Escape special chars in indexes

We now have the functionality via sub-insets, and unexperienced users
are surprised when they don't get a literal !.

The unescaped chars can still be used via ERT.

lib/doc/UserGuide.lyx

@@ -35194,10 +35194,33 @@ Note also that in \SpecialChar LaTeX
 .
  This also works in \SpecialChar LyX
 ,
- although you have to take care that the character is not formatted.
+ although you 
- For this reason,
+\change_deleted -712698321 1723474901
+have
+\change_inserted -712698321 1723474901
+need
+\change_unchanged
+ to 
+\change_inserted -712698321 1723474911
+insert 
+\change_deleted -712698321 1723474914
+take care that
+\change_unchanged
+ the character 
+\change_deleted -712698321 1723474917
+is not formatted
+\change_inserted -712698321 1723474926
+in \SpecialChar TeX
+ mode
+\change_unchanged
+.
+ 
+\change_deleted -712698321 1723474935
+For this reason,
  a real exclamation mark has to be inserted to an index entry in a specific way.
- See section
+ 
+\change_unchanged
+See section
 \begin_inset space ~
 \end_inset
 
@@ -35493,10 +35516,21 @@ Note that in \SpecialChar LaTeX
 .
  This also works in \SpecialChar LyX
 ,
- although you have to take care that the character is not formatted.
+ although you 
- For this reason,
+\change_deleted -712698321 1723475013
+have to take care that the character is not formatted
+\change_inserted -712698321 1723475013
+need to insert  the character in \SpecialChar TeX
+ mode
+\change_unchanged
+.
+ 
+\change_deleted -712698321 1723475019
+For this reason,
  a real @ character has to be inserted to an index entry in a specific way.
- See section
+ 
+\change_unchanged
+See section
 \begin_inset space ~
 \end_inset
 
@@ -35894,13 +35928,39 @@ status collapsed
 \end_inset
 
 .
- If you use them literally,
+ 
+\change_deleted -712698321 1723475049
+If you use them literally,
  you might get surprising results or even a non-working index.
  In any case,
- you will not get the character itself.
+ you will not get the character itself
+\change_inserted -712698321 1723475153
+Since the respective functionality is achievable in \SpecialChar LyX
+ via sub-insets,
+ these characters are escaped by 
+\begin_inset Flex Code
+status collapsed
+
+\begin_layout Plain Layout
+
+\change_inserted -712698321 1723475136
+\begin_inset Quotes qld
+\end_inset
+
+
+\end_layout
+
+\end_inset
+
+ if you insert them to indexes,
+ so you get the literal character in the output
+\change_unchanged
+.
 \end_layout
 
 \begin_layout Standard
+
+\change_deleted -712698321 1723475160
 In order to use these characters in inset entries,
  they have to be 
 \begin_inset Quotes els
@@ -36002,11 +36062,27 @@ status collapsed
 \end_inset
 
 ,
+\change_inserted -712698321 1723475173
+If you want the special meaning instead,
+ you
+\change_unchanged
  
 \emph on
 must
 \emph default
- be inserted in a TeX code box (see section
+ 
+\change_deleted -712698321 1723475181
+be 
+\change_unchanged
+insert
+\change_deleted -712698321 1723475183
+ed
+\change_unchanged
+ 
+\change_inserted -712698321 1723475191
+these special characters 
+\change_unchanged
+in a TeX code box (see section
 \begin_inset space ~
 \end_inset
 
@@ -37370,6 +37446,8 @@ status collapsed
 
 \backslash
 protect
+\change_unchanged
+
 \end_layout
 
 \end_inset

lib/doc/de/UserGuide.lyx

@@ -34426,10 +34426,8 @@ Beachten Sie außerdem,
  markiert werden.
  Das funktioniert auch in \SpecialChar LyX
 ,
- allerdings müssen Sie darauf achten,
+ allerdings müssen Sie das Zeichen hierfür im \SpecialChar TeX
- dass das Zeichen nicht formatiert ist.
+-Modus eingeben.
- Da das Ausrufezeichen in Indexeinträgen diese spezielle Bedeutung hat,
- müssen echte Anführungszeichen in einer spezifischen Art und Weise eingegeben werden.
  Dies wird in Abschnitt
 \begin_inset space ~
 \end_inset
@@ -34734,10 +34732,9 @@ Beachten Sie,
  abgetrennt werden.
  Das funktioniert auch in \SpecialChar LyX
 ,
- allerdings müssen Sie darauf achten,
+ ,
- dass das Zeichen nicht formatiert ist.
+ allerdings müssen Sie das Zeichen hierfür im \SpecialChar TeX
- Da das @-Zeichen in Indexeinträgen diese spezielle Bedeutung hat,
+-Modus eingeben.
- müssen echte Klammeraffen in einer spezifischen Art und Weise eingegeben werden.
  Dies wird in Abschnitt
 \begin_inset space ~
 \end_inset
@@ -35215,112 +35212,29 @@ status collapsed
 \end_inset
 
 .
- Wenn Sie diese verwenden,
+ Da die damit erzielte Funktionalität in \SpecialChar LyX
- bekommen Sie unter Umständen überraschende Ergebnisse oder die Erstellung des Stichwortverzeichnisses schlägt sogar fehl.
+ durch Untereinfügungen zugänglich ist,
- Jedenfalls bekommen Sie nicht das Zeichen selbst.
+ werden diese Sonderzeichen mit 
+\begin_inset Flex Code
+status collapsed
+
+\begin_layout Plain Layout
+\begin_inset Quotes qld
+\end_inset
+
+
+\end_layout
+
+\end_inset
+
+ maskiert,
+ wenn Sie sie in einen Index eingeben.
+ Dadurch bekommen Sie das Zeichen selbst im Index ausgegeben.
 \end_layout
 
 \begin_layout Standard
-Um diese Zeichen in Stichworteinträgen zu verwenden,
+Wenn Sie die Zeichen in ihrer Spezialbedeutung verwenden wollen,
- müssen Sie spezifisch markiert werden:
+ müssen Sie sie über ein TeX-Code-Kästchen eingeben (siehe Abschnitt
- Ihnen muss ein sog.
- Escape-Zeichen vorangestellt werden,
- das dem Indexprozessor mitteilt,
- dass dieses Zeichen nicht in seiner speziellen Bedeutung gemeint ist.
- Per Voreinstellung ist dieses Escape-Zeichen 
-\begin_inset Flex Code
-status collapsed
-
-\begin_layout Plain Layout
-\begin_inset Quotes qld
-\end_inset
-
-
-\end_layout
-
-\end_inset
-
- (aber das kann man auch ändern).
- Geben Sie also 
-\begin_inset Flex Code
-status collapsed
-
-\begin_layout Plain Layout
-\begin_inset Quotes qld
-\end_inset
-
-!
-\end_layout
-
-\end_inset
-
-,
- 
-\begin_inset Flex Code
-status collapsed
-
-\begin_layout Plain Layout
-\begin_inset Quotes qld
-\end_inset
-
-@
-\end_layout
-
-\end_inset
-
-,
- 
-\begin_inset Flex Code
-status collapsed
-
-\begin_layout Plain Layout
-\begin_inset Quotes qld
-\end_inset
-
-
-\begin_inset Quotes qrd
-\end_inset
-
-
-\end_layout
-
-\end_inset
-
-,
- oder 
-\begin_inset Flex Code
-status collapsed
-
-\begin_layout Plain Layout
-\begin_inset Quotes qld
-\end_inset
-
-|
-\end_layout
-
-\end_inset
-
- ein,
- um das entsprechende Zeichen im Stichwortverzeichnis zu bekommen.
- Beachten Sie,
- dass das Escape-Zeichen 
-\begin_inset Flex Code
-status collapsed
-
-\begin_layout Plain Layout
-\begin_inset Quotes qld
-\end_inset
-
-
-\end_layout
-
-\end_inset
-
- dabei über ein TeX-Code-Kästchen eingegeben werden 
-\emph on
-muss
-\emph default
- (siehe Abschnitt
 \begin_inset space ~
 \end_inset
 

lib/layouts/stdinsets.inc

@@ -420,7 +420,7 @@ InsetLayout Index
 	CustomPars            false
 	ForcePlain            true
 	ContentAsLabel        true
-	PassThruChars         @|!
+	EscapeChars           \"@|!
 End
 
 InsetLayout IndexMacro:see

lib/lyx2lyx/lyx_2_5.py

@@ -34,8 +34,8 @@ from lyx2lyx_tools import (
 # Uncomment only what you need to import, please (parser_tools):
 #    check_token, count_pars_in_inset, del_complete_lines, 
 #    del_value, find_complete_lines, find_end_of, 
-#    find_re, find_substring, find_token_backwards, find_token_exact,
+#    find_re, find_token_backwards, find_token_exact,
-#    find_tokens, get_bool_value, get_containing_inset,
+#    find_tokens, get_bool_value, 
 #    get_containing_layout, get_option_value,
 #    is_in_inset, set_bool_value
 from parser_tools import (
@@ -43,7 +43,9 @@ from parser_tools import (
     find_end_of_inset,
     find_end_of_layout,
     find_re,
+    find_substring,
     find_token,
+    get_containing_inset,
     get_quoted_value,
     get_value
 )
@@ -620,6 +622,93 @@ def revert_nomencl(document):
         i += 1
 
 
+def convert_index_sc(document):
+    """Convert index special characters to ERT."""
+
+    i = 0
+    while True:
+        i = find_token(document.body, "\\begin_inset Index", i)
+        if i == -1:
+            return
+
+        j = find_end_of_inset(document.body, i)
+        if j == -1:
+            document.warning(
+                "Malformed LyX document: Can't find end of index inset at line %d" % i
+            )
+            i += 1
+            continue
+
+        escchars = ["!", "@", "|"]
+        for ec in escchars:
+            k = i;
+            while True:
+                j = find_end_of_inset(document.body, i)
+                k = find_substring(document.body, ec, k, j)
+                if k == -1:
+                    break
+                if get_containing_inset(document.body, k)[0] == "ERT":
+                    k += 1
+                    continue
+                    
+                line = document.body[k]
+                chunks = line.split(ec)
+                repl = []
+                if line[0] == ec:
+                    repl = put_cmd_in_ert(ec)
+                chunks_len = len(chunks)-1
+                for ch in chunks[:-1]:
+                    repl += [ch]
+                    repl += put_cmd_in_ert(ec)
+                repl += chunks[-1:]
+                if line[-1:] == ec:
+                    repl += put_cmd_in_ert(ec)
+                document.body[k:k+1] = repl
+        i += 1
+
+
+def revert_index_sc(document):
+    """Escape index special characters."""
+
+    i = 0
+    while True:
+        i = find_token(document.body, "\\begin_inset Index", i)
+        if i == -1:
+            return
+
+        j = find_end_of_inset(document.body, i)
+        if j == -1:
+            document.warning(
+                "Malformed LyX document: Can't find end of index inset at line %d" % i
+            )
+            i += 1
+            continue
+
+        escchars = ["!", "@", "|"]
+        for ec in escchars:
+            k = i;
+            while True:
+                j = find_end_of_inset(document.body, i)
+                k = find_substring(document.body, ec, k, j)
+                if k == -1:
+                    break
+                if get_containing_inset(document.body, k)[0] == "ERT":
+                    k += 1
+                    continue
+                    
+                line = document.body[k]
+                chunks = line.split(ec)
+                repl = []
+                chunks_len = len(chunks)-1
+                for ch in chunks[:-1]:
+                    repl += [ch]
+                    repl += put_cmd_in_ert("\"")
+                    repl += [ec]
+                repl += chunks[-1:]
+                document.body[k:k+1] = repl
+                k += len(repl)
+        i += 1
+
 ##
 # Conversion hub
 #
@@ -632,12 +721,12 @@ convert = [
     [624, [convert_biblatex_chicago]],
     [625, []],
     [626, []],
-    [627, [convert_nomencl]]
+    [627, [convert_nomencl, convert_index_sc]]
 ]
 
 
 revert = [
-    [626, [revert_nomencl]],
+    [626, [revert_nomencl, revert_index_sc]],
     [625, [revert_nomencl_textwidth]],
     [624, [revert_nptextcite]],
     [623, [revert_biblatex_chicago]],

src/Layout.cpp

@@ -669,9 +669,11 @@ bool Layout::readIgnoreForcelocal(Lexer & lex, TextClass const & tclass,
 			lex >> pass_thru;
 			break;
 
-		case LT_ESCAPE_CHARS:
+		case LT_ESCAPE_CHARS: {
-			lex >> escape_chars;
+			lex.next(true);
+			escape_chars = lex.getDocString();
 			break;
+		}
 
 		case LT_PASS_THRU_CHARS:
 			lex >> pass_thru_chars;

src/insets/InsetIndex.cpp

@@ -190,6 +190,11 @@ void InsetIndex::latex(otexstream & ios, OutputParams const & runparams_in) cons
 		// These are the LaTeX and plaintext representations
 		docstring latexstr = ourlatex.str();
 		docstring plainstr = ourplain.str();
+
+		// get the escape char from the layout
+		docstring const escape_char = (runparams.escape_chars.empty())
+				? from_ascii("\"")
+				: runparams.escape_chars.substr(0,1);
 	
 		// This will get what follows | if anything does,
 		// the command (e.g., see, textbf) for pagination
@@ -205,10 +210,10 @@ void InsetIndex::latex(otexstream & ios, OutputParams const & runparams_in) cons
 			cmd = from_utf8(params_.pagefmt);
 		} else {
 			// Check for the | separator to strip the cmd.
-			// This goes wrong on an escaped "|", but as the escape
+			// Consider escaped "|"
-			// character can be changed in style files, we cannot
-			// prevent that.
 			size_t pos = latexstr.find(from_ascii("|"));
+			while (pos > 0 && pos < docstring::npos && prefixIs(latexstr.substr(pos - 1), escape_char))
+				pos = latexstr.find(from_ascii("|"), pos + 1);
 			if (pos != docstring::npos) {
 				// Put the bit after "|" into cmd...
 				cmd = latexstr.substr(pos + 1);
@@ -228,23 +233,22 @@ void InsetIndex::latex(otexstream & ios, OutputParams const & runparams_in) cons
 		getSubentries(otsub, runparams, ourlatex.str());
 		if (subentries.str().empty()) {
 			// Separate the entries and subentries, i.e., split on "!".
-			// This goes wrong on an escaped "!", but as the escape
+			// Consider escaped "!"
-			// character can be changed in style files, we cannot
+			// temporarily replace "! with SUBST character
-			// prevent that.
+			docstring const sub_latexstring =
+					subst(latexstr, escape_char + from_ascii("!"), docstring(1, 0x001A));
 			std::vector<docstring> const levels =
-					getVectorFromString(latexstr, from_ascii("!"), true);
+					getVectorFromString(sub_latexstring, from_ascii("!"), true);
-			std::vector<docstring> const levels_plain =
-					getVectorFromString(plainstr, from_ascii("!"), true);
 		
 			vector<docstring>::const_iterator it = levels.begin();
 			vector<docstring>::const_iterator end = levels.end();
-			vector<docstring>::const_iterator it2 = levels_plain.begin();
 			bool first = true;
 			for (; it != end; ++it) {
-				if ((*it).empty()) {
+				// replace back "!
+				docstring const thislevel =
+					subst(*it, docstring(1, 0x001A), escape_char + from_ascii("!"));
+				if (thislevel.empty()) {
 					emptySubentriesWarning(ourlatex.str());
-					if (it2 < levels_plain.end())
-						++it2;
 					continue;
 				}
 				// The separator needs to be put back when
@@ -261,15 +265,12 @@ void InsetIndex::latex(otexstream & ios, OutputParams const & runparams_in) cons
 				// e.g. \index{LyX@\LyX}, \index{text@\textbf{text}}.
 				// We do this on all levels.
 				// We don't do it if the level already contains a '@', though.
-				// Plaintext might return nothing (e.g. for ERTs).
+				// We use a somewhat "plain" representation for this
-				// In that case, we use LaTeX.
+				docstring const spart = Encodings::convertLaTeXCommands(thislevel);
-				docstring const spart = (levels_plain.empty() || (*it2).empty()) ? *it : *it2;
+				processLatexSorting(os, runparams, thislevel, spart, escape_char);
-				processLatexSorting(os, runparams, *it, spart);
-				if (it2 < levels_plain.end())
-					++it2;
 			}
 		} else {
-			processLatexSorting(os, runparams, latexstr, plainstr);
+			processLatexSorting(os, runparams, latexstr, plainstr, escape_char);
 			os << subentries.str();
 		}
 
@@ -295,9 +296,13 @@ void InsetIndex::latex(otexstream & ios, OutputParams const & runparams_in) cons
 
 
 void InsetIndex::processLatexSorting(otexstream & os, OutputParams const & runparams,
-				docstring const & latex, docstring const & spart) const
+				     docstring const & latex, docstring const & spart,
+				     docstring const & esc) const
 {
-	if (contains(latex, '\\') && !contains(latex, '@')) {
+	size_t at_pos = latex.find(from_ascii("@"));
+	while (at_pos > 0 && at_pos < docstring::npos && latex.substr(at_pos - 1, at_pos) == esc)
+		at_pos = latex.find(from_ascii("|"), at_pos + 1);
+	if (contains(latex, '\\') && at_pos == docstring::npos) {
 		// Now we need to validate that all characters in
 		// the sorting part are representable in the current
 		// encoding. If not try the LaTeX macro which might

src/insets/InsetIndex.h

@@ -81,7 +81,8 @@ private:
 	void latex(otexstream &, OutputParams const &) const override;
 	///
 	void processLatexSorting(otexstream &, OutputParams const &,
-				 docstring const &, docstring const &) const;
+				 docstring const &, docstring const &,
+				 docstring const &) const;
 	///
 	bool showInsetDialog(BufferView *) const override;
 	///

src/insets/InsetLayout.cpp

@@ -386,9 +386,11 @@ bool InsetLayout::read(Lexer & lex, TextClass const & tclass,
 			lex >> forceplain_;
 			readCustomOrPlain = true;
 			break;
-		case IL_ESCAPE_CHARS:
+		case IL_ESCAPE_CHARS: {
-			lex >> escape_chars_;
+			lex.next(true);
+			escape_chars_ = lex.getDocString();
 			break;
+		}
 		case IL_PASSTHRU:
 			lex >> passthru_;
 			break;
@@ -928,7 +930,7 @@ void InsetLayout::readArgument(Lexer & lex)
 		} else if (tok == "labelfont") {
 			arg.labelfont = lyxRead(lex, arg.labelfont);
 		} else if (tok == "escapechars") {
-			lex.next();
+			lex.next(true);
 			arg.escape_chars = lex.getDocString();
 		} else if (tok == "passthruchars") {
 			lex.next();