feedc0de/boost_detail
1
0
Fork 0
forked from boostorg/detail
Code Pull Requests Activity

Fixed inconsistent linkage specification on MSVC. Code cleanup.

Browse Source 
cc32906071 introduced inconsistent
BOOST_UTF8_DECL markup between function declarations and definitions,
which caused MSVC build errors. This commit fixes this.

Also, replaced unneeded BOOST_UTF8_DECL markup with inline for functions
that are only used internally. Made get_cont_octet_out_count static.

Additionally, removed duplicate octet1_modifier_table definitions
and moved it to a function-local static.

Whitespace and formatting cleanup.

Closes https://github.com/boostorg/filesystem/issues/204.
This commit is contained in:
Andrey Semashev
2021-08-18 00:39:45 +03:00
parent db4c8248c0
commit 87489a4346
2 changed files with 108 additions and 94 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
include/boost/detail/utf8_codecvt_facet.hpp
View File

@ -113,24 +113,25 @@ struct BOOST_SYMBOL_VISIBLE utf8_codecvt_facet :
public std::codecvt<wchar_t, char, std::mbstate_t>
{
public:
explicit utf8_codecvt_facet(std::size_t no_locale_manage=0);
virtual ~utf8_codecvt_facet();
BOOST_UTF8_DECL explicit utf8_codecvt_facet(std::size_t no_locale_manage = 0);
BOOST_UTF8_DECL virtual ~utf8_codecvt_facet();
protected:
virtual std::codecvt_base::result do_in(
BOOST_UTF8_DECL virtual std::codecvt_base::result do_in(
std::mbstate_t& state,
const char * from,
const char * from_end,
const char * & from_next,
wchar_t * to,
wchar_t * to_end,
wchar_t*& to_next
wchar_t * & to_next
) const;
virtual std::codecvt_base::result do_out(
BOOST_UTF8_DECL virtual std::codecvt_base::result do_out(
std::mbstate_t & state,
const wchar_t * from,
const wchar_t * from_end,
const wchar_t* & from_next,
const wchar_t * & from_next,
char * to,
char * to_end,
char * & to_next
@ -150,11 +151,11 @@ protected:
return get_octet_count(lead_octet) - 1;
}
static unsigned int get_octet_count(unsigned char lead_octet);
BOOST_UTF8_DECL static unsigned int get_octet_count(unsigned char lead_octet);
// How many "continuing octets" will be needed for this word
// == total octets - 1.
int get_cont_octet_out_count(wchar_t word) const ;
BOOST_UTF8_DECL static int get_cont_octet_out_count(wchar_t word);
virtual bool do_always_noconv() const BOOST_NOEXCEPT_OR_NOTHROW {
return false;
@ -162,7 +163,7 @@ protected:
// UTF-8 isn't really stateful since we rewind on partial conversions
virtual std::codecvt_base::result do_unshift(
std::mbstate_t&,
std::mbstate_t &,
char * from,
char * /*to*/,
char * & next
@ -178,7 +179,7 @@ protected:
// How many char objects can I process to get <= max_limit
// wchar_t objects?
virtual int do_length(
BOOST_UTF8_DECL virtual int do_length(
std::mbstate_t &,
const char * from,
const char * from_end,

59
include/boost/detail/utf8_codecvt_facet.ipp
View File

@ -30,6 +30,19 @@ BOOST_UTF8_BEGIN_NAMESPACE
/////////1/////////2/////////3/////////4/////////5/////////6/////////7/////////8
// implementation for wchar_t
namespace detail {
inline const wchar_t * get_octet1_modifier_table() BOOST_NOEXCEPT
{
static const wchar_t octet1_modifier_table[] = {
0x00, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc
};
return octet1_modifier_table;
}
} // namespace detail
BOOST_UTF8_DECL utf8_codecvt_facet::utf8_codecvt_facet(
std::size_t no_locale_manage
) :
@ -57,10 +70,11 @@ BOOST_UTF8_DECL std::codecvt_base::result utf8_codecvt_facet::do_in(
// then mash the whole lot together. Note that each continuing
// octet only uses 6 bits as unique values, so only shift by
// multiples of 6 to combine.
const wchar_t * const octet1_modifier_table = detail::get_octet1_modifier_table();
while (from != from_end && to != to_end) {
// Error checking on the first octet
if (invalid_leading_octet(*from)){
if (invalid_leading_octet(*from)) {
from_next = from;
to_next = to;
return std::codecvt_base::error;
@ -69,21 +83,18 @@ BOOST_UTF8_DECL std::codecvt_base::result utf8_codecvt_facet::do_in(
// The first octet is adjusted by a value dependent upon
// the number of "continuing octets" encoding the character
const int cont_octet_count = get_cont_octet_count(*from);
const wchar_t octet1_modifier_table[] = {
0x00, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc
};
// The unsigned char conversion is necessary in case char is
// signed (I learned this the hard way)
wchar_t ucs_result =
(unsigned char)(*from++) - octet1_modifier_table[cont_octet_count];
// Invariants :
// Invariants:
// 1) At the start of the loop, 'i' continuing characters have been
// processed
// 2) *from points to the next continuing character to be processed.
int i = 0;
while(i != cont_octet_count && from != from_end) {
while (i != cont_octet_count && from != from_end) {
// Error checking on continuing characters
if (invalid_continuing_octet(*from)) {
@ -103,7 +114,7 @@ BOOST_UTF8_DECL std::codecvt_base::result utf8_codecvt_facet::do_in(
// If the buffer ends with an incomplete unicode character...
if (from == from_end && i != cont_octet_count) {
// rewind "from" to before the current character translation
from_next = from - (i+1);
from_next = from - (i + 1);
to_next = to;
return std::codecvt_base::partial;
}
@ -113,8 +124,10 @@ BOOST_UTF8_DECL std::codecvt_base::result utf8_codecvt_facet::do_in(
to_next = to;
// Were we done converting or did we run out of destination space?
if(from == from_end) return std::codecvt_base::ok;
else return std::codecvt_base::partial;
if (from == from_end)
return std::codecvt_base::ok;
else
return std::codecvt_base::partial;
}
BOOST_UTF8_DECL std::codecvt_base::result utf8_codecvt_facet::do_out(
@ -127,11 +140,7 @@ BOOST_UTF8_DECL std::codecvt_base::result utf8_codecvt_facet::do_out(
char * & to_next
) const
{
// RG - consider merging this table with the other one
const wchar_t octet1_modifier_table[] = {
0x00, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc
};
const wchar_t * const octet1_modifier_table = detail::get_octet1_modifier_table();
wchar_t max_wchar = (std::numeric_limits<wchar_t>::max)();
while (from != from_end && to != to_end) {
@ -145,7 +154,7 @@ BOOST_UTF8_DECL std::codecvt_base::result utf8_codecvt_facet::do_out(
int cont_octet_count = get_cont_octet_out_count(*from);
// RG - comment this formula better
int shift_exponent = (cont_octet_count) * 6;
int shift_exponent = cont_octet_count * 6;
// Process the first character
*to++ = static_cast<char>(octet1_modifier_table[cont_octet_count] +
@ -163,18 +172,21 @@ BOOST_UTF8_DECL std::codecvt_base::result utf8_codecvt_facet::do_out(
++i;
}
// If we filled up the out buffer before encoding the character
if(to == to_end && i != cont_octet_count) {
if (to == to_end && i != cont_octet_count) {
from_next = from;
to_next = to - (i+1);
to_next = to - (i + 1);
return std::codecvt_base::partial;
}
++from;
}
from_next = from;
to_next = to;
// Were we done or did we run out of destination space
if(from == from_end) return std::codecvt_base::ok;
else return std::codecvt_base::partial;
if (from == from_end)
return std::codecvt_base::ok;
else
return std::codecvt_base::partial;
}
// How many char objects can I process to get <= max_limit
@ -203,7 +215,7 @@ BOOST_UTF8_DECL int utf8_codecvt_facet::do_length(
BOOST_UTF8_DECL unsigned int utf8_codecvt_facet::get_octet_count(
unsigned char lead_octet
){
) {
// if the 0-bit (MSB) is 0, then 1 character
if (lead_octet <= 0x7f) return 1;
@ -220,7 +232,7 @@ BOOST_UTF8_DECL unsigned int utf8_codecvt_facet::get_octet_count(
namespace detail {
template<std::size_t s>
BOOST_UTF8_DECL int get_cont_octet_out_count_impl(wchar_t word){
inline int get_cont_octet_out_count_impl(wchar_t word) {
if (word < 0x80) {
return 0;
}
@ -231,7 +243,7 @@ BOOST_UTF8_DECL int get_cont_octet_out_count_impl(wchar_t word){
}
template<>
BOOST_UTF8_DECL int get_cont_octet_out_count_impl<4>(wchar_t word){
inline int get_cont_octet_out_count_impl<4>(wchar_t word) {
if (word < 0x80) {
return 0;
}
@ -275,9 +287,10 @@ BOOST_UTF8_DECL int get_cont_octet_out_count_impl<4>(wchar_t word){
// == total octets - 1.
BOOST_UTF8_DECL int utf8_codecvt_facet::get_cont_octet_out_count(
wchar_t word
) const {
) {
return detail::get_cont_octet_out_count_impl<sizeof(wchar_t)>(word);
}
BOOST_UTF8_END_NAMESPACE
#endif