diff options
author | jmisinco | 2013-12-17 15:37:47 +0000 |
---|---|---|
committer | Gerrit Code Review @ Eclipse.org | 2013-12-18 20:33:37 +0000 |
commit | e3200c768c2983d2e36882ce8cb1ee4b41c6ecd6 (patch) | |
tree | 354f907b3e924be306ade30c8adf0e47bce3c82c /plugins | |
parent | 59a104e254f2200dc7fa951ca8b105961cc246c7 (diff) | |
download | org.eclipse.osee-e3200c768c2983d2e36882ce8cb1ee4b41c6ecd6.tar.gz org.eclipse.osee-e3200c768c2983d2e36882ce8cb1ee4b41c6ecd6.tar.xz org.eclipse.osee-e3200c768c2983d2e36882ce8cb1ee4b41c6ecd6.zip |
bug: Fixes to search tagger
Change-Id: Ib13d00935e2b9c2bcf5cc4d75be86b83d5db529a
Diffstat (limited to 'plugins')
6 files changed, 23 insertions, 19 deletions
diff --git a/plugins/org.eclipse.osee.framework.jdk.core/src/org/eclipse/osee/framework/jdk/core/util/io/xml/XmlTextInputStream.java b/plugins/org.eclipse.osee.framework.jdk.core/src/org/eclipse/osee/framework/jdk/core/util/io/xml/XmlTextInputStream.java index 743608afb41..6273a9bbcd4 100644 --- a/plugins/org.eclipse.osee.framework.jdk.core/src/org/eclipse/osee/framework/jdk/core/util/io/xml/XmlTextInputStream.java +++ b/plugins/org.eclipse.osee.framework.jdk.core/src/org/eclipse/osee/framework/jdk/core/util/io/xml/XmlTextInputStream.java @@ -22,12 +22,12 @@ import org.eclipse.osee.framework.jdk.core.util.ReservedCharacters; * @author Roberto E. Escobar */ public class XmlTextInputStream extends BufferedInputStream { - private static final String START_PARAGRAPH = "<w:p"; - private static final String STOP_PARAGRAPH = "</w:p"; - private static final String START_WORDML_TEXT = "<w:t>"; + private static final String START_PARAGRAPH_REGEX = "<w:p( .+)?>"; + private static final String STOP_PARAGRAPH = "</w:p>"; + private static final String START_WORDML_TEXT_REGEX = "<w:t( .+)?>"; private static final String END_WORDML_TEXT = "</w:t>"; private static final String LINE_BREAK = "<w:br/>"; - private static final String TAB = "<w:tab/>"; + private static final String TAB_REGEX = "<w:tab( .+)?/>"; private IReadHelper readHelper; @@ -236,17 +236,17 @@ public class XmlTextInputStream extends BufferedInputStream { if ((char) value == '>') { partOfTag = false; String tag = buffer.toString(); - if (tag.equals(START_WORDML_TEXT)) { + if (tag.matches(START_WORDML_TEXT_REGEX)) { collect = true; } else if (tag.equals(END_WORDML_TEXT)) { collect = false; - } else if (tag.startsWith(START_PARAGRAPH)) { + } else if (tag.matches(START_PARAGRAPH_REGEX)) { isStartOfParagraph = true; } else if (tag.startsWith(STOP_PARAGRAPH)) { isStartOfParagraph = false; } else if (tag.startsWith(LINE_BREAK)) { isBreak = true; - } else if (tag.startsWith(TAB)) { + } else if (tag.matches(TAB_REGEX)) { isBreak = true; } diff --git a/plugins/org.eclipse.osee.orcs.db.test/src/org/eclipse/osee/orcs/db/internal/search/tagger/data/test8.data.xml b/plugins/org.eclipse.osee.orcs.db.test/src/org/eclipse/osee/orcs/db/internal/search/tagger/data/test8.data.xml index eca2b5629df..6f9dc24b028 100644 --- a/plugins/org.eclipse.osee.orcs.db.test/src/org/eclipse/osee/orcs/db/internal/search/tagger/data/test8.data.xml +++ b/plugins/org.eclipse.osee.orcs.db.test/src/org/eclipse/osee/orcs/db/internal/search/tagger/data/test8.data.xml @@ -2,4 +2,6 @@ <w:r><w:rPr><w:sz-cs w:val="18"/></w:rPr><w:t>5 = ARC-210</w:t></w:r> <w:r><w:rPr><w:sz-cs w:val="18"/></w:rPr><w:br/></w:r> <w:r><w:rPr><w:sz w:val="18"/><w:sz-cs w:val="18"/></w:rPr><w:t>6 = RESERVED</w:t></w:r><w:r><w:tab/><w:t>yes it is.</w:t></w:r> +<w:p><w:r w:rsidR="00CE6B16"><w:rPr><w:rFonts w:ascii="Helvetica" w:hAnsi="Helvetica" w:cs="Helvetica"/></w:rPr><w:t xml:space="preserve"> The quick brown fox.</w:t></w:r> +<w:tab wx:wTab="2880" wx:tlc="none" wx:cTlc="14"/><w:t>Hello</w:t></w:p> </w:p> diff --git a/plugins/org.eclipse.osee.orcs.db.test/src/org/eclipse/osee/orcs/db/internal/search/tagger/data/test8.expected.txt b/plugins/org.eclipse.osee.orcs.db.test/src/org/eclipse/osee/orcs/db/internal/search/tagger/data/test8.expected.txt index 70e6343a0f7..bd7768b0ef4 100644 --- a/plugins/org.eclipse.osee.orcs.db.test/src/org/eclipse/osee/orcs/db/internal/search/tagger/data/test8.expected.txt +++ b/plugins/org.eclipse.osee.orcs.db.test/src/org/eclipse/osee/orcs/db/internal/search/tagger/data/test8.expected.txt @@ -1 +1 @@ -5 = ARC-210 6 = RESERVED yes it is.
\ No newline at end of file +5 = ARC-210 6 = RESERVED yes it is. The quick brown fox. Hello
\ No newline at end of file diff --git a/plugins/org.eclipse.osee.orcs.db.test/src/org/eclipse/osee/orcs/db/internal/search/tagger/data/test8.tags.txt b/plugins/org.eclipse.osee.orcs.db.test/src/org/eclipse/osee/orcs/db/internal/search/tagger/data/test8.tags.txt index ef15220213b..7588c3c7f86 100644 --- a/plugins/org.eclipse.osee.orcs.db.test/src/org/eclipse/osee/orcs/db/internal/search/tagger/data/test8.tags.txt +++ b/plugins/org.eclipse.osee.orcs.db.test/src/org/eclipse/osee/orcs/db/internal/search/tagger/data/test8.tags.txt @@ -6,3 +6,8 @@ reserved -539428360 yes 6654 it 434 is 402 +the 3866 +quick 4182975 +brown 1431435 +fox 16239 +hello 1520625
\ No newline at end of file diff --git a/plugins/org.eclipse.osee.orcs.db/src/org/eclipse/osee/orcs/db/internal/search/indexer/data/IndexerDataSourceImpl.java b/plugins/org.eclipse.osee.orcs.db/src/org/eclipse/osee/orcs/db/internal/search/indexer/data/IndexerDataSourceImpl.java index 201c24d1cef..7c5359f33e8 100644 --- a/plugins/org.eclipse.osee.orcs.db/src/org/eclipse/osee/orcs/db/internal/search/indexer/data/IndexerDataSourceImpl.java +++ b/plugins/org.eclipse.osee.orcs.db/src/org/eclipse/osee/orcs/db/internal/search/indexer/data/IndexerDataSourceImpl.java @@ -16,7 +16,6 @@ import java.io.InputStream; import java.net.URI; import org.eclipse.osee.framework.jdk.core.type.OseeCoreException; import org.eclipse.osee.framework.jdk.core.type.PropertyStore; -import org.eclipse.osee.framework.jdk.core.util.Lib; import org.eclipse.osee.framework.jdk.core.util.Strings; import org.eclipse.osee.framework.resource.management.IResource; import org.eclipse.osee.framework.resource.management.IResourceLocator; @@ -100,15 +99,9 @@ public class IndexerDataSourceImpl implements IndexedResource { toReturn = resource.getContent(); } catch (OseeCoreException ex) { throw new IOException(ex); - } finally { - Lib.close(toReturn); - } - } else { - try { - toReturn = new ByteArrayInputStream(getStringValue().getBytes("UTF-8")); - } finally { - Lib.close(toReturn); } + } else if (Strings.isValid(getStringValue())) { + toReturn = new ByteArrayInputStream(getStringValue().getBytes("UTF-8")); } return toReturn; } diff --git a/plugins/org.eclipse.osee.orcs.db/src/org/eclipse/osee/orcs/db/internal/search/tagger/TagProcessor.java b/plugins/org.eclipse.osee.orcs.db/src/org/eclipse/osee/orcs/db/internal/search/tagger/TagProcessor.java index fb71215fc6b..777454073c4 100644 --- a/plugins/org.eclipse.osee.orcs.db/src/org/eclipse/osee/orcs/db/internal/search/tagger/TagProcessor.java +++ b/plugins/org.eclipse.osee.orcs.db/src/org/eclipse/osee/orcs/db/internal/search/tagger/TagProcessor.java @@ -40,8 +40,12 @@ public class TagProcessor { public void collectFromInputStream(InputStream inputStream, TagCollector tagCollector) { if (inputStream != null) { Scanner scanner = new Scanner(inputStream, "UTF-8"); - while (scanner.hasNext()) { - processWord(scanner.next(), tagCollector); + try { + while (scanner.hasNext()) { + processWord(scanner.next(), tagCollector); + } + } finally { + scanner.close(); } } } |