
goos, goarch = "js", "wasm"
}
if goos != "" {
ctxt.GOOS = goos
}
if goarch != "" {
ctxt.GOARCH = goarch
}
pkginfo, err := ctxt.ImportDir(abspath, 0)
// continue if there are no Go source files; we still want the directory info
if _, nogo := err.(*build.NoGoError); err != nil && !nogo {
info.Err = err
return info
}
// collect package files
pkgname := pkginfo.Name
pkgfiles := append(pkginfo.GoFiles, pkginfo.CgoFiles...)
if len(pkgfiles) == 0 {
// Commands written in C have no .go files in the build.
// Instead, documentation may be found in an ignored file.
// The file may be ignored via an explicit +build ignore
// constraint (recommended), or by defining the package
// documentation (historic).
pkgname = "main" // assume package main since pkginfo.Name == ""
pkgfiles = pkginfo.IgnoredGoFiles
}
// get package information, if any
if len(pkgfiles) > 0 {
// build package AST
fset := token.NewFileSet()
files, err := h.c.parseFiles(fset, relpath, abspath, pkgfiles)
if err != nil {
info.Err = err
return info
}
// ignore any errors - they are due to unresolved identifiers
pkg, _ := ast.NewPackage(fset, files, poorMansImporter, nil)
// extract package documentation
info.FSet = fset
if mode&ShowSource == 0 {
// show extracted documentation
var m doc.Mode
if mode&NoFiltering != 0 {
m |= doc.AllDecls
}
if mode&AllMethods != 0 {
m |= doc.AllMethods
}
info.PDoc = doc.New(pkg, pathpkg.Clean(relpath), m) // no trailing '/' in importpath
if mode&NoTypeAssoc != 0 {
for _, t := range info.PDoc.Types {
info.PDoc.Consts = append(info.PDoc.Consts, t.Consts...)
info.PDoc.Vars = append(info.PDoc.Vars, t.Vars...)
info.PDoc.Funcs = append(info.PDoc.Funcs, t.Funcs...)
t.Consts = nil
t.Vars = nil
t.Funcs = nil
}
// for now we cannot easily sort consts and vars since
// go/doc.Value doesn't export the order information
sort.Sort(funcsByName(info.PDoc.Funcs))
}
// collect examples
testfiles := append(pkginfo.TestGoFiles, pkginfo.XTestGoFiles...)
files, err = h.c.parseFiles(fset, relpath, abspath, testfiles)
if err != nil {
log.Println("parsing examples:", err)
}
info.Examples = collectExamples(h.c, pkg, files)
// collect any notes that we want to show
if info.PDoc.Notes != nil {
// could regexp.Compile only once per godoc, but probably not worth it
if rx := h.p.NotesRx; rx != nil {
for m, n := range info.PDoc.Notes {
if rx.MatchString(m) {
if info.Notes == nil {
info.Notes = make(map[string][]*doc.Note)
}
info.Notes[m] = n
}
}
}
}
} else {
// show source code
// TODO(gri) Consider eliminating export filtering in this mode,
// or perhaps eliminating the mode altogether.
if mode&NoFiltering == 0 {
packageExports(fset, pkg)
}
info.PAst = files
}
info.IsMain = pkgname == "main"
}
// get directory information, if any
var dir *Directory
var timestamp time.Time
if tree, ts := h.c.fsTree.Get(); tree != nil && tree.(*Directory) != nil {
// directory tree is present; lookup respective directory
// (may still fail if the file system was updated and the
// new directory tree has not yet been computed)
dir = tree.(*Directory).lookup(abspath)
timestamp = ts
}
if dir == nil {
// TODO(agnivade): handle this case better, now since there is no CLI mode.
// no directory tree present (happens in command-line mode);
// compute 2 levels for this page. The second level is to
// get the synopses of sub-directories.
// note: cannot use path filter here because in general
// it doesn't contain the FSTree path
dir = h.c.newDirectory(abspath, 2)
timestamp = time.Now()
}
info.Dirs = dir.listing(true, func(path string) bool { return h.includePath(path, mode) })
info.DirTime = timestamp
info.DirFlat = mode&FlatDir != 0
return info
}
func (h *handlerServer) includePath(path string, mode PageInfoMode) (r bool) {
// if the path is under one of the exclusion paths, don't list.
for _, e := range h.exclude {
if strings.HasPrefix(path, e) {
return false
}
}
// if the path includes 'internal', don't list unless we are in the NoFiltering mode.
if mode&NoFiltering != 0 {
return true
}
if strings.Contains(path, "internal") || strings.Contains(path, "vendor") {
for _, c := range strings.Split(filepath.Clean(path), string(os.PathSeparator)) {
if c == "internal" || c == "vendor" {
return false
}
}
}
return true
}
type funcsByName []*doc.Func
func (s funcsByName) Len() int { return len(s) }
func (s funcsByName) Swap(i, j int) { s[i], s[j] = s[j], s[i] }
func (s funcsByName) Less(i, j int) bool { return s[i].Name < s[j].Name }
func (h *handlerServer) ServeHTTP(w http.ResponseWriter, r *http.Request) {
if redirect(w, r) {
return
}
relpath := pathpkg.Clean(r.URL.Path[len(h.stripPrefix)+1:])
if !h.corpusInitialized() {
h.p.ServeError(w, r, relpath, errors.New("Scan is not yet complete. Please retry after a few moments"))
return
}
abspath := pathpkg.Join(h.fsRoot, relpath)
mode := h.p.GetPageInfoMode(r)
if relpath == builtinPkgPath {
// The fake built-in package contains unexported identifiers,
// but we want to show them. Also, disable type association,
// since it's not helpful for this fake package (see issue 6645).
mode |= NoFiltering | NoTypeAssoc
}
info := h.GetPageInfo(abspath, relpath, mode, r.FormValue("GOOS"), r.FormValue("GOARCH"))
if info.Err != nil {
log.Print(info.Err)
h.p.ServeError(w, r, relpath, info.Err)
return
}
var tabtitle, title, subtitle string
switch {
case info.PAst != nil:
for _, ast := range info.PAst {
tabtitle = ast.Name.Name
break
}
case info.PDoc != nil:
tabtitle = info.PDoc.Name
default:
tabtitle = info.Dirname
title = "Directory "
if h.p.ShowTimestamps {
subtitle = "Last update: " + info.DirTime.String()
}
}
if title == "" {
if info.IsMain {
// assume that the directory name is the command name
_, tabtitle = pathpkg.Split(relpath)
title = "Command "
} else {
title = "Package "
}
}
title += tabtitle
// special cases for top-level package/command directories
switch tabtitle {
case "/src":
title = "Packages"
tabtitle = "Packages"
case "/src/cmd":
title = "Commands"
tabtitle = "Commands"
}
// Emit JSON array for type information.
pi := h.c.Analysis.PackageInfo(relpath)
hasTreeView := len(pi.CallGraph) != 0
info.CallGraphIndex = pi.CallGraphIndex
info.CallGraph = htmltemplate.JS(marshalJSON(pi.CallGraph))
info.AnalysisData = htmltemplate.JS(marshalJSON(pi.Types))
info.TypeInfoIndex = make(map[string]int)
for i, ti := range pi.Types {
info.TypeInfoIndex[ti.Name] = i
}
info.GoogleCN = googleCN(r)
var body []byte
if info.Dirname == "/src" {
body = applyTemplate(h.p.PackageRootHTML, "packageRootHTML", info)
} else {
body = applyTemplate(h.p.PackageHTML, "packageHTML", info)
}
h.p.ServePage(w, Page{
Title: title,
Tabtitle: tabtitle,
Subtitle: subtitle,
Body: body,
GoogleCN: info.GoogleCN,
TreeView: hasTreeView,
})
}
func (h *handlerServer) corpusInitialized() bool {
h.c.initMu.RLock()
defer h.c.initMu.RUnlock()
return h.c.initDone
}
type PageInfoMode uint
const (
PageInfoModeQueryString = "m" // query string where PageInfoMode is stored
NoFiltering PageInfoMode = 1 << iota // do not filter exports
AllMethods // show all embedded methods
ShowSource // show source code, do not extract documentation
FlatDir // show directory in a flat (non-indented) manner
NoTypeAssoc // don't associate consts, vars, and factory functions with types (not exposed via ?m= query parameter, used for package builtin, see issue 6645)
)
// modeNames defines names for each PageInfoMode flag.
var modeNames = map[string]PageInfoMode{
"all": NoFiltering,
"methods": AllMethods,
"src": ShowSource,
"flat": FlatDir,
}
// generate a query string for persisting PageInfoMode between pages.
func modeQueryString(mode PageInfoMode) string {
if modeNames := mode.names(); len(modeNames) > 0 {
return "?m=" + strings.Join(modeNames, ",")
}
return ""
}
// alphabetically sorted names of active flags for a PageInfoMode.
func (m PageInfoMode) names() []string {
var names []string
for name, mode := range modeNames {
if m&mode != 0 {
names = append(names, name)
}
}
sort.Strings(names)
return names
}
// GetPageInfoMode computes the PageInfoMode flags by analyzing the request
// URL form value "m". It is value is a comma-separated list of mode names
// as defined by modeNames (e.g.: m=src,text).
func (p *Presentation) GetPageInfoMode(r *http.Request) PageInfoMode {
var mode PageInfoMode
for _, k := range strings.Split(r.FormValue(PageInfoModeQueryString), ",") {
if m, found := modeNames[strings.TrimSpace(k)]; found {
mode |= m
}
}
if p.AdjustPageInfoMode != nil {
mode = p.AdjustPageInfoMode(r, mode)
}
return mode
}
// poorMansImporter returns a (dummy) package object named
// by the last path component of the provided package path
// (as is the convention for packages). This is sufficient
// to resolve package identifiers without doing an actual
// import. It never returns an error.
//
func poorMansImporter(imports map[string]*ast.Object, path string) (*ast.Object, error) {
pkg := imports[path]
if pkg == nil {
// note that strings.LastIndex returns -1 if there is no "/"
pkg = ast.NewObj(ast.Pkg, path[strings.LastIndex(path, "/")+1:])
pkg.Data = ast.NewScope(nil) // required by ast.NewPackage for dot-import
imports[path] = pkg
}
return pkg, nil
}
// globalNames returns a set of the names declared by all package-level
// declarations. Method names are returned in the form Receiver_Method.
func globalNames(pkg *ast.Package) map[string]bool {
names := make(map[string]bool)
for _, file := range pkg.Files {
for _, decl := range file.Decls {
addNames(names, decl)
}
}
return names
}
// collectExamples collects examples for pkg from testfiles.
func collectExamples(c *Corpus, pkg *ast.Package, testfiles map[string]*ast.File) []*doc.Example {
var files []*ast.File
for _, f := range testfiles {
files = append(files, f)
}
var examples []*doc.Example
globals := globalNames(pkg)
for _, e := range doc.Examples(files...) {
name := stripExampleSuffix(e.Name)
if name == "" || globals[name] {
examples = append(examples, e)
} else if c.Verbose {
log.Printf("skipping example 'Example%s' because '%s' is not a known function or type", e.Name, e.Name)
}
}
return examples
}
// addNames adds the names declared by decl to the names set.
// Method names are added in the form ReceiverTypeName_Method.
func addNames(names map[string]bool, decl ast.Decl) {
switch d := decl.(type) {
case *ast.FuncDecl:
name := d.Name.Name
if d.Recv != nil {
var typeName string
switch r := d.Recv.List[0].Type.(type) {
case *ast.StarExpr:
typeName = r.X.(*ast.Ident).Name
case *ast.Ident:
typeName = r.Name
}
name = typeName + "_" + name
}
names[name] = true
case *ast.GenDecl:
for _, spec := range d.Specs {
switch s := spec.(type) {
case *ast.TypeSpec:
names[s.Name.Name] = true
case *ast.ValueSpec:
for _, id := range s.Names {
names[id.Name] = true
}
}
}
}
}
// packageExports is a local implementation of ast.PackageExports
// which correctly updates each package file's comment list.
// (The ast.PackageExports signature is frozen, hence the local
// implementation).
//
func packageExports(fset *token.FileSet, pkg *ast.Package) {
for _, src := range pkg.Files {
cmap := ast.NewCommentMap(fset, src, src.Comments)
ast.FileExports(src)
src.Comments = cmap.Filter(src).Comments()
}
}
func applyTemplate(t *template.Template, name string, data interface{}) []byte {
var buf bytes.Buffer
if err := t.Execute(&buf, data); err != nil {
log.Printf("%s.Execute: %s", name, err)
}
return buf.Bytes()
}
type writerCapturesErr struct {
w io.Writer
err error
}
func (w *writerCapturesErr) Write(p []byte) (int, error) {
n, err := w.w.Write(p)
if err != nil {
w.err = err
}
return n, err
}
// applyTemplateToResponseWriter uses an http.ResponseWriter as the io.Writer
// for the call to template.Execute. It uses an io.Writer wrapper to capture
// errors from the underlying http.ResponseWriter. Errors are logged only when
// they come from the template processing and not the Writer; this avoid
// polluting log files with error messages due to networking issues, such as
// client disconnects and http HEAD protocol violations.
func applyTemplateToResponseWriter(rw http.ResponseWriter, t *template.Template, data interface{}) {
w := &writerCapturesErr{w: rw}
err := t.Execute(w, data)
// There are some cases where template.Execute does not return an error when
// rw returns an error, and some where it does. So check w.err first.
if w.err == nil && err != nil {
// Log template errors.
log.Printf("%s.Execute: %s", t.Name(), err)
}
}
func redirect(w http.ResponseWriter, r *http.Request) (redirected bool) {
canonical := pathpkg.Clean(r.URL.Path)
if !strings.HasSuffix(canonical, "/") {
canonical += "/"
}
if r.URL.Path != canonical {
url := *r.URL
url.Path = canonical
http.Redirect(w, r, url.String(), http.StatusMovedPermanently)
redirected = true
}
return
}
func redirectFile(w http.ResponseWriter, r *http.Request) (redirected bool) {
c := pathpkg.Clean(r.URL.Path)
c = strings.TrimRight(c, "/")
if r.URL.Path != c {
url := *r.URL
url.Path = c
http.Redirect(w, r, url.String(), http.StatusMovedPermanently)
redirected = true
}
return
}
func (p *Presentation) serveTextFile(w http.ResponseWriter, r *http.Request, abspath, relpath, title string) {
src, err := vfs.ReadFile(p.Corpus.fs, abspath)
if err != nil {
log.Printf("ReadFile: %s", err)
p.ServeError(w, r, relpath, err)
return
}
if r.FormValue(PageInfoModeQueryString) == "text" {
p.ServeText(w, src)
return
}
h := r.FormValue("h")
s := RangeSelection(r.FormValue("s"))
var buf bytes.Buffer
if pathpkg.Ext(abspath) == ".go" {
// Find markup links for this file (e.g. "/src/fmt/print.go").
fi := p.Corpus.Analysis.FileInfo(abspath)
buf.WriteString("<script type='text/javascript'>document.ANALYSIS_DATA = ")
buf.Write(marshalJSON(fi.Data))
buf.WriteString(";</script>\n")
if status := p.Corpus.Analysis.Status(); status != "" {
buf.WriteString("<a href='/lib/godoc/analysis/help.html'>Static analysis features</a> ")
// TODO(adonovan): show analysis status at per-file granularity.
fmt.Fprintf(&buf, "<span style='color: grey'>[%s]</span><br/>", htmlpkg.EscapeString(status))
}
buf.WriteString("<pre>")
formatGoSource(&buf, src, fi.Links, h, s)
buf.WriteString("</pre>")
} else {
buf.WriteString("<pre>")
FormatText(&buf, src, 1, false, h, s)
buf.WriteString("</pre>")
}
fmt.Fprintf(&buf, `<p><a href="/%s?m=text">View as plain text</a></p>`, htmlpkg.EscapeString(relpath))
p.ServePage(w, Page{
Title: title,
SrcPath: relpath,
Tabtitle: relpath,
Body: buf.Bytes(),
GoogleCN: googleCN(r),
})
}
// formatGoSource HTML-escapes Go source text and writes it to w,
// decorating it with the specified analysis links.
//
func formatGoSource(buf *bytes.Buffer, text []byte, links []analysis.Link, pattern string, selection Selection) {
// Emit to a temp buffer so that we can add line anchors at the end.
saved, buf := buf, new(bytes.Buffer)
var i int
var link analysis.Link // shared state of the two funcs below
segmentIter := func() (seg Segment) {
if i < len(links) {
link = links[i]
i++
seg = Segment{link.Start(), link.End()}
}
return
}
linkWriter := func(w io.Writer, offs int, start bool) {
link.Write(w, offs, start)
}
comments := tokenSelection(text, token.COMMENT)
var highlights Selection
if pattern != "" {
highlights = regexpSelection(text, pattern)
}
FormatSelections(buf, text, linkWriter, segmentIter, selectionTag, comments, highlights, selection)
// Now copy buf to saved, adding line anchors.
// The lineSelection mechanism can't be composed with our
// linkWriter, so we have to add line spans as another pass.
n := 1
for _, line := range bytes.Split(buf.Bytes(), []byte("\n")) {
// The line numbers are inserted into the document via a CSS ::before
// pseudo-element. This prevents them from being copied when users
// highlight and copy text.
// ::before is supported in 98% of browsers: https://caniuse.com/#feat=css-gencontent
// This is also the trick Github uses to hide line numbers.
//
// The first tab for the code snippet needs to start in column 9, so
// it indents a full 8 spaces, hence the two nbsp's. Otherwise the tab
// character only indents a short amount.
//
// Due to rounding and font width Firefox might not treat 8 rendered
// characters as 8 characters wide, and subsequently may treat the tab
// character in the 9th position as moving the width from (7.5 or so) up
// to 8. See
// https://github.com/webcompat/web-bugs/issues/17530#issuecomment-402675091
// for a fuller explanation. The solution is to add a CSS class to
// explicitly declare the width to be 8 characters.
fmt.Fprintf(saved, `<span id="L%d" class="ln">%6d&nbsp;&nbsp;</span>`, n, n)
n++
saved.Write(line)
saved.WriteByte('\n')
}
}
func (p *Presentation) serveDirectory(w http.ResponseWriter, r *http.Request, abspath, relpath string) {
if redirect(w, r) {
return
}
list, err := p.Corpus.fs.ReadDir(abspath)
if err != nil {
p.ServeError(w, r, relpath, err)
return
}
p.ServePage(w, Page{
Title: "Directory",
SrcPath: relpath,
Tabtitle: relpath,
Body: applyTemplate(p.DirlistHTML, "dirlistHTML", list),
GoogleCN: googleCN(r),
})
}
func (p *Presentation) ServeHTMLDoc(w http.ResponseWriter, r *http.Request, abspath, relpath string) {
// get HTML body contents
src, err := vfs.ReadFile(p.Corpus.fs, abspath)
if err != nil {
log.Printf("ReadFile: %s", err)
p.ServeError(w, r, relpath, err)
return
}
// if it begins with "<!DOCTYPE " assume it is standalone
// html that doesn't need the template wrapping.
if bytes.HasPrefix(src, doctype) {
w.Write(src)
return
}
// if it begins with a JSON blob, read in the metadata.
meta, src, err := extractMetadata(src)
if err != nil {
log.Printf("decoding metadata %s: %v", relpath, err)
}
page := Page{
Title: meta.Title,
Subtitle: meta.Subtitle,
GoogleCN: googleCN(r),
}
// evaluate as template if indicated
if meta.Template {
tmpl, err := template.New("main").Funcs(p.TemplateFuncs()).Parse(string(src))
if err != nil {
log.Printf("parsing template %s: %v", relpath, err)
p.ServeError(w, r, relpath, err)
return
}
var buf bytes.Buffer
if err := tmpl.Execute(&buf, page); err != nil {
log.Printf("executing template %s: %v", relpath, err)
p.ServeError(w, r, relpath, err)
return
}
src = buf.Bytes()
}
// if it's the language spec, add tags to EBNF productions
if strings.HasSuffix(abspath, "go_spec.html") {
var buf bytes.Buffer
Linkify(&buf, src)
src = buf.Bytes()
}
page.Body = src
p.ServePage(w, page)
}
func (p *Presentation) ServeFile(w http.ResponseWriter, r *http.Request) {
p.serveFile(w, r)
}
func (p *Presentation) serveFile(w http.ResponseWriter, r *http.Request) {
if strings.HasSuffix(r.URL.Path, "/index.html") {
// We'll show index.html for the directory.
// Use the dir/ version as canonical instead of dir/index.html.
http.Redirect(w, r, r.URL.Path[0:len(r.URL.Path)-len("index.html")], http.StatusMovedPermanently)
return
}
// Check to see if we need to redirect or serve another file.
relpath := r.URL.Path
if m := p.Corpus.MetadataFor(relpath); m != nil {
if m.Path != relpath {
// Redirect to canonical path.
http.Redirect(w, r, m.Path, http.StatusMovedPermanently)
return
}
// Serve from the actual filesystem path.
relpath = m.filePath
}
abspath := relpath
relpath = relpath[1:] // strip leading slash
switch pathpkg.Ext(relpath) {
case ".html":
p.ServeHTMLDoc(w, r, abspath, relpath)
return
case ".go":
p.serveTextFile(w, r, abspath, relpath, "Source file")
return
}
dir, err := p.Corpus.fs.Lstat(abspath)
if err != nil {
log.Print(err)
p.ServeError(w, r, relpath, err)
return
}
if dir != nil && dir.IsDir() {
if redirect(w, r) {
return
}
if index := pathpkg.Join(abspath, "index.html"); util.IsTextFile(p.Corpus.fs, index) {
p.ServeHTMLDoc(w, r, index, index)
return
}
p.serveDirectory(w, r, abspath, relpath)
return
}
if util.IsTextFile(p.Corpus.fs, abspath) {
if redirectFile(w, r) {
return
}
p.serveTextFile(w, r, abspath, relpath, "Text file")
return
}
p.fileServer.ServeHTTP(w, r)
}
func (p *Presentation) ServeText(w http.ResponseWriter, text []byte) {
w.Header().Set("Content-Type", "text/plain; charset=utf-8")
w.Write(text)
}
func marshalJSON(x interface{}) []byte {
var data []byte
var err error
const indentJSON = false // for easier debugging
if indentJSON {
data, err = json.MarshalIndent(x, "", " ")
} else {
data, err = json.Marshal(x)
}
if err != nil {
panic(fmt.Sprintf("json.Marshal failed: %s", err))
}
return data
}