You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@poi.apache.org by fa...@apache.org on 2019/12/26 06:12:56 UTC
svn commit: r1871986 - in
/poi/trunk/src/scratchpad/src/org/apache/poi/hsmf/extractor:
OutlookTextExtactor.java OutlookTextExtractor.java
Author: fanningpj
Date: Thu Dec 26 06:12:56 2019
New Revision: 1871986
URL: http://svn.apache.org/viewvc?rev=1871986&view=rev
Log:
fix spelling of OutlookTextExtractor class name
Added:
poi/trunk/src/scratchpad/src/org/apache/poi/hsmf/extractor/OutlookTextExtractor.java
- copied, changed from r1871985, poi/trunk/src/scratchpad/src/org/apache/poi/hsmf/extractor/OutlookTextExtactor.java
Modified:
poi/trunk/src/scratchpad/src/org/apache/poi/hsmf/extractor/OutlookTextExtactor.java
Modified: poi/trunk/src/scratchpad/src/org/apache/poi/hsmf/extractor/OutlookTextExtactor.java
URL: http://svn.apache.org/viewvc/poi/trunk/src/scratchpad/src/org/apache/poi/hsmf/extractor/OutlookTextExtactor.java?rev=1871986&r1=1871985&r2=1871986&view=diff
==============================================================================
--- poi/trunk/src/scratchpad/src/org/apache/poi/hsmf/extractor/OutlookTextExtactor.java (original)
+++ poi/trunk/src/scratchpad/src/org/apache/poi/hsmf/extractor/OutlookTextExtactor.java Thu Dec 26 06:12:56 2019
@@ -16,174 +16,46 @@
==================================================================== */
package org.apache.poi.hsmf.extractor;
-import static org.apache.poi.util.StringUtil.startsWithIgnoreCase;
+import org.apache.poi.hsmf.MAPIMessage;
+import org.apache.poi.poifs.filesystem.DirectoryNode;
+import org.apache.poi.poifs.filesystem.POIFSFileSystem;
+import org.apache.poi.util.Removal;
import java.io.File;
import java.io.IOException;
import java.io.InputStream;
-import java.text.SimpleDateFormat;
-import java.util.Locale;
-
-import org.apache.poi.extractor.POIOLE2TextExtractor;
-import org.apache.poi.hsmf.MAPIMessage;
-import org.apache.poi.hsmf.datatypes.AttachmentChunks;
-import org.apache.poi.hsmf.datatypes.StringChunk;
-import org.apache.poi.hsmf.exceptions.ChunkNotFoundException;
-import org.apache.poi.poifs.filesystem.DirectoryNode;
-import org.apache.poi.poifs.filesystem.POIFSFileSystem;
-import org.apache.poi.util.LocaleUtil;
-import org.apache.poi.util.StringUtil.StringsIterator;
/**
* A text extractor for HSMF (Outlook) .msg files.
* Outputs in a format somewhat like a plain text email.
+ *
+ * @deprecated use @{link OutlookTextExtractor} instead
*/
-public class OutlookTextExtactor extends POIOLE2TextExtractor {
- public OutlookTextExtactor(MAPIMessage msg) {
- super(msg);
- }
- public OutlookTextExtactor(DirectoryNode poifsDir) throws IOException {
- this(new MAPIMessage(poifsDir));
- }
- public OutlookTextExtactor(POIFSFileSystem fs) throws IOException {
- this(new MAPIMessage(fs));
- }
- public OutlookTextExtactor(InputStream inp) throws IOException {
- this(new MAPIMessage(inp));
- }
-
- public static void main(String[] args) throws Exception {
- for(String filename : args) {
- try (POIFSFileSystem poifs = new POIFSFileSystem(new File(filename));
- OutlookTextExtactor extractor = new OutlookTextExtactor(poifs)) {
- System.out.println(extractor.getText());
- }
- }
- }
-
- /**
- * Returns the underlying MAPI message
- */
- public MAPIMessage getMAPIMessage() {
- return (MAPIMessage)document;
- }
-
- /**
- * Outputs something a little like a RFC822 email
- */
- public String getText() {
- MAPIMessage msg = (MAPIMessage)document;
- StringBuilder s = new StringBuilder();
-
- // See if we can get a suitable encoding for any
- // non unicode text in the file
- msg.guess7BitEncoding();
-
- // Off we go
- StringsIterator emails;
- try {
- emails = new StringsIterator(
- msg.getRecipientEmailAddressList()
- );
- } catch(ChunkNotFoundException e) {
- emails = new StringsIterator(new String[0]);
- }
-
- try {
- s.append("From: " + msg.getDisplayFrom() + "\n");
- } catch(ChunkNotFoundException e) {}
-
- // For To, CC and BCC, try to match the names
- // up with their email addresses. Relies on the
- // Recipient Chunks being in the same order as
- // people in To + CC + BCC.
- try {
- handleEmails(s, "To", msg.getDisplayTo(), emails);
- } catch(ChunkNotFoundException e) {}
- try {
- handleEmails(s, "CC", msg.getDisplayCC(), emails);
- } catch(ChunkNotFoundException e) {}
- try {
- handleEmails(s, "BCC", msg.getDisplayBCC(), emails);
- } catch(ChunkNotFoundException e) {}
-
- // Date - try two ways to find it
- try {
- // First try via the proper chunk
- SimpleDateFormat f = new SimpleDateFormat("E, d MMM yyyy HH:mm:ss Z", Locale.ROOT);
- f.setTimeZone(LocaleUtil.getUserTimeZone());
- s.append("Date: ").append(f.format(msg.getMessageDate().getTime())).append("\n");
- } catch(ChunkNotFoundException e) {
- try {
- // Failing that try via the raw headers
- String[] headers = msg.getHeaders();
- for(String header: headers) {
- if(startsWithIgnoreCase(header, "date:")) {
- s.append("Date:").append(header, header.indexOf(':')+1, header.length()).append("\n");
- break;
- }
- }
- } catch(ChunkNotFoundException he) {
- // We can't find the date, sorry...
- }
- }
-
- try {
- s.append("Subject: ").append(msg.getSubject()).append("\n");
- } catch(ChunkNotFoundException e) {}
-
- // Display attachment names
- // To get the attachments, use ExtractorFactory
- for(AttachmentChunks att : msg.getAttachmentFiles()) {
- StringChunk name = att.getAttachLongFileName();
- if (name == null) name = att.getAttachFileName();
- String attName = name == null ? null : name.getValue();
-
- if(att.getAttachMimeTag() != null &&
- att.getAttachMimeTag().getValue() != null) {
- attName = att.getAttachMimeTag().getValue() + " = " + attName;
- }
- s.append("Attachment: ").append(attName).append("\n");
- }
-
- try {
- s.append("\n").append(msg.getTextBody()).append("\n");
- } catch(ChunkNotFoundException e) {}
-
- return s.toString();
- }
-
- /**
- * Takes a Display focused string, eg "Nick; Jim" and an iterator
- * of emails, and does its best to return something like
- * "Nick <ni...@example.com>; Jim <ji...@example.com>"
- */
- protected void handleEmails(StringBuilder s, String type, String displayText, StringsIterator emails) {
- if(displayText == null || displayText.length() == 0) {
- return;
- }
-
- String[] names = displayText.split(";\\s*");
- boolean first = true;
-
- s.append(type).append(": ");
- for(String name : names) {
- if(first) {
- first = false;
- } else {
- s.append("; ");
- }
-
- s.append(name);
- if(emails.hasNext()) {
- String email = emails.next();
- // Append the email address in <>, assuming
- // the name wasn't already the email address
- if(! email.equals(name)) {
- s.append(" <").append(email).append(">");
+@Deprecated
+@Removal(version = "5.0.0")
+public class OutlookTextExtactor extends OutlookTextExtractor {
+ public OutlookTextExtactor(MAPIMessage msg) {
+ super(msg);
+ }
+
+ public OutlookTextExtactor(DirectoryNode poifsDir) throws IOException {
+ super(new MAPIMessage(poifsDir));
+ }
+
+ public OutlookTextExtactor(POIFSFileSystem fs) throws IOException {
+ super(new MAPIMessage(fs));
+ }
+
+ public OutlookTextExtactor(InputStream inp) throws IOException {
+ super(new MAPIMessage(inp));
+ }
+
+ public static void main(String[] args) throws Exception {
+ for (String filename : args) {
+ try (POIFSFileSystem poifs = new POIFSFileSystem(new File(filename));
+ OutlookTextExtractor extractor = new OutlookTextExtractor(poifs)) {
+ System.out.println(extractor.getText());
}
- }
- }
- s.append("\n");
- }
+ }
+ }
}
Copied: poi/trunk/src/scratchpad/src/org/apache/poi/hsmf/extractor/OutlookTextExtractor.java (from r1871985, poi/trunk/src/scratchpad/src/org/apache/poi/hsmf/extractor/OutlookTextExtactor.java)
URL: http://svn.apache.org/viewvc/poi/trunk/src/scratchpad/src/org/apache/poi/hsmf/extractor/OutlookTextExtractor.java?p2=poi/trunk/src/scratchpad/src/org/apache/poi/hsmf/extractor/OutlookTextExtractor.java&p1=poi/trunk/src/scratchpad/src/org/apache/poi/hsmf/extractor/OutlookTextExtactor.java&r1=1871985&r2=1871986&rev=1871986&view=diff
==============================================================================
--- poi/trunk/src/scratchpad/src/org/apache/poi/hsmf/extractor/OutlookTextExtactor.java (original)
+++ poi/trunk/src/scratchpad/src/org/apache/poi/hsmf/extractor/OutlookTextExtractor.java Thu Dec 26 06:12:56 2019
@@ -16,14 +16,6 @@
==================================================================== */
package org.apache.poi.hsmf.extractor;
-import static org.apache.poi.util.StringUtil.startsWithIgnoreCase;
-
-import java.io.File;
-import java.io.IOException;
-import java.io.InputStream;
-import java.text.SimpleDateFormat;
-import java.util.Locale;
-
import org.apache.poi.extractor.POIOLE2TextExtractor;
import org.apache.poi.hsmf.MAPIMessage;
import org.apache.poi.hsmf.datatypes.AttachmentChunks;
@@ -32,158 +24,178 @@ import org.apache.poi.hsmf.exceptions.Ch
import org.apache.poi.poifs.filesystem.DirectoryNode;
import org.apache.poi.poifs.filesystem.POIFSFileSystem;
import org.apache.poi.util.LocaleUtil;
+import org.apache.poi.util.Removal;
import org.apache.poi.util.StringUtil.StringsIterator;
+import java.io.File;
+import java.io.IOException;
+import java.io.InputStream;
+import java.text.SimpleDateFormat;
+import java.util.Locale;
+
+import static org.apache.poi.util.StringUtil.startsWithIgnoreCase;
+
/**
* A text extractor for HSMF (Outlook) .msg files.
* Outputs in a format somewhat like a plain text email.
+ *
+ * @since 4.1.2
*/
-public class OutlookTextExtactor extends POIOLE2TextExtractor {
- public OutlookTextExtactor(MAPIMessage msg) {
- super(msg);
- }
- public OutlookTextExtactor(DirectoryNode poifsDir) throws IOException {
- this(new MAPIMessage(poifsDir));
- }
- public OutlookTextExtactor(POIFSFileSystem fs) throws IOException {
- this(new MAPIMessage(fs));
- }
- public OutlookTextExtactor(InputStream inp) throws IOException {
- this(new MAPIMessage(inp));
- }
-
- public static void main(String[] args) throws Exception {
- for(String filename : args) {
- try (POIFSFileSystem poifs = new POIFSFileSystem(new File(filename));
- OutlookTextExtactor extractor = new OutlookTextExtactor(poifs)) {
- System.out.println(extractor.getText());
- }
- }
- }
-
- /**
- * Returns the underlying MAPI message
- */
- public MAPIMessage getMAPIMessage() {
- return (MAPIMessage)document;
- }
-
- /**
- * Outputs something a little like a RFC822 email
- */
- public String getText() {
- MAPIMessage msg = (MAPIMessage)document;
- StringBuilder s = new StringBuilder();
-
- // See if we can get a suitable encoding for any
- // non unicode text in the file
- msg.guess7BitEncoding();
-
- // Off we go
- StringsIterator emails;
- try {
- emails = new StringsIterator(
- msg.getRecipientEmailAddressList()
- );
- } catch(ChunkNotFoundException e) {
- emails = new StringsIterator(new String[0]);
- }
-
- try {
- s.append("From: " + msg.getDisplayFrom() + "\n");
- } catch(ChunkNotFoundException e) {}
-
- // For To, CC and BCC, try to match the names
- // up with their email addresses. Relies on the
- // Recipient Chunks being in the same order as
- // people in To + CC + BCC.
- try {
- handleEmails(s, "To", msg.getDisplayTo(), emails);
- } catch(ChunkNotFoundException e) {}
- try {
- handleEmails(s, "CC", msg.getDisplayCC(), emails);
- } catch(ChunkNotFoundException e) {}
- try {
- handleEmails(s, "BCC", msg.getDisplayBCC(), emails);
- } catch(ChunkNotFoundException e) {}
-
- // Date - try two ways to find it
- try {
- // First try via the proper chunk
- SimpleDateFormat f = new SimpleDateFormat("E, d MMM yyyy HH:mm:ss Z", Locale.ROOT);
- f.setTimeZone(LocaleUtil.getUserTimeZone());
- s.append("Date: ").append(f.format(msg.getMessageDate().getTime())).append("\n");
- } catch(ChunkNotFoundException e) {
- try {
- // Failing that try via the raw headers
- String[] headers = msg.getHeaders();
- for(String header: headers) {
- if(startsWithIgnoreCase(header, "date:")) {
- s.append("Date:").append(header, header.indexOf(':')+1, header.length()).append("\n");
- break;
- }
+public class OutlookTextExtractor extends POIOLE2TextExtractor {
+ public OutlookTextExtractor(MAPIMessage msg) {
+ super(msg);
+ }
+
+ public OutlookTextExtractor(DirectoryNode poifsDir) throws IOException {
+ this(new MAPIMessage(poifsDir));
+ }
+
+ public OutlookTextExtractor(POIFSFileSystem fs) throws IOException {
+ this(new MAPIMessage(fs));
+ }
+
+ public OutlookTextExtractor(InputStream inp) throws IOException {
+ this(new MAPIMessage(inp));
+ }
+
+ public static void main(String[] args) throws Exception {
+ for (String filename : args) {
+ try (POIFSFileSystem poifs = new POIFSFileSystem(new File(filename));
+ OutlookTextExtractor extractor = new OutlookTextExtractor(poifs)) {
+ System.out.println(extractor.getText());
}
- } catch(ChunkNotFoundException he) {
- // We can't find the date, sorry...
- }
- }
-
- try {
- s.append("Subject: ").append(msg.getSubject()).append("\n");
- } catch(ChunkNotFoundException e) {}
-
- // Display attachment names
- // To get the attachments, use ExtractorFactory
- for(AttachmentChunks att : msg.getAttachmentFiles()) {
- StringChunk name = att.getAttachLongFileName();
- if (name == null) name = att.getAttachFileName();
- String attName = name == null ? null : name.getValue();
-
- if(att.getAttachMimeTag() != null &&
- att.getAttachMimeTag().getValue() != null) {
- attName = att.getAttachMimeTag().getValue() + " = " + attName;
- }
- s.append("Attachment: ").append(attName).append("\n");
- }
-
- try {
- s.append("\n").append(msg.getTextBody()).append("\n");
- } catch(ChunkNotFoundException e) {}
-
- return s.toString();
- }
-
- /**
- * Takes a Display focused string, eg "Nick; Jim" and an iterator
- * of emails, and does its best to return something like
- * "Nick <ni...@example.com>; Jim <ji...@example.com>"
- */
- protected void handleEmails(StringBuilder s, String type, String displayText, StringsIterator emails) {
- if(displayText == null || displayText.length() == 0) {
- return;
- }
-
- String[] names = displayText.split(";\\s*");
- boolean first = true;
-
- s.append(type).append(": ");
- for(String name : names) {
- if(first) {
- first = false;
- } else {
- s.append("; ");
- }
-
- s.append(name);
- if(emails.hasNext()) {
- String email = emails.next();
- // Append the email address in <>, assuming
- // the name wasn't already the email address
- if(! email.equals(name)) {
- s.append(" <").append(email).append(">");
+ }
+ }
+
+ /**
+ * Returns the underlying MAPI message
+ */
+ public MAPIMessage getMAPIMessage() {
+ return (MAPIMessage) document;
+ }
+
+ /**
+ * Outputs something a little like a RFC822 email
+ */
+ public String getText() {
+ MAPIMessage msg = (MAPIMessage) document;
+ StringBuilder s = new StringBuilder();
+
+ // See if we can get a suitable encoding for any
+ // non unicode text in the file
+ msg.guess7BitEncoding();
+
+ // Off we go
+ StringsIterator emails;
+ try {
+ emails = new StringsIterator(
+ msg.getRecipientEmailAddressList()
+ );
+ } catch (ChunkNotFoundException e) {
+ emails = new StringsIterator(new String[0]);
+ }
+
+ try {
+ s.append("From: " + msg.getDisplayFrom() + "\n");
+ } catch (ChunkNotFoundException e) {
+ }
+
+ // For To, CC and BCC, try to match the names
+ // up with their email addresses. Relies on the
+ // Recipient Chunks being in the same order as
+ // people in To + CC + BCC.
+ try {
+ handleEmails(s, "To", msg.getDisplayTo(), emails);
+ } catch (ChunkNotFoundException e) {
+ }
+ try {
+ handleEmails(s, "CC", msg.getDisplayCC(), emails);
+ } catch (ChunkNotFoundException e) {
+ }
+ try {
+ handleEmails(s, "BCC", msg.getDisplayBCC(), emails);
+ } catch (ChunkNotFoundException e) {
+ }
+
+ // Date - try two ways to find it
+ try {
+ // First try via the proper chunk
+ SimpleDateFormat f = new SimpleDateFormat("E, d MMM yyyy HH:mm:ss Z", Locale.ROOT);
+ f.setTimeZone(LocaleUtil.getUserTimeZone());
+ s.append("Date: ").append(f.format(msg.getMessageDate().getTime())).append("\n");
+ } catch (ChunkNotFoundException e) {
+ try {
+ // Failing that try via the raw headers
+ String[] headers = msg.getHeaders();
+ for (String header : headers) {
+ if (startsWithIgnoreCase(header, "date:")) {
+ s.append("Date:").append(header, header.indexOf(':') + 1, header.length()).append("\n");
+ break;
+ }
+ }
+ } catch (ChunkNotFoundException he) {
+ // We can't find the date, sorry...
+ }
+ }
+
+ try {
+ s.append("Subject: ").append(msg.getSubject()).append("\n");
+ } catch (ChunkNotFoundException e) {
+ }
+
+ // Display attachment names
+ // To get the attachments, use ExtractorFactory
+ for (AttachmentChunks att : msg.getAttachmentFiles()) {
+ StringChunk name = att.getAttachLongFileName();
+ if (name == null) name = att.getAttachFileName();
+ String attName = name == null ? null : name.getValue();
+
+ if (att.getAttachMimeTag() != null &&
+ att.getAttachMimeTag().getValue() != null) {
+ attName = att.getAttachMimeTag().getValue() + " = " + attName;
+ }
+ s.append("Attachment: ").append(attName).append("\n");
+ }
+
+ try {
+ s.append("\n").append(msg.getTextBody()).append("\n");
+ } catch (ChunkNotFoundException e) {
+ }
+
+ return s.toString();
+ }
+
+ /**
+ * Takes a Display focused string, eg "Nick; Jim" and an iterator
+ * of emails, and does its best to return something like
+ * "Nick <ni...@example.com>; Jim <ji...@example.com>"
+ */
+ protected void handleEmails(StringBuilder s, String type, String displayText, StringsIterator emails) {
+ if (displayText == null || displayText.length() == 0) {
+ return;
+ }
+
+ String[] names = displayText.split(";\\s*");
+ boolean first = true;
+
+ s.append(type).append(": ");
+ for (String name : names) {
+ if (first) {
+ first = false;
+ } else {
+ s.append("; ");
+ }
+
+ s.append(name);
+ if (emails.hasNext()) {
+ String email = emails.next();
+ // Append the email address in <>, assuming
+ // the name wasn't already the email address
+ if (!email.equals(name)) {
+ s.append(" <").append(email).append(">");
+ }
}
- }
- }
- s.append("\n");
- }
+ }
+ s.append("\n");
+ }
}
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@poi.apache.org
For additional commands, e-mail: commits-help@poi.apache.org