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

trim unicode characters from url if any

Browse Source
This commit is contained in:
Athou
2013-04-22 13:20:17 +02:00
parent c2e56ccce0
commit faaf18f5cf
3 changed files with 42 additions and 33 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
src/main/java/com/commafeed/backend/feeds/FeedRefreshWorker.java
View File

@@ -123,7 +123,8 @@ public class FeedRefreshWorker {
feed.setDisabledUntil(disabledUntil);
if (fetchedFeed != null) {
feed.setLink(fetchedFeed.getLink());
feed.setLink(FeedUtils.trimUnicodeSurrogateCharacters(fetchedFeed
.getLink()));
feed.setLastModifiedHeader(fetchedFeed.getLastModifiedHeader());
feed.setEtagHeader(fetchedFeed.getEtagHeader());
feedUpdateService.updateEntries(feed, fetchedFeed.getEntries());

37
src/main/java/com/commafeed/backend/feeds/FeedUtils.java Normal file
View File

@@ -0,0 +1,37 @@
package com.commafeed.backend.feeds;
import org.apache.commons.lang.StringUtils;
import org.jsoup.Jsoup;
import org.jsoup.nodes.Document.OutputSettings;
import org.jsoup.nodes.Entities.EscapeMode;
import org.jsoup.safety.Whitelist;
public class FeedUtils {
public static String handleContent(String content) {
if (StringUtils.isNotBlank(content)) {
content = trimUnicodeSurrogateCharacters(content);
Whitelist whitelist = Whitelist.relaxed();
whitelist.addEnforcedAttribute("a", "target", "_blank");
whitelist.addTags("iframe");
whitelist.addAttributes("iframe", "src", "height", "width",
"allowfullscreen", "frameborder");
content = Jsoup.clean(content, "", whitelist,
new OutputSettings().escapeMode(EscapeMode.base));
}
return content;
}
public static String trimUnicodeSurrogateCharacters(String text) {
StringBuilder sb = new StringBuilder();
for (int i = 0; i < text.length(); i++) {
char ch = text.charAt(i);
if (!Character.isHighSurrogate(ch) && !Character.isLowSurrogate(ch)) {
sb.append(ch);
}
}
return sb.toString();
}
}

35
src/main/java/com/commafeed/backend/services/FeedUpdateService.java
View File

@@ -10,15 +10,12 @@ import javax.inject.Inject;
import org.apache.commons.lang.ObjectUtils;
import org.apache.commons.lang.StringUtils;
import org.jsoup.Jsoup;
import org.jsoup.nodes.Document.OutputSettings;
import org.jsoup.nodes.Entities.EscapeMode;
import org.jsoup.safety.Whitelist;
import com.commafeed.backend.dao.FeedDAO;
import com.commafeed.backend.dao.FeedEntryDAO;
import com.commafeed.backend.dao.FeedEntryStatusDAO;
import com.commafeed.backend.dao.FeedSubscriptionDAO;
import com.commafeed.backend.feeds.FeedUtils;
import com.commafeed.backend.model.Feed;
import com.commafeed.backend.model.FeedEntry;
import com.commafeed.backend.model.FeedEntryContent;
@@ -60,9 +57,9 @@ public class FeedUpdateService {
}
if (foundEntry == null) {
FeedEntryContent content = entry.getContent();
content.setContent(handleContent(content.getContent()));
content.setContent(FeedUtils.handleContent(content.getContent()));
String title = handleContent(content.getTitle());
String title = FeedUtils.handleContent(content.getTitle());
if (title != null) {
content.setTitle(title.substring(0,
Math.min(2048, title.length())));
@@ -100,30 +97,4 @@ public class FeedUpdateService {
}
private String handleContent(String content) {
if (StringUtils.isNotBlank(content)) {
content = trimUnicodeSurrogateCharacters(content);
Whitelist whitelist = Whitelist.relaxed();
whitelist.addEnforcedAttribute("a", "target", "_blank");
whitelist.addTags("iframe");
whitelist.addAttributes("iframe", "src", "height", "width",
"allowfullscreen", "frameborder");
content = Jsoup.clean(content, "", whitelist,
new OutputSettings().escapeMode(EscapeMode.base));
}
return content;
}
private String trimUnicodeSurrogateCharacters(String text) {
StringBuilder sb = new StringBuilder();
for (int i = 0; i < text.length(); i++) {
char ch = text.charAt(i);
if (!Character.isHighSurrogate(ch) && !Character.isLowSurrogate(ch)) {
sb.append(ch);
}
}
return sb.toString();
}
}