X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2Fsupport%2Fdocstream.cpp;h=e8839e07f61420104c3ff49826f7dc4997123ce6;hb=faa87bf9f30b943397429a04254d96963bbf38bc;hp=82ec8dc54d270e25a9b85381db0599a647426abb;hpb=58729e1117a77aa5d1088284ab910e16e73f3790;p=lyx.git diff --git a/src/support/docstream.cpp b/src/support/docstream.cpp index 82ec8dc54d..e8839e07f6 100644 --- a/src/support/docstream.cpp +++ b/src/support/docstream.cpp @@ -11,16 +11,27 @@ #include #include "support/docstream.h" +#include "support/lstrings.h" #include "support/unicode.h" +#include #include #include +#include #include #include using namespace std; using lyx::ucs4_codeset; +using lyx::support::contains; +using lyx::support::split; + + +#if defined(_MSC_VER) && (_MSC_VER >= 1600) +std::locale::id numpunct::id; +#endif + namespace { @@ -83,13 +94,50 @@ protected: extern_type * to, extern_type * to_end, extern_type *& to_next) const { +#define WORKAROUND_ICONV_BUG 1 +#if WORKAROUND_ICONV_BUG + // Due to a bug in some iconv versions, when the last char + // in the buffer is a wide char and the output encoding is + // ISO-2022-JP and we are going to switch to another encoding, + // the appropriate escape sequence for changing the character + // set is not output (see bugs 5216, 5280, and also 5489). + // As a workaround, we append a nul char in order to force + // a switch to ASCII, and then remove it from output after + // the conversion. + intern_type * from_new = 0; + intern_type const * from_old = from; + size_t extra = 0; + if (*(from_end - 1) >= 0x80 && encoding_ == "ISO-2022-JP") { + size_t len = from_end - from; + from_new = new intern_type[len + 1]; + memcpy(from_new, from, len * sizeof(intern_type)); + from_new[len] = 0; + from_end = from_new + len + 1; + from = from_new; + extra = 1; + } +#endif size_t inbytesleft = (from_end - from) * sizeof(intern_type); size_t outbytesleft = (to_end - to) * sizeof(extern_type); +#if WORKAROUND_ICONV_BUG + outbytesleft += extra * sizeof(extern_type); +#endif from_next = from; to_next = to; result const retval = do_iconv(out_cd_, reinterpret_cast(&from_next), &inbytesleft, &to_next, &outbytesleft); +#if WORKAROUND_ICONV_BUG + // Remove from output the nul char that we inserted at the end + // of the input buffer in order to circumvent an iconv bug. + if (from_new) { + --to_next; + --from_next; + from_next = from_old + (from_next - from); + from = from_old; + delete[] from_new; + } +#endif if (retval == base::error) { fprintf(stderr, "Error %d returned from iconv when converting from %s to %s: %s\n", @@ -204,32 +252,7 @@ protected: } virtual int do_max_length() const throw() { - // FIXME: this information should be transferred to lib/encodings - // UTF8 uses at most 4 bytes to represent one UCS4 code point - // (see RFC 3629). RFC 2279 specifies 6 bytes, but that - // information is outdated, and RFC 2279 has been superseded by - // RFC 3629. - // The CJK encodings use (different) multibyte representation as well. - // All other encodings encode one UCS4 code point in one byte - // (and can therefore only encode a subset of UCS4) - // Note that BIG5 and SJIS do not work with LaTeX (see lib/encodings). - // Furthermore, all encodings that use shifting (like SJIS) do not work with - // iconv_codecvt_facet. - if (encoding_ == "UTF-8" || - encoding_ == "GB" || - encoding_ == "EUC-TW") - return 4; - else if (encoding_ == "EUC-JP") - return 3; - else if (encoding_ == "BIG5" || - encoding_ == "EUC-KR" || - encoding_ == "EUC-CN" || - encoding_ == "SJIS" || - encoding_ == "GBK" || - encoding_ == "JIS" ) - return 2; - else - return 1; + return lyx::max_encoded_bytes(encoding_); } private: /// Do the actual conversion. The interface is equivalent to that of @@ -242,6 +265,14 @@ private: inbytesleft, to, outbytesleft); if (converted == (size_t)(-1)) { switch(errno) { + case 0: + // As strange as it may seem, this + // does happen on windows when parsing + // comments with accented chars in + // tex2lyx. See the following thread + // for details + // http://thread.gmane.org/gmane.editors.lyx.devel/117636 + break; case EINVAL: case E2BIG: return base::partial; @@ -281,13 +312,13 @@ const char * iconv_codecvt_facet_exception::what() const throw() } -idocfstream::idocfstream(string const & encoding) : base() +ifdocstream::ifdocstream() : base() { - setEncoding(*this, encoding, in); + setEncoding(*this, "UTF-8", in); } -idocfstream::idocfstream(const char* s, ios_base::openmode mode, +ifdocstream::ifdocstream(const char* s, ios_base::openmode mode, string const & encoding) : base() { @@ -296,13 +327,13 @@ idocfstream::idocfstream(const char* s, ios_base::openmode mode, } -odocfstream::odocfstream(): base() +ofdocstream::ofdocstream(): base() { setEncoding(*this, "UTF-8", out); } -odocfstream::odocfstream(const char* s, ios_base::openmode mode, +ofdocstream::ofdocstream(const char* s, ios_base::openmode mode, string const & encoding) : base() { @@ -311,7 +342,7 @@ odocfstream::odocfstream(const char* s, ios_base::openmode mode, } -void odocfstream::reset(string const & encoding) +void ofdocstream::reset(string const & encoding) { setEncoding(*this, encoding, out); } @@ -344,6 +375,28 @@ odocstream & operator<<(odocstream & os, SetEnc e) } +//CHECKME: I just copied the code above, and have no idea whether it +//is correct... (JMarc) +idocstream & operator<<(idocstream & is, SetEnc e) +{ + if (has_facet(is.rdbuf()->getloc())) { + // This stream must be a file stream, since we never imbue + // any other stream with a locale having a iconv_codecvt_facet. + // Flush the stream so that all pending output is written + // with the old encoding. + //is.flush(); + locale locale(is.rdbuf()->getloc(), + new iconv_codecvt_facet(e.encoding, ios_base::in)); + // FIXME Does changing the codecvt facet of an open file + // stream always work? It does with gcc 4.1, but I have read + // somewhere that it does not with MSVC. + // What does the standard say? + is.imbue(locale); + } + return is; +} + + #if ! defined(USE_WCHAR_T) odocstream & operator<<(odocstream & os, char c) { @@ -352,6 +405,157 @@ odocstream & operator<<(odocstream & os, char c) } #endif + +void otexstream::put(char_type const & c) +{ + if (protectspace_) { + if (!canbreakline_ && c == ' ') + os_ << "{}"; + protectspace_ = false; + } + os_.put(c); + lastchar_ = c; + if (c == '\n') { + texrow_.newline(); + canbreakline_ = false; + } else + canbreakline_ = true; +} + + +BreakLine breakln; +SafeBreakLine safebreakln; + + +otexstream & operator<<(otexstream & ots, BreakLine) +{ + if (ots.canBreakLine()) { + ots.os().put('\n'); + ots.lastChar('\n'); + ots.canBreakLine(false); + ots.texrow().newline(); + } + ots.protectSpace(false); + return ots; +} + + +otexstream & operator<<(otexstream & ots, SafeBreakLine) +{ + if (ots.canBreakLine()) { + ots.os() << "%\n"; + ots.lastChar('\n'); + ots.canBreakLine(false); + ots.texrow().newline(); + } + ots.protectSpace(false); + return ots; +} + + +otexstream & operator<<(otexstream & ots, odocstream_manip pf) +{ + ots.os() << pf; + if (pf == static_cast(endl)) { + ots.lastChar('\n'); + ots.texrow().newline(); + } + return ots; +} + + +otexstream & operator<<(otexstream & ots, docstring const & s) +{ + size_t const len = s.length(); + + // Check whether there's something to output + if (len == 0) + return ots; + + if (ots.protectSpace()) { + if (!ots.canBreakLine() && s[0] == ' ') + ots.os() << "{}"; + ots.protectSpace(false); + } + + if (contains(s, 0xF0000)) { + // Some encoding changes for the underlying stream are embedded + // in the docstring. The encoding names to be used are enclosed + // between the code points 0xF0000 and 0xF0001, the first two + // characters of plane 15, which is a Private Use Area whose + // codepoints don't have any associated glyph. + docstring s1; + docstring s2 = split(s, s1, 0xF0000); + while (true) { + if (!s1.empty()) + ots.os() << s1; + if (s2.empty()) + break; + docstring enc; + docstring const s3 = split(s2, enc, 0xF0001); + if (!contains(s2, 0xF0001)) + s2 = split(enc, s1, 0xF0000); + else { + ots.os() << setEncoding(to_ascii(enc)); + s2 = split(s3, s1, 0xF0000); + } + } + } else + ots.os() << s; + + ots.lastChar(s[len - 1]); + ots.texrow().newlines(count(s.begin(), s.end(), '\n')); + ots.canBreakLine(s[len - 1] != '\n'); + return ots; +} + + +otexstream & operator<<(otexstream & ots, string const & s) +{ + ots << from_utf8(s); + return ots; +} + + +otexstream & operator<<(otexstream & ots, char const * s) +{ + ots << from_utf8(s); + return ots; +} + + +otexstream & operator<<(otexstream & ots, char c) +{ + if (ots.protectSpace()) { + if (!ots.canBreakLine() && c == ' ') + ots.os() << "{}"; + ots.protectSpace(false); + } + ots.os() << c; + ots.lastChar(c); + if (c == '\n') + ots.texrow().newline(); + ots.canBreakLine(c != '\n'); + return ots; +} + + +template +otexstream & operator<<(otexstream & ots, Type value) +{ + ots.os() << value; + ots.lastChar(0); + ots.canBreakLine(true); + ots.protectSpace(false); + return ots; +} + +template otexstream & operator<< (otexstream & os, SetEnc); +template otexstream & operator<< (otexstream &, double); +template otexstream & operator<< (otexstream &, int); +template otexstream & operator<< (otexstream &, unsigned int); +template otexstream & operator<< (otexstream &, unsigned long); + } #if ! defined(USE_WCHAR_T) && defined(__GNUC__)