Fb2RSS

commit 53b909191ac05f4bcc69746c185aca93aa41444b
parent 16b892d26bfa0edf70cf45c30385520f6a9e2475
Author: Dominik Schmidt <das1993@hotmail.com>
Date:   Sat, 13 May 2017 11:09:49 +0200

Correct the faulty XML delivered by Facebook prior to parsing

Diffstat:
Makefile
 | 
2
+-
fbstream.d
 | 
14
++++++++++++++

2 files changed, 15 insertions(+), 1 deletion(-)
diff --git a/Makefile b/Makefile
@@ -14,7 +14,7 @@ standardpaths/libstandardpaths.a: standardpaths/source/standardpaths.o
 	$(DMD) $(IOPTS) -c $< -of$@
 .PHONY:
 DRSS/drss.a: 
-	cd DRSS/; make DMD="$(DMD)" OPTS="$(OPTS) -version=FORGIVING" drss.a
+	cd DRSS/; make DMD="$(DMD)" drss.a
 clean:
 	rm -f *.o standardpaths/source/standardpaths.o
 	cd DRSS/; make clean
diff --git a/fbstream.d b/fbstream.d
@@ -125,6 +125,20 @@ class FBStream : DRSS!(Post){
 	override public void parse(string document){
 		XmlNode[] arr;
 		XmlNode root;
+		
+		//Make the HTML valid for the parser
+		import std.regex;
+		/*
+		 * Scripts aren't properly commented, so just replace them with comments
+		 * We don't need them anyways
+		 */
+		auto script_start=ctRegex!"<script[^>]*>";
+		auto script_end=ctRegex!"</script>";
+		document=document.replaceAll(script_start, "<!--").replaceAll(script_end, "-->");
+		
+		//Add important End-Tags
+		document~="</body></HTML>";
+		
 		root=readDocument(document);
 		if(!captchaSolved(document)){
 			throw new CaptchaException("Captcha has not been solved yet. "