From 6f98b4dfb2b80e3eb3072dd3f140351581746db2 Mon Sep 17 00:00:00 2001
From: Jonas Smedegaard <dr@jones.dk>
Date: Tue, 30 Dec 2014 01:50:08 +0100
Subject: Move mediawiki tweaks to separate scripts.

---
 Makefile | 18 ++++--------------
 1 file changed, 4 insertions(+), 14 deletions(-)

(limited to 'Makefile')

diff --git a/Makefile b/Makefile
index c0da69d..be72240 100644
--- a/Makefile
+++ b/Makefile
@@ -2,12 +2,6 @@ stem = eut
 source_baseurl = http://euwiki.org/
 source_basename = EUT/2nd-edition
 
-# work around in mediawiki bug parsing blockquote tags on same line
-re_blockquote = s|\n*(</?blockquote>)\n*|\n$$1\n|g
-
-# work around modest URL parsing <https://github.com/jgm/pandoc/issues/1838>
-re_urlencode = s|\bhttps?://$$RE{net}{domain}{-nospace}{-rfc1101}(?::\d+)?/\K(\S*?)(?=[.,;]?[\s<\"\|}])|uri_escape(uri_unescape($$1), "^A-Za-z0-9/\.")|eg
-
 templates = template.tex header.tex before.tex after.tex
 
 filters = ./pandoc-memoir ./pandoc-cs1
@@ -49,14 +43,10 @@ download:
 	wget -O $(stem).raw '$(source_baseurl)w/index.php?title=$(source_basename)&action=edit'
 
 $(stem).mediawiki: $(stem).raw
-	perl -C -0777 -MHTML::Entities -MRegexp::Common=net -MURI::Escape -p \
-	-e 's|.*<textarea[^>]*>||s; s|</textarea.*||s;' \
-	-e 'decode_entities($$_);' \
-	-e 's|.*?\n= |= |s;' \
-	-e 's|<!--.*-->||s;' \
-	-e '$(re_blockquote);' \
-	-e '$(re_urlencode);' \
-	< $< > $@
+	./mediawiki-extract $< $@
+	./mediawiki-trim $@
+	./mediawiki-blockquote $@
+	./mediawiki-uri-escape $@
 
 $(stem).native: $(stem).mediawiki $(localfilters)
 	pandoc -f mediawiki $(args_meta) $(args_filter) -o $@ $<
-- 
cgit v1.2.3