From e0d2ede90d34fc35440c414228a536c1c869dbfc Mon Sep 17 00:00:00 2001
From: Silvan Jegen <s.jegen@gmail.com>
Date: Thu, 1 Dec 2016 22:27:15 +0100
Subject: Make sure to get text in sub-elements

We use the simpler API offered by the Go standard library to extract
the article titles. I could not find another way to make sure I get the
nested sub-elements.
---
 goencxml.go | 51 ++++++++++++++++++++++++++++++++++++++++++---------
 1 file changed, 42 insertions(+), 9 deletions(-)

diff --git a/goencxml.go b/goencxml.go
index b4fb7cf..35ddf5c 100644
--- a/goencxml.go
+++ b/goencxml.go
@@ -2,23 +2,56 @@ package main
 
 import (
 	"bufio"
+	"bytes"
 	"encoding/xml"
 	"fmt"
+	"io"
 	"os"
 )
 
-type article struct {
-	Title string `xml:"front>article-meta>title-group>article-title"`
-}
-
 func process(r *bufio.Reader) {
-	var a article
+	var (
+		intitle       bool
+		inarticlemeta bool
+		buffer        bytes.Buffer
+	)
 
-	err := xml.NewDecoder(r).Decode(&a)
-	if err != nil {
-		fmt.Fprintf(os.Stderr, "Error when decoding XML file %q\n", err)
+	dec := xml.NewDecoder(r)
+	for {
+		token, err := dec.Token()
+		if err == io.EOF {
+			break
+		}
+		if err != nil {
+			fmt.Fprintf(os.Stderr, "Error when decoding XML file %q\n", err)
+			os.Exit(1)
+		}
+		switch t := token.(type) {
+		case xml.StartElement:
+			if t.Name.Local == "article-meta" {
+				inarticlemeta = true
+				continue
+			}
+			if t.Name.Local == "article-title" && inarticlemeta {
+				intitle = true
+			}
+		case xml.CharData:
+			if !intitle || !inarticlemeta {
+				continue
+			}
+			buffer.Write(t)
+		case xml.EndElement:
+			if t.Name.Local == "article-meta" {
+				inarticlemeta = false
+				continue
+			}
+			if t.Name.Local == "article-title" && inarticlemeta {
+				intitle = false
+				fmt.Printf("article-title: %s\n", buffer.String())
+				buffer.Reset()
+			}
+		}
 	}
-	fmt.Printf("article-title: %s\n", a.Title)
 }
 
 func main() {
-- 
cgit v1.2.3