From 8e27aa2fb7c0b9469ac83f0af49e1f7bc59006f7 Mon Sep 17 00:00:00 2001
From: =?utf8?q?Jos=C3=A9=20Matox?= <jamatos@lyx.org>
Date: Fri, 29 Oct 2004 23:08:04 +0000
Subject: [PATCH] Sanitize ids for SGML/XML.

git-svn-id: svn://svn.lyx.org/lyx/lyx-devel/trunk@9142 a592a061-630c-0410-9148-cb99ea01b6c8
---
 src/ChangeLog               | 11 ++++++
 src/insets/ChangeLog        |  9 +++++
 src/insets/insetgraphics.C  | 13 ++-----
 src/insets/insetlabel.C     |  5 ++-
 src/insets/insetref.C       |  9 +++--
 src/mathed/ChangeLog        |  5 +++
 src/mathed/math_hullinset.C | 48 +++++++++++++++++-------
 src/mathed/ref_inset.C      | 14 +++++--
 src/output_docbook.C        |  3 +-
 src/paragraph.C             |  5 ++-
 src/paragraph.h             |  4 +-
 src/sgml.C                  | 75 +++++++++++++++++++++++++++++++++++--
 src/sgml.h                  |  6 +++
 13 files changed, 165 insertions(+), 42 deletions(-)

diff --git a/src/ChangeLog b/src/ChangeLog
index 72e7e17643..0955a93515 100644
--- a/src/ChangeLog
+++ b/src/ChangeLog
@@ -1,3 +1,14 @@
+2004-10-29  José Matos  <jamatos@lyx.org>
+
+	* output_docbook.C (makeEnvironment):
+	* sgml.C (openTag):
+	* paragraph.[Ch] (getID): rename function, and return it enclosed in id="...".
+
+2004-10-29 Andreas Vox  <vox@isp.uni-luebeck.de>
+
+	* sgml.[Ch] (uniqueID): returns a unique id for a given label.
+	(cleanID): sanitize any id.
+
 2004-10-29  Georg Baum  <Georg.Baum@post.rwth-aachen.de>
 
 	* buffer.C, lyxlex_pimpl.C:
diff --git a/src/insets/ChangeLog b/src/insets/ChangeLog
index 0bd9e7c3d0..1d3430135c 100644
--- a/src/insets/ChangeLog
+++ b/src/insets/ChangeLog
@@ -1,3 +1,12 @@
+2004-10-29  José Matos  <jamatos@lyx.org>
+
+	* insetgraphics.C (uniqueID): transfered to sgml.C
+
+2004-10-29 Andreas Vox  <vox@isp.uni-luebeck.de>
+
+	* insetlabel.C (docbook, linuxdoc):
+	* insetref.C (docbook, linuxdoc) sanitize id.
+
 2004-10-29  Georg Baum  <Georg.Baum@post.rwth-aachen.de>
 
 	* ExternalSupport.C: s/getExtFromContents/getFormatFromContents/
diff --git a/src/insets/insetgraphics.C b/src/insets/insetgraphics.C
index 956819d798..71f4a3d531 100644
--- a/src/insets/insetgraphics.C
+++ b/src/insets/insetgraphics.C
@@ -69,6 +69,7 @@ TODO
 #include "metricsinfo.h"
 #include "mover.h"
 #include "outputparams.h"
+#include "sgml.h"
 
 #include "frontends/Alert.h"
 #include "frontends/LyXView.h"
@@ -78,7 +79,6 @@ TODO
 #include "support/lyxlib.h" // float_equal
 #include "support/os.h"
 #include "support/systemcall.h"
-#include "support/tostr.h"
 
 #include <boost/bind.hpp>
 #include <boost/tuple/tuple.hpp>
@@ -126,13 +126,6 @@ string const RemoveExtension(string const & filename)
 }
 
 
-string const uniqueID()
-{
-	static unsigned int seed = 1000;
-	return "graph" + tostr(++seed);
-}
-
-
 string findTargetFormat(string const & format, OutputParams const & runparams)
 {
 	// Are we using latex or pdflatex?
@@ -160,7 +153,7 @@ string findTargetFormat(string const & format, OutputParams const & runparams)
 
 
 InsetGraphics::InsetGraphics()
-	: graphic_label(uniqueID()),
+	: graphic_label(sgml::uniqueID("graph")),
 	  graphic_(new RenderGraphic(this))
 {}
 
@@ -168,7 +161,7 @@ InsetGraphics::InsetGraphics()
 InsetGraphics::InsetGraphics(InsetGraphics const & ig)
 	: InsetOld(ig),
 	  boost::signals::trackable(),
-	  graphic_label(uniqueID()),
+	  graphic_label(sgml::uniqueID("graph")),
 	  graphic_(new RenderGraphic(*ig.graphic_, this))
 {
 	setParams(ig.params());
diff --git a/src/insets/insetlabel.C b/src/insets/insetlabel.C
index 816821f9c7..c3643f5b95 100644
--- a/src/insets/insetlabel.C
+++ b/src/insets/insetlabel.C
@@ -20,6 +20,7 @@
 #include "lyxtext.h"
 #include "paragraph.h"
 #include "pariterator.h"
+#include "sgml.h"
 
 #include "frontends/LyXView.h"
 
@@ -135,7 +136,7 @@ int InsetLabel::plaintext(Buffer const &, ostream & os,
 int InsetLabel::linuxdoc(Buffer const &, ostream & os,
 			 OutputParams const &) const
 {
-	os << "<label id=\"" << getContents() << "\" >";
+	os << "<label id=\"" << sgml::cleanID(getContents()) << "\" >";
 	return 0;
 }
 
@@ -143,6 +144,6 @@ int InsetLabel::linuxdoc(Buffer const &, ostream & os,
 int InsetLabel::docbook(Buffer const &, ostream & os,
 			OutputParams const &) const
 {
-	os << "<!-- anchor id=\"" << getContents() << "\" -->";
+	os << "<!-- anchor id=\"" << sgml::cleanID(getContents()) << "\" -->";
 	return 0;
 }
diff --git a/src/insets/insetref.C b/src/insets/insetref.C
index f30f071022..b5f82e6748 100644
--- a/src/insets/insetref.C
+++ b/src/insets/insetref.C
@@ -19,6 +19,7 @@
 #include "gettext.h"
 #include "LaTeXFeatures.h"
 #include "outputparams.h"
+#include "sgml.h"
 
 #include "frontends/LyXView.h"
 
@@ -104,7 +105,7 @@ int InsetRef::plaintext(Buffer const &, ostream & os,
 int InsetRef::linuxdoc(Buffer const &, ostream & os,
 		       OutputParams const &) const
 {
-	os << "<ref id=\"" << getContents()
+	os << "<ref id=\"" << sgml::cleanID(getContents())
 	   << "\" name=\"" << getOptions() << "\" >";
 	return 0;
 }
@@ -114,11 +115,11 @@ int InsetRef::docbook(Buffer const &, ostream & os,
 		      OutputParams const & runparams) const
 {
 	if (getOptions().empty() && runparams.flavor == OutputParams::XML) {
-		os << "<xref linkend=\"" << getContents() << "\" />";
+		os << "<xref linkend=\"" << sgml::cleanID(getContents()) << "\" />";
 	} else if (getOptions().empty()) {
-		os << "<xref linkend=\"" << getContents() << "\">";
+		os << "<xref linkend=\"" << sgml::cleanID(getContents()) << "\">";
 	} else {
-		os << "<link linkend=\"" << getContents()
+		os << "<link linkend=\"" << sgml::cleanID(getContents())
 		   << "\">" << getOptions() << "</link>";
 	}
 
diff --git a/src/mathed/ChangeLog b/src/mathed/ChangeLog
index 09d22478e3..63dc990baf 100644
--- a/src/mathed/ChangeLog
+++ b/src/mathed/ChangeLog
@@ -1,3 +1,8 @@
+2004-10-29  Andreas Vox  <vox@isp.uni-luebeck.de>
+
+	* math_hullinset.C (docbook):
+	* ref_inset.C (docbook): sanitize ids, clean exported docbook.
+
 2004-10-15  Georg Baum  <Georg.Baum@post.rwth-aachen.de>
 
 	* math_hullinset.C (mutate): fix endless loop for unknown types
diff --git a/src/mathed/math_hullinset.C b/src/mathed/math_hullinset.C
index 5d068841c4..4248ad291d 100644
--- a/src/mathed/math_hullinset.C
+++ b/src/mathed/math_hullinset.C
@@ -30,6 +30,7 @@
 #include "lyx_main.h"
 #include "lyxrc.h"
 #include "outputparams.h"
+#include "sgml.h"
 #include "textpainter.h"
 #include "undo.h"
 
@@ -40,11 +41,14 @@
 #include "graphics/PreviewImage.h"
 #include "graphics/PreviewLoader.h"
 
+#include "support/lstrings.h"
+
 #include <boost/bind.hpp>
 
 #include <sstream>
 
 using lyx::cap::grabAndEraseSelection;
+using lyx::support::subst;
 
 using std::endl;
 using std::max;
@@ -1122,7 +1126,6 @@ bool MathHullInset::getStatus(LCursor & cur, FuncRequest const & cmd,
 #include "frontends/LyXView.h"
 #include "frontends/Dialogs.h"
 
-#include "support/lstrings.h"
 #include "support/lyxlib.h"
 
 
@@ -1342,25 +1345,44 @@ int MathHullInset::docbook(Buffer const & buf, ostream & os,
 		name = "informalequation";
 
 	string bname = name;
-	if (! label(0).empty()) bname += " id=\"" + label(0)+ "\"";
+	if (!label(0).empty()) 
+		bname += " id=\"" + sgml::cleanID(label(0)) + "\"";
 	ms << MTag(bname.c_str());
 
+	ostringstream ls;
 	if (runparams.flavor == OutputParams::XML) {
-		ms <<   MTag("math");
-		MathGridInset::mathmlize(ms);
-		ms <<   ETag("math");
-		ms <<   MTag("alt role=\"tex\" ");
-		ostringstream ls;
+		ms << MTag("alt role=\"tex\" ");
+		// Workaround for db2latex: db2latex always includes equations with
+		// \ensuremath{} or \begin{display}\end{display}
+		// so we strip LyX' math environment
 		WriteStream wi(ls, false, false);
 		MathGridInset::write(wi);
-		ms << ls.str();
-		ms <<   ETag("alt");
+		ms << subst(subst(ls.str(), "&", "&amp;"), "<", "&lt;");
+		ms << ETag("alt");
+		ms << MTag("math");
+		MathGridInset::mathmlize(ms);
+		ms << ETag("math");
 	} else {
-		ms <<   MTag("alt role=\"tex\" ");
-		res = latex(buf, ms.os(), runparams);
-		ms <<   ETag("alt");
+		ms << MTag("alt role=\"tex\"");		
+		res = latex(buf, ls, runparams);
+		ms << subst(subst(ls.str(), "&", "&amp;"), "<", "&lt;");
+		ms << ETag("alt");
 	}
-
+	
+	ms <<  "<graphic fileref=\"eqn/";
+	if ( !label(0).empty()) 
+		ms << sgml::cleanID(label(0));
+	else {
+		// Some arbitrary unique number for this os. 
+		// Note that each call of math_hullinset::docbook()
+		// will increase the os position by at least 60 chars or more
+		ms << sgml::uniqueID("anon");
+	}
+	if (runparams.flavor == OutputParams::XML) 
+		ms << "\"/>";
+	else 
+		ms << "\">";
+		
 	ms << ETag(name.c_str());
 	return ms.line() + res;
 }
diff --git a/src/mathed/ref_inset.C b/src/mathed/ref_inset.C
index 2706f76bbc..0aac3bbe28 100644
--- a/src/mathed/ref_inset.C
+++ b/src/mathed/ref_inset.C
@@ -21,6 +21,8 @@
 #include "math_data.h"
 #include "math_factory.h"
 #include "math_support.h"
+#include "outputparams.h"
+#include "sgml.h"
 
 #include "frontends/LyXView.h"
 #include "frontends/Dialogs.h"
@@ -131,17 +133,21 @@ int RefInset::plaintext(std::ostream & os, OutputParams const &) const
 int RefInset::linuxdoc(std::ostream & os, OutputParams const &) const
 {
 	os << "<ref id=\"" << asString(cell(0))
-	   << "\" name=\"" << asString(cell(1)) << "\" >";
+	   << "\" name=\"" << asString(cell(1)) << "\">";
 	return 0;
 }
 
 
-int RefInset::docbook(std::ostream & os, OutputParams const &) const
+int RefInset::docbook(std::ostream & os, OutputParams const & runparams) const
 {
 	if (cell(1).empty()) {
-		os << "<xref linkend=\"" << asString(cell(0)) << "\">";
+		os << "<xref linkend=\"" << sgml::cleanID(asString(cell(0)));
+		if (runparams.flavor == OutputParams::XML) 
+			os << "\"/>";
+		else
+			os << "\">";
 	} else {
-		os << "<link linkend=\"" << asString(cell(0))
+		os << "<link linkend=\"" << sgml::cleanID(asString(cell(0)))
 		   << "\">" << asString(cell(1)) << "</link>";
 	}
 
diff --git a/src/output_docbook.C b/src/output_docbook.C
index 593071c11a..69e04977d0 100644
--- a/src/output_docbook.C
+++ b/src/output_docbook.C
@@ -133,8 +133,7 @@ ParagraphList::const_iterator makeEnvironment(Buffer const & buf,
 	while (par != pend) {
 		LyXLayout_ptr const & style = par->layout();
 		ParagraphList::const_iterator send;
-		string id = par->getDocbookId();
-		id = id.empty()? "" : " id = \"" + id + "\"";
+		string id = par->getID();
 		string wrapper = "";
 		pos_type sep = 0;
 
diff --git a/src/paragraph.C b/src/paragraph.C
index 72d7358303..95d8bbfb97 100644
--- a/src/paragraph.C
+++ b/src/paragraph.C
@@ -1331,14 +1331,15 @@ void Paragraph::simpleLinuxDocOnePar(Buffer const & buf,
 }
 
 
-string Paragraph::getDocbookId() const
+string Paragraph::getID() const
 {
 	for (pos_type i = 0; i < size(); ++i) {
 		if (isInset(i)) {
 			InsetBase const * inset = getInset(i);
 			InsetBase::Code lyx_code = inset->lyxCode();
 			if (lyx_code == InsetBase::LABEL_CODE) {
-				return static_cast<InsetCommand const *>(inset)->getContents();
+				string const id = static_cast<InsetCommand const *>(inset)->getContents();
+				return "id=\"" + sgml::cleanID(id) + "\"";
 			}
 		}
 
diff --git a/src/paragraph.h b/src/paragraph.h
index f937fdbd07..a204661bc6 100644
--- a/src/paragraph.h
+++ b/src/paragraph.h
@@ -129,8 +129,8 @@ public:
 				  OutputParams const & runparams,
 				  lyx::depth_type depth) const;
 
-	/// Get the id of the paragraph, usefull for docbook
-	std::string getDocbookId() const;
+	/// Get the id of the paragraph, usefull for docbook and linuxdoc
+	std::string getID() const;
 
 	// Get the first word of a paragraph, return the position where it left
 	lyx::pos_type getFirstWord(Buffer const & buf,
diff --git a/src/sgml.C b/src/sgml.C
index 12f045193a..12f3ae53e3 100644
--- a/src/sgml.C
+++ b/src/sgml.C
@@ -25,12 +25,13 @@
 
 #include <boost/tuple/tuple.hpp>
 
+#include <map>
 #include <sstream>
 
 using lyx::support::subst;
 
 using std::make_pair;
-
+using std::map;
 using std::ostream;
 using std::ostringstream;
 using std::pair;
@@ -112,6 +113,75 @@ string escapeString(string const & raw)
 }
 
 
+string const uniqueID(string const label)
+{
+	static unsigned int seed = 1000;
+	return label + tostr(++seed);
+}
+
+
+string cleanID(std::string const & orig, std::string const & allowed)
+{
+	// The standard DocBook SGML declaration only allows letters,
+	// digits, '-' and '.' in a name.
+	// Since users might change that declaration one has to cater
+	// for additional allowed characters.
+	// This routine replaces illegal characters by '-' or '.'
+	// and adds a number for uniqueness.
+	// If you know what you are doing, you can set allowed==""
+	// to disable this mangling.
+	
+	string::const_iterator it  = orig.begin();
+	string::const_iterator end = orig.end();
+
+	string content;
+
+	if (allowed.empty()) {
+		return orig;
+	}
+
+	typedef map<string, string> MangledMap;
+	static MangledMap mangledNames;
+	static int mangleID = 1;
+
+	MangledMap::const_iterator const known = mangledNames.find(orig);
+	if (known != mangledNames.end())
+		return (*known).second;
+
+	// make sure it starts with a letter
+	if (!isalpha(*it) && allowed.find(*it) >= allowed.size())
+		content += "x";
+	
+	bool mangle = false;	
+	for (; it != end; ++it) {
+		char c = *it;
+		if (isalpha(c) || isdigit(c) || c == '-' || c == '.' || allowed.find(c) < allowed.size())
+			content += c;
+		else if (c == '_' || c == ' ') {
+			mangle = true;
+			content += "-";
+		}
+		else if (c == ':' || c == ',' || c == ';' || c == '!') {
+			mangle = true;
+			content += ".";
+		}
+		else {
+			mangle = true;
+		}
+	}
+	if (mangle) {
+		content += "-" + tostr(mangleID++);
+	}
+	else if (isdigit(content[content.size()-1])) {
+		content += ".";
+	}
+
+	mangledNames[orig] = content;
+
+	return content;
+}
+
+
 void openTag(ostream & os, string const & name, string const & attribute)
 {
 	// This should be fixed in layout files later.
@@ -141,8 +211,7 @@ void openTag(Buffer const & buf, ostream & os, Paragraph const & par)
 	string param = style->latexparam();
 	Counters & counters = buf.params().getLyXTextClass().counters();
 
-	string id = par.getDocbookId();
-	id = id.empty()? "" : " id = \"" + id + "\"";
+	string id = par.getID();
 
 	string attribute;
 	if(!id.empty()) {
diff --git a/src/sgml.h b/src/sgml.h
index ff495d7882..e6b2e4e487 100644
--- a/src/sgml.h
+++ b/src/sgml.h
@@ -34,6 +34,12 @@ std::pair<bool, std::string> escapeChar(char c);
 /// Escape a word instead of a single character
 std::string escapeString(std::string const & raw);
 
+/// replaces illegal chars like ':' or '_' from SGML ID attributes
+std::string cleanID(std::string const & orig, std::string const & allowed = std::string());
+
+/// returns a uniq numeric id
+std::string const uniqueID(std::string const label);
+
 /// Opens tag
 void openTag(std::ostream & os, std::string const & name,
 	    std::string const & attribute = std::string());
-- 
2.39.2