Make the output of \inputencoding commands (and the change of output
encodings) more fine grained: From paragraph level to individual character
level. The inputenc package supports that since at least 2000.
* src/insets/insetbase.h
(latex): Document the differences between output to a string stream
and a file stream
* src/buffer.h
(writeLaTeXSource): ditto
* src/output_latex.h
(latexParagraphs): ditto
(switchEncoding): new function to change the encoding of a stream
(and output \inputencoding commands)
* src/paragraph_pimpl.C
(Paragraph::Pimpl::simpleTeXSpecialC): Adjust to latexWriteEndChanges
changes
* src/support/docstream.[Ch]
(setEncoding, operator<<): New stream modifier that changes the
encoding of file streams
* src/lyxfont.[Ch]
(LyXFont::latexWriteStartChanges): Change stream encoding if needed
(LyXFont::latexWriteEndChanges): Change stream encoding if needed
* src/lyxfont.h
(public:):
* src/paragraph.C
(Paragraph::simpleTeXOnePar): Adjust to latexWriteStartChanges and latexWriteEndChanges changes
(bool Paragraph::simpleTeXOnePar):
(bool Paragraph::simpleTeXOnePar):
(bool Paragraph::simpleTeXOnePar):
(bool Paragraph::simpleTeXOnePar):
(bool Paragraph::simpleTeXOnePar):
* src/output_latex.C
(TeXOnePar): Remove the ugly hack to for wencoding changes and use
switchEncoding instead. A nice side effect is that the old hack would
not work if the main language encoding is latin1 and a character
would be mapped to a cedilla in the "fake ucs4" encoding, because
iconv refuses to convert such a character to latin1, although it
exists in latin1 (it wants to attach it to a base character).
git-svn-id: svn://svn.lyx.org/lyx/lyx-devel/trunk@16633
a592a061-630c-0410-9148-
cb99ea01b6c8