changeset 213:dd95cf01602c

working around limit
author edhoprima@gmail.com <edhoprima@gmail.com>
date Tue, 30 Jun 2009 09:23:02 +0000
parents da4c03c7377c
children a6624fb9b317
files moefetch.sh
diffstat 1 files changed, 12 insertions(+), 4 deletions(-) [+]
line wrap: on
line diff
--- a/moefetch.sh	Mon Jun 29 17:35:43 2009 +0000
+++ b/moefetch.sh	Tue Jun 30 09:23:02 2009 +0000
@@ -84,11 +84,15 @@
 # generate link by transforming xml
 Generate_Link() {
 	printf "\nFetching xml file\n"
-	wget "http://${SITE}/post/index.xml?tags=${TAGS}&offset=0&limit=100000" -O "${TEMP_PREFIX}-xml" -e continue=off
-	printf "Processing XML file..."
-	# xslt evilry
+	TEMPNUM=1001
+	_i=1
 	> "${TEMP_PREFIX}-list"
-	xsltproc - "${TEMP_PREFIX}-xml" <<EOF | sed 's/.*\(http.*\)\(\/[a-f0-9]\{32\}\).*\.\([^\.]*\)/\1\2.\3/g' | grep ^http > "${TEMP_PREFIX}-list" 2>/dev/null
+	while [ "${TEMPNUM}" -ge 1000 ]; do
+		wget "http://${SITE}/post/index.xml?tags=${TAGS}&offset=0&limit=1000&page=${_i}" -O "${TEMP_PREFIX}-xml" -e continue=off
+		printf "Processing XML file..."
+		# xslt evilry
+		> "${TEMP_PREFIX}-templist"
+		xsltproc - "${TEMP_PREFIX}-xml" <<EOF | sed 's/.*\(http.*\)\(\/[a-f0-9]\{32\}\).*\.\([^\.]*\)/\1\2.\3/g' | grep ^http > "${TEMP_PREFIX}-templist" 2>/dev/null
 <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
 <xsl:output method="xml" indent="yes"/>
 <xsl:template match="post">
@@ -96,6 +100,10 @@
 </xsl:template>
 </xsl:stylesheet>
 EOF
+		TEMPNUM=$(echo $(wc -l < "${TEMP_PREFIX}-templist"))
+		_i=$((_i+1))
+		cat "${TEMP_PREFIX}-templist" >> "${TEMP_PREFIX}-list"
+	done
 	NUMFILES=$(echo $(wc -l < "${TEMP_PREFIX}-list"))
 	[ "${NUMFILES}" -gt 0 ] || Err_Fatal "Error in processing list or no files can be found with specified tag(s) or site"
 	echo " ${NUMFILES} file(s) available on server"