Half decent error handling in scripts with project2:handler[try/catch/finally]

Component to validate a date string
Merge some code (fixes sqlmerge missing attachments to ifs)
Minor fixes
New changelog importer in GB

2 files changed, 85 insertions, 40 deletions

diff --git a/gentoobrowse/console/changelogs.xml b/gentoobrowse/console/changelogs.xml
new file mode 100644
index 0000000..11235c4
--- /dev/null
+++ b/gentoobrowse/console/changelogs.xml
@@ -0,0 +1,81 @@
+<?xml version="1.0"?>
+<packageimport name="bugassociate" xmlns:xi="http://www.w3.org/2001/XInclude"
+		xmlns:project2="http://project2.randomdan.homeip.net">
+	<project2:sqlrows name="updatedlogs" datasource="postgres">
+		<sql>
+			SELECT f.filename, ft.searchroot || f.filename AS path, p.packageid
+			FROM files f, filetypes ft, packages p, categories c
+			WHERE f.filetypeid = ft.filetypeid
+			AND (f.cachedat IS NULL OR f.cachedat != f.moddate)
+			AND ft.filetypeid = 2
+			AND p.name = SPLIT_PART(filename, '/', 2)
+			AND c.name = SPLIT_PART(filename, '/', 1)
+			AND c.categoryid = p.categoryid
+		</sql>
+	</project2:sqlrows>
+	<project2:filerows name="logcontent" encoding="utf-8" newline="" fieldSep="">
+		<path source="parent" name="path" depth="1" />
+		<columns><column>text</column></columns>
+	</project2:filerows>
+	<project2:regexrows name="logentries">
+		<sourceText source="parent" name="text" depth="1" />
+		<regex>(\d+ *(?:jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\w* \d+); +([^&lt;]+)? *&lt;([^ &gt;]+)[&gt;\s]\s*(?:[^:]*:)?\s*(.*?)(?:\R\s*|$)</regex>
+		<columns>
+			<column>date</column>
+			<column>person</column>
+			<column>email</column>
+			<column>comment</column>
+		</columns>
+	</project2:regexrows>
+	<project2:sqlmerge name="changelogs" datasource="postgres" targettable="newchangelogs">
+		<updatewhere>
+			a.packageid IN (
+					SELECT p.packageid
+					FROM files f, packages p, categories c
+					WHERE f.filetypeid = 2
+					AND (f.cachedat IS NULL OR f.cachedat != f.moddate)
+					AND p.name = SPLIT_PART(filename, '/', 2)
+					AND c.name = SPLIT_PART(filename, '/', 1)
+					AND c.categoryid = p.categoryid)
+		</updatewhere>
+		<project2:iterate name="seelogs" source="updatedlogs">
+			<project2:iterate name="seelog" source="logcontent">
+				<project2:iterate name="seeentries" source="logentries">
+					<project2:if>
+						<project2:validdatecheck format="%d %b %Y">
+							<apply-to source="parent" name="date" depth="1" warn="no" />
+						</project2:validdatecheck>
+						<project2:regexcheck regex=".+@.+">
+							<apply-to source="parent" name="email" depth="1" />
+						</project2:regexcheck>
+						<project2:sqlmergeinsert>
+							<parameters>
+								<param name="packageid"><value source="parent" name="packageid" depth="3" /></param>
+								<param name="date"><value source="parent" name="date" depth="1" /></param>
+								<param name="n"><value source="parent" attribute="rownum" depth="1" /></param>
+								<param name="person"><value source="parent" name="person" depth="1" /></param>
+								<param name="email"><value source="parent" name="email" depth="1" /></param>
+								<param name="comment"><value source="parent" name="comment" depth="1" /></param>
+							</parameters>
+						</project2:sqlmergeinsert>
+					</project2:if>
+				</project2:iterate>
+			</project2:iterate>
+		</project2:iterate>
+		<columns>
+			<column key="true">packageid</column>
+			<column key="true">date</column>
+			<column key="true">n</column>
+			<column>person</column>
+			<column>email</column>
+			<column>comment</column>
+		</columns>
+	</project2:sqlmerge>
+	<project2:sqltask name="update" datasource="postgres">
+		<sql>
+			UPDATE files SET
+				cachedat = moddate
+			WHERE filetypeid = 2
+		</sql>
+	</project2:sqltask>
+</packageimport>
diff --git a/gentoobrowse/console/packageimport.xml b/gentoobrowse/console/packageimport.xml
index b2c48ab..5dba528 100644
--- a/gentoobrowse/console/packageimport.xml
+++ b/gentoobrowse/console/packageimport.xml
@@ -169,46 +169,10 @@
 		</columns>
 	</project2:sqlmerge>
 
-	<project2:sqlmerge name="changelogs" datasource="postgres" targettable="package_changelogs" useview="no">
-		<sql>
-			SELECT packageid,
-				CAST(r[1] AS DATE) AS DATE,
-				RANK() OVER(PARTITION BY packageid, CAST(r[1] AS DATE) ORDER BY s.linebegin) AS n,
-				r[2] AS person,
-				r[3] AS email,
-				(SELECT TRIM(SPLIT_PART(STRING_AGG(m.content, ' ' ORDER BY m.line), ':', 2))
-				 FROM filecache m
-				 WHERE m.fileid=s.fileid
-				 AND m.line &gt;= s.linebegin
-				 AND m.content not like '*%'
-				 AND (m.line &lt; s.lineend OR s.lineend IS NULL)) AS comment
-			FROM (SELECT f.fileid,
-					p.packageid,
-					fc.line AS linebegin,
-					REGEXP_MATCHES(content, '([0-9]{2} [[:alnum:]]+ [0-9]{2,4}); ?(.+)? &lt;([^ &gt;]+)') AS r,
-					(SELECT MIN(line)
-					 FROM filecache nl
-					 WHERE nl.fileid = f.fileid
-					 AND nl.content ~ '([0-9]{2} [[:alnum:]]+ [0-9]{2,4}); ?(.+)? &lt;([^ &gt;]+)'
-					 AND nl.line &gt; fc.line) AS lineend
-					FROM filecache fc, files f, packages p, categories c
-					WHERE f.fileid=fc.fileid
-					AND f.filetypeid = 2
-					AND c.categoryid = p.categoryid
-					AND p.name = SPLIT_PART(f.filename, '/', 2)
-					AND c.name = SPLIT_PART(f.filename, '/', 1)) AS s
-			WHERE IS_DATE(s.r[1])
-		</sql>
-		<columns>
-			<column key="true">packageid</column>
-			<column key="true">date</column>
-			<column key="true">n</column>
-			<column>person</column>
-			<column>email</column>
-			<column>comment</column>
-		</columns>
-	</project2:sqlmerge>
-
+	<xi:include href="changelogs.xml" />
+	<xi:include href="packagelicenses.xml" />
+	<xi:include href="use.global.xml" />
+	<xi:include href="use.local.xml" />
 	<xi:include href="categorymetadata.xml" />
 	<xi:include href="packagemetadata.xml" />
 	<xi:include href="packagemasks.xml" />