After this patch I don't think it pays to fiddle a lot more with
unicode.C. Except moving away from iconv that is.
If you have the chance, please test this.
I will commit this quite soon.
Index: src/support/unicode.C
===================================================================
--- src/support/unicode.C (revision 14966)
+++ src/support/unicode.C (working copy)
@@ -20,10 +20,8 @@
#include <cerrno>
#include <iomanip>
-#include <string>
using std::endl;
-using std::string;
namespace {
@@ -35,36 +33,42 @@
char const * ucs2_codeset = "UCS-2LE";
#endif
-std::vector<char>
-iconv_convert(std::string const & tocode, std::string const & fromcode,
- std::vector<char> const & buf)
-{
- if (buf.empty())
- return std::vector<char>();
-
- iconv_t cd = iconv_open(tocode.c_str(), fromcode.c_str());
- if (cd == (iconv_t)(-1)) {
- lyxerr << "Error returned from iconv_open" << endl;
- switch (errno) {
- case EINVAL:
- lyxerr << "EINVAL The conversion from " << fromcode
- << " to " << tocode
- << " is not supported by the implementation."
- << endl;
- break;
- default:
- lyxerr << "\tSome other error: " << errno << endl;
- break;
+template<typename RetType, typename InType>
+std::vector<RetType>
+iconv_convert(iconv_t * cd,
+ char const * tocode,
+ char const * fromcode,
+ InType const * buf,
+ size_t buflen)
+{
+ if (buflen == 0)
+ return std::vector<RetType>();
+
+ if (*cd == (iconv_t)(-1)) {
+ *cd = iconv_open(tocode, fromcode);
+ if (*cd == (iconv_t)(-1)) {
+ lyxerr << "Error returned from iconv_open" << endl;
+ switch (errno) {
+ case EINVAL:
+ lyxerr << "EINVAL The conversion from " << fromcode
+ << " to " << tocode
+ << " is not supported by the implementation."
+ << endl;
+ break;
+ default:
+ lyxerr << "\tSome other error: " << errno << endl;
+ break;
+ }
}
}
- char ICONV_CONST * inbuf = const_cast<char ICONV_CONST *>(&buf[0]);
- size_t inbytesleft = buf.size();
+ char ICONV_CONST * inbuf = const_cast<char ICONV_CONST *>(reinterpret_cast<char const *>(buf));
+ size_t inbytesleft = buflen * sizeof(InType);
static char out[1000];
char * outbuf = out;
size_t outbytesleft = 1000;
- size_t res = iconv(cd, &inbuf, &inbytesleft, &outbuf, &outbytesleft);
+ size_t res = iconv(*cd, &inbuf, &inbytesleft, &outbuf, &outbytesleft);
if (res == (size_t)(-1)) {
lyxerr << "Error returned from iconv" << endl;
@@ -78,7 +82,7 @@
<< "When converting from " << fromcode
<< " to " << tocode << ".\n";
lyxerr << "Input: " << std::hex;
- for (size_t i = 0; i < buf.size(); ++i) {
+ for (size_t i = 0; i < buflen; ++i) {
unsigned char const b = buf[i];
lyxerr << "0x" << int(b) << " ";
}
@@ -90,7 +94,7 @@
<< "When converting from " << fromcode
<< " to " << tocode << ".\n";
lyxerr << "Input: " << std::hex;
- for (size_t i = 0; i < buf.size(); ++i) {
+ for (size_t i = 0; i < buflen; ++i) {
unsigned char const b = buf[i];
lyxerr << "0x" << int(b) << " ";
}
@@ -100,11 +104,12 @@
lyxerr << "\tSome other error: " << errno << endl;
break;
}
- }
-
- if (iconv_close(cd) == -1) {
- lyxerr << "Error returned from iconv_close("
- << errno << ")" << endl;
+ // We got an error so we close down the conversion engine
+ if (iconv_close(*cd) == -1) {
+ lyxerr << "Error returned from iconv_close("
+ << errno << ")" << endl;
+ }
+ *cd = (iconv_t)(-1);
}
//lyxerr << std::dec;
@@ -112,22 +117,8 @@
//lyxerr << "Outbytesleft: " << outbytesleft << endl;
int bytes = 1000 - outbytesleft;
- std::vector<char> outvec(out, out + bytes);
- return outvec;
-}
-
-
-std::vector<boost::uint32_t> bytes_to_ucs4(std::vector<char> const & bytes)
-{
- boost::uint32_t const * tmp = reinterpret_cast<uint32_t const *>(&bytes[0]);
- return std::vector<boost::uint32_t>(tmp, tmp + bytes.size() / 4);
-}
-
-
-std::vector<unsigned short> bytes_to_ucs2(std::vector<char> const & bytes)
-{
- unsigned short const * tmp = reinterpret_cast<unsigned short const *>(&bytes[0]);
- return std::vector<unsigned short>(tmp, tmp + bytes.size() / 2);
+ RetType const * tmp = reinterpret_cast<RetType const *>(out);
+ return std::vector<RetType>(tmp, tmp + bytes / sizeof(RetType));
}
} // anon namespace
@@ -135,69 +126,55 @@
std::vector<boost::uint32_t> utf8_to_ucs4(std::vector<char> const & utf8str)
{
- //lyxerr << "Buff = " << string(utf8str.begin(), utf8str.end())
- // << " (" << utf8str.size() << ")" << endl;
- //lyxerr << "Res = " << string(res.begin(), res.end())
- // << " (" << res.size() << ")" << endl;
-
- std::vector<char> res = iconv_convert(ucs4_codeset, "UTF-8", utf8str);
- return bytes_to_ucs4(res);
+ static iconv_t cd = (iconv_t)(-1);
+ return iconv_convert<boost::uint32_t>(&cd, ucs4_codeset, "UTF-8",
+ &utf8str[0], utf8str.size());
}
std::vector<boost::uint32_t>
ucs2_to_ucs4(std::vector<unsigned short> const & ucs2str)
{
- char const * tin = reinterpret_cast<char const *>(&ucs2str[0]);
- std::vector<char> in(tin, tin + ucs2str.size() * 2);
- std::vector<char> res = iconv_convert(ucs4_codeset, ucs2_codeset, in);
- return bytes_to_ucs4(res);
+ static iconv_t cd = (iconv_t)(-1);
+ return iconv_convert<boost::uint32_t>(&cd, ucs4_codeset, ucs2_codeset,
+ &ucs2str[0], ucs2str.size());
}
std::vector<unsigned short>
ucs4_to_ucs2(std::vector<boost::uint32_t> const & ucs4str)
{
- char const * tin = reinterpret_cast<char const *>(&ucs4str[0]);
- std::vector<char> in(tin, tin + ucs4str.size() * 4);
- std::vector<char> res = iconv_convert(ucs2_codeset, ucs4_codeset, in);
- return bytes_to_ucs2(res);
+ return ucs4_to_ucs2(&ucs4str[0], ucs4str.size());
}
std::vector<unsigned short>
ucs4_to_ucs2(boost::uint32_t const * s, size_t ls)
{
- char const * tin = reinterpret_cast<char const *>(s);
- std::vector<char> in(tin, tin + ls * 4);
- std::vector<char> res = iconv_convert(ucs2_codeset, ucs4_codeset, in);
- return bytes_to_ucs2(res);
+ static iconv_t cd = (iconv_t)(-1);
+ return iconv_convert<unsigned short>(&cd, ucs2_codeset, ucs4_codeset,
+ s, ls);
}
unsigned short
ucs4_to_ucs2(boost::uint32_t c)
{
- char const * tin = reinterpret_cast<char const *>(&c);
- std::vector<char> in(tin, tin + 4);
- std::vector<char> res = iconv_convert(ucs2_codeset, ucs4_codeset, in);
- return bytes_to_ucs2(res)[0];
+ boost::uint32_t tmp[] = { c, 0 };
+ return ucs4_to_ucs2(tmp, 1)[0];
}
std::vector<char> ucs4_to_utf8(std::vector<boost::uint32_t> const & ucs4str)
{
- char const * tin = reinterpret_cast<char const *>(&ucs4str[0]);
- std::vector<char> in(tin, tin + ucs4str.size() * 4);
- std::vector<char> res = iconv_convert("UTF-8", ucs4_codeset, in);
- return res;
+ static iconv_t cd = (iconv_t)(-1);
+ return iconv_convert<char>(&cd, "UTF-8", ucs4_codeset,
+ &ucs4str[0], ucs4str.size());
}
std::vector<char> ucs4_to_utf8(boost::uint32_t c)
{
- char const * tin = reinterpret_cast<char const *>(&c);
- std::vector<char> in(tin, tin + 4);
- std::vector<char> res = iconv_convert("UTF-8", ucs4_codeset, in);
- return res;
+ static iconv_t cd = (iconv_t)(-1);
+ return iconv_convert<char>(&cd, "UTF-8", ucs4_codeset, &c, 1);
}
--
Lgb
