X-Git-Url: https://git.saurik.com/apt.git/blobdiff_plain/b74d2d3af3238210bfafe57562d3670327918cd9..34e8a998eb9e4f75f989acc6f8b9db8010e56b7c:/apt-pkg/contrib/strutl.cc diff --git a/apt-pkg/contrib/strutl.cc b/apt-pkg/contrib/strutl.cc index 1683868c8..b285a9f2e 100644 --- a/apt-pkg/contrib/strutl.cc +++ b/apt-pkg/contrib/strutl.cc @@ -43,9 +43,10 @@ bool UTF8ToCodeset(const char *codeset, const string &orig, string *dest) { iconv_t cd; const char *inbuf; - char *inptr, *outbuf, *outptr; - size_t insize, outsize; - + char *inptr, *outbuf; + size_t insize, bufsize; + dest->clear(); + cd = iconv_open(codeset, "UTF-8"); if (cd == (iconv_t)(-1)) { // Something went wrong @@ -55,22 +56,49 @@ bool UTF8ToCodeset(const char *codeset, const string &orig, string *dest) else perror("iconv_open"); - // Clean the destination string - *dest = ""; - return false; } - insize = outsize = orig.size(); + insize = bufsize = orig.size(); inbuf = orig.data(); inptr = (char *)inbuf; - outbuf = new char[insize+1]; - outptr = outbuf; + outbuf = new char[bufsize]; + size_t lastError = -1; - iconv(cd, &inptr, &insize, &outptr, &outsize); - *outptr = '\0'; + while (insize != 0) + { + char *outptr = outbuf; + size_t outsize = bufsize; + size_t const err = iconv(cd, &inptr, &insize, &outptr, &outsize); + dest->append(outbuf, outptr - outbuf); + if (err == (size_t)(-1)) + { + switch (errno) + { + case EILSEQ: + insize--; + inptr++; + // replace a series of unknown multibytes with a single "?" + if (lastError != insize) { + lastError = insize - 1; + dest->append("?"); + } + break; + case EINVAL: + insize = 0; + break; + case E2BIG: + if (outptr == outbuf) + { + bufsize *= 2; + delete[] outbuf; + outbuf = new char[bufsize]; + } + break; + } + } + } - *dest = outbuf; delete[] outbuf; iconv_close(cd); @@ -972,6 +1000,23 @@ bool TokSplitString(char Tok,char *Input,char **List, return true; } /*}}}*/ +// ExplodeString - Split a string up into a vector /*{{{*/ +// --------------------------------------------------------------------- +/* This can be used to split a given string up into a vector, so the + propose is the same as in the method above and this one is a bit slower + also, but the advantage is that we an iteratable vector */ +vector ExplodeString(string const &haystack, char const &split) { + string::const_iterator start = haystack.begin(); + string::const_iterator end = start; + vector exploded; + do { + for (; end != haystack.end() && *end != split; ++end); + exploded.push_back(string(start, end)); + start = end; + } while (end != haystack.end() && (++end) != haystack.end()); + return exploded; +} + /*}}}*/ // RegexChoice - Simple regex list/list matcher /*{{{*/ // --------------------------------------------------------------------- /* */