Properly communicate forced encodings

This is currently only relevant fo InsetListings, which falls back to
a fixed-width encoding under specific conditions. It is now possible
to query the inset about that and report the correct encoding in
DocIterator::getEncoding.

Addresses the second part of #10995

diff --git a/src/DocIterator.cpp b/src/DocIterator.cpp
index beb4b2ad50332f2bcc1c7fefc0e74dc80bcde72b..895ebc68c2211cd44b67b14a28e493b1d0200e27 100644 (file)
--- a/src/DocIterator.cpp
+++ b/src/DocIterator.cpp
@@ -731,6 +731,24 @@ Encoding const * DocIterator::getEncoding() const
                (customenc && lang->encoding()->package() != Encoding::CJK)
                ? &bp.encoding() : lang->encoding();
 
+       // Some insets force specific encodings sometimes (e.g., listings in
+       // multibyte context forces singlebyte).
+       if (inset().forcedEncoding(enc, encodings.fromLyXName("iso8859-1"))) {
+               // Get the language outside the inset
+               size_t const n = depth();
+               for (size_t i = 0; i < n; ++i) {
+                       Text const & otext = *slices_[i].text();
+                       Language const * olang =
+                                       otext.getPar(slices_[i].pit()).getFont(bp, slices_[i].pos(),
+                                                                                                                  otext.outerFont(slices_[i].pit())).language();
+                       Encoding const * oenc = olang->encoding();
+                       if (oenc->name() != "inherit")
+                               return inset().forcedEncoding(enc, oenc);
+               }
+               // Fall back to buffer encoding if no outer lang was found.
+               return inset().forcedEncoding(enc, &bp.encoding());
+       }
+
        // Inherited encoding (latex_language) is determined by the context
        // Look for the first outer encoding that is not itself "inherit"
        if (lang->encoding()->name() == "inherit") {

diff --git a/src/insets/Inset.h b/src/insets/Inset.h
index 633005b4a3a4d4d17d462e0955b3e3ad1b8ada5f..ecd171dc45de302844373c860c378171d198f662 100644 (file)
--- a/src/insets/Inset.h
+++ b/src/insets/Inset.h
@@ -39,6 +39,7 @@ class Cursor;
 class CursorSlice;
 class Dimension;
 class DocIterator;
+class Encoding;
 class FuncRequest;
 class FuncStatus;
 class InsetArgument;
@@ -422,6 +423,10 @@ public:
        /// if this inset has paragraphs should they be forced to use a
        /// local font language switch?
        virtual bool forceLocalFontSwitch() const { return false; }
+       /// Does the inset force a specific encoding?
+       virtual Encoding const * forcedEncoding(Encoding const *, Encoding const *) const
+       { return 0; }
+
 
        /// Is the content of this inset part of the output document?
        virtual bool producesOutput() const { return true; }

diff --git a/src/insets/InsetListings.cpp b/src/insets/InsetListings.cpp
index 633a2441b1c13345c96d16126a8aee2aa0720ab4..d66112e6c81e63bee7727060268eaadfdce98b75 100644 (file)
--- a/src/insets/InsetListings.cpp
+++ b/src/insets/InsetListings.cpp
@@ -119,6 +119,26 @@ void InsetListings::read(Lexer & lex)
 }
 
 
+Encoding const * InsetListings::forcedEncoding(Encoding const * inner_enc,
+                                                                                          Encoding const * outer_enc) const
+{
+       // The listings package cannot deal with multi-byte-encoded
+       // glyphs, except if full-unicode aware backends
+       // such as XeTeX or LuaTeX are used, and with pLaTeX.
+       // Minted can deal with all encodings.
+       if (buffer().params().use_minted
+               || (buffer().params().encoding().package() == Encoding::japanese
+                       && inner_enc->package() == Encoding::japanese)
+               || inner_enc->hasFixedWidth())
+               return 0;
+
+       // We try if there's a singlebyte encoding for the outer
+       // language; if not, fall back to latin1.
+       return (outer_enc->hasFixedWidth()) ?
+                       outer_enc : encodings.fromLyXName("iso8859-1");
+}
+
+
 void InsetListings::latex(otexstream & os, OutputParams const & runparams) const
 {
        string param_string = params().params();
@@ -160,30 +180,19 @@ void InsetListings::latex(otexstream & os, OutputParams const & runparams) const
 
        bool encoding_switched = false;
        Encoding const * const save_enc = runparams.encoding;
-       // The listings package cannot deal with multi-byte-encoded
-       // glyphs, except if full-unicode aware backends
-       // such as XeTeX or LuaTeX are used, and with pLaTeX.
-       bool const multibyte_possible = use_minted || runparams.isFullUnicode()
-           || (buffer().params().encoding().package() == Encoding::japanese
-               && runparams.encoding->package() == Encoding::japanese);
 
-       if (!multibyte_possible && !runparams.encoding->hasFixedWidth()) {
+       Encoding const * const outer_encoding =
+               (runparams.local_font != 0) ?
+                       runparams.local_font->language()->encoding()
+                       : buffer().params().language->encoding();
+       Encoding const * fixedlstenc = forcedEncoding(runparams.encoding, outer_encoding);
+       if (fixedlstenc) {
                // We need to switch to a singlebyte encoding, due to
                // the restrictions of the listings package (see above).
                // This needs to be consistent with
                // LaTeXFeatures::getTClassI18nPreamble().
-               Language const * const outer_language =
-                       (runparams.local_font != 0) ?
-                               runparams.local_font->language()
-                               : buffer().params().language;
-               // We try if there's a singlebyte encoding for the current
-               // language; if not, fall back to latin1.
-               Encoding const * const lstenc =
-                       (outer_language->encoding()->hasFixedWidth()) ?
-                               outer_language->encoding()
-                               : encodings.fromLyXName("iso8859-1");
-               switchEncoding(os.os(), buffer().params(), runparams, *lstenc, true);
-               runparams.encoding = lstenc;
+               switchEncoding(os.os(), buffer().params(), runparams, *fixedlstenc, true);
+               runparams.encoding = fixedlstenc;
                encoding_switched = true;
        }
 
@@ -334,7 +343,7 @@ void InsetListings::latex(otexstream & os, OutputParams const & runparams) const
        if (!uncodable.empty() && !runparams.silent) {
                // issue a warning about omitted characters
                // FIXME: should be passed to the error dialog
-               if (!multibyte_possible && !runparams.encoding->hasFixedWidth())
+               if (fixedlstenc)
                        frontend::Alert::warning(_("Uncodable characters in listings inset"),
                                bformat(_("The following characters in one of the program listings are\n"
                                          "not representable in the current encoding and have been omitted:\n%1$s.\n"

diff --git a/src/insets/InsetListings.h b/src/insets/InsetListings.h
index 6c90de0a38a20dbaa24a16978295101616b68efa..4be28f850dd942c108d5e83b021f3e9e7f1927a1 100644 (file)
--- a/src/insets/InsetListings.h
+++ b/src/insets/InsetListings.h
@@ -79,6 +79,8 @@ private:
        TexString getCaption(OutputParams const &) const;
        ///
        bool insetAllowed(InsetCode c) const { return c == CAPTION_CODE || c == QUOTE_CODE; }
+       ///
+       Encoding const * forcedEncoding(Encoding const *, Encoding const *) const;
 
        ///
        InsetListingsParams params_;