boxtec
/
yacy_search_server
mirror of https://github.com/yacy/yacy_search_server
4
1
Fork 0
Code Issues Projects Releases Wiki Activity

fix for parser/encoding Exception

Browse Source 
see http://forum.yacy-websuche.de/viewtopic.php?f=5&t=1111&hilit=&sid=55a320b54e1e3bda9410e7c50b5147f1&p=7431#p7431

git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@4741 6c8d7289-2bf4-0310-a012-ef5d649a1542
pull/1/head
orbiter 17 years ago
parent cf042e6957
commit 685794e7e7
1 changed files with 13 additions and 13 deletions
Show Stats Download Patch File Download Diff File

26
source/de/anomic/plasma/plasmaParser.java
Unescape View File

@ -338,21 +338,21 @@ public final class plasmaParser {
if ((encoding == null) || (encoding.length() == 0)) return "ISO-8859-1"; if ((encoding == null) || (encoding.length() == 0)) return "ISO-8859-1";
// trim encoding string // trim encoding string
encoding = encoding.trim(); encoding = encoding.toUpperCase().trim();
if (encoding.toLowerCase().startsWith("windows") && encoding.length() > 7) { if (encoding.toLowerCase().startsWith("WINDOWS") && encoding.length() > 7) {
char c = encoding.charAt(7); char c = encoding.charAt(7);
if (c == '_') encoding = "windows-" + encoding.substring(8); if (c == '_') encoding = "WINDOWS-" + encoding.substring(8);
else if ((c >= '0') && (c <= '9')) encoding = "windows-" + encoding.substring(7); else if ((c >= '0') && (c <= '9')) encoding = "WINDOWS-" + encoding.substring(7);
} }
if (encoding.toLowerCase().startsWith("iso") && encoding.length() > 3) { if (encoding.toLowerCase().startsWith("ISO") && encoding.length() > 3) {
char c = encoding.charAt(3); char c = encoding.charAt(3);
if (c == '_') encoding = "ISO-" + encoding.substring(4); if (c == '_') encoding = "ISO-" + encoding.substring(4);
else if ((c >= '0') && (c <= '9')) encoding = "ISO-" + encoding.substring(3); else if ((c >= '0') && (c <= '9')) encoding = "ISO-" + encoding.substring(3);
} }
if (encoding.toLowerCase().startsWith("iso") && encoding.length() > 8) { if (encoding.toLowerCase().startsWith("ISO") && encoding.length() > 8) {
char c = encoding.charAt(8); char c = encoding.charAt(8);
if (c == '_') encoding = encoding.substring(0,8) + "-" + encoding.substring(9); if (c == '_') encoding = encoding.substring(0,8) + "-" + encoding.substring(9);
else if ((c >= '0') && (c <= '9')) encoding = encoding.substring(0,8) + "-" + encoding.substring(8); else if ((c >= '0') && (c <= '9')) encoding = encoding.substring(0,8) + "-" + encoding.substring(8);
@ -360,17 +360,17 @@ public final class plasmaParser {
// converting cp\d{4} -> windows-\d{4} // converting cp\d{4} -> windows-\d{4}
if (encoding.toLowerCase().matches("cp([_-])?125[0-8]")) { if (encoding.toLowerCase().matches("CP([_-])?125[0-8]")) {
char c = encoding.charAt(2); char c = encoding.charAt(2);
if (c == '_' || c == '-') encoding = "windows-" + encoding.substring(3); if (c == '_' || c == '-') encoding = "WINDOWS-" + encoding.substring(3);
else if ((c >= '0') && (c <= '9')) encoding = "windows-" + encoding.substring(2); else if ((c >= '0') && (c <= '9')) encoding = "WINDOWS-" + encoding.substring(2);
} }
if (encoding.toLowerCase().matches("gb[_-]?2312([-_]80)?")) { if (encoding.toLowerCase().matches("GB[_-]?2312([-_]80)?")) {
encoding = "x-EUC-CN"; encoding = "X-EUC-CN";
} }
if (encoding.toLowerCase().matches(".*utf[-_]?8.*")) { if (encoding.toLowerCase().matches(".*UTF[-_]?8.*")) {
encoding = "UTF-8"; encoding = "UTF-8";
} }
@ -636,7 +636,7 @@ public final class plasmaParser {
} catch (UnsupportedEncodingException e) { } catch (UnsupportedEncodingException e) {
String errorMsg = "Unsupported charset encoding: " + e.getMessage(); String errorMsg = "Unsupported charset encoding: " + e.getMessage();
this.theLogger.logSevere("Unable to parse '" + location + "'. " + errorMsg); this.theLogger.logSevere("Unable to parse '" + location + "'. " + errorMsg, e);
throw new ParserException(errorMsg,location,plasmaCrawlEURL.DENIED_UNSUPPORTED_CHARSET); throw new ParserException(errorMsg,location,plasmaCrawlEURL.DENIED_UNSUPPORTED_CHARSET);
} catch (Exception e) { } catch (Exception e) {
// Interrupted- and Parser-Exceptions should pass through // Interrupted- and Parser-Exceptions should pass through

Write Preview
Loading…
Cancel
Save