Archives/Athou_commafeed
1
0
Fork 1
mirror of https://github.com/Athou/commafeed.git synced 2026-03-21 21:37:29 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix encoding issues with idiots using word for writing content

Browse Source
This commit is contained in:
Athou
2013-04-06 16:54:49 +02:00
parent 381733deeb
commit b2421eb1ff
1 changed files with 13 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
src/main/java/com/commafeed/backend/feeds/FeedParser.java
View File

@@ -7,6 +7,8 @@ import java.util.List;
import org.apache.commons.lang.StringUtils; import org.apache.commons.lang.StringUtils;
import org.apache.commons.lang.SystemUtils; import org.apache.commons.lang.SystemUtils;
import org.jsoup.Jsoup; import org.jsoup.Jsoup;
import org.jsoup.nodes.Document.OutputSettings;
import org.jsoup.nodes.Entities.EscapeMode;
import org.jsoup.safety.Whitelist; import org.jsoup.safety.Whitelist;
import org.xml.sax.InputSource; import org.xml.sax.InputSource;
@@ -28,8 +30,14 @@ public class FeedParser {
feed.setLastUpdated(Calendar.getInstance().getTime()); feed.setLastUpdated(Calendar.getInstance().getTime());
try { try {
SyndFeed rss = new SyndFeedInput().build(new InputSource( InputSource source = new InputSource(new ByteArrayInputStream(xml));
new ByteArrayInputStream(xml))); if (new String(xml).split(SystemUtils.LINE_SEPARATOR)[0]
.toUpperCase().contains("ISO-8859-1")) {
// they probably use word, we need to handle curly quotes and
// other word special characters
source.setEncoding("windows-1252");
}
SyndFeed rss = new SyndFeedInput().build(source);
feed.setUrl(feedUrl); feed.setUrl(feedUrl);
feed.setTitle(rss.getTitle()); feed.setTitle(rss.getTitle());
feed.setLink(rss.getLink()); feed.setLink(rss.getLink());
@@ -38,7 +46,7 @@ public class FeedParser {
FeedEntry entry = new FeedEntry(); FeedEntry entry = new FeedEntry();
entry.setGuid(item.getUri()); entry.setGuid(item.getUri());
entry.setTitle(handleContent(item.getTitle())); entry.setTitle(handleContent(item.getTitle()));
entry.setContent(getContent(item)); entry.setContent(handleContent(getContent(item)));
entry.setUrl(item.getLink()); entry.setUrl(item.getLink());
entry.setUpdated(item.getUpdatedDate() != null ? item entry.setUpdated(item.getUpdatedDate() != null ? item
.getUpdatedDate() : item.getPublishedDate()); .getUpdatedDate() : item.getPublishedDate());
@@ -67,7 +75,6 @@ public class FeedParser {
} }
}), SystemUtils.LINE_SEPARATOR); }), SystemUtils.LINE_SEPARATOR);
} }
content = handleContent(content);
return content; return content;
} }
@@ -81,7 +88,8 @@ public class FeedParser {
whitelist.addAttributes("iframe", "src", "height", "width", whitelist.addAttributes("iframe", "src", "height", "width",
"allowfullscreen", "frameborder"); "allowfullscreen", "frameborder");
content = Jsoup.clean(content, whitelist); content = Jsoup.clean(content, "", whitelist,
new OutputSettings().escapeMode(EscapeMode.extended));
} }
return content; return content;
} }