Amebis/ZRCola
Amebis/ZRCola
1
1
Fork 0
Code Issues 2 Pull Requests Actions Packages Projects Releases 49 Wiki Activity

Accessing translation data standardized

Browse Source
This commit is contained in:
Simon Rozman 2016-03-02 10:45:22 +01:00
parent c44b449a36
commit 4e5811803a
3 changed files with 24 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
ZRColaCompile/dbsource.h
View File

@ -36,8 +36,8 @@ namespace ZRCola {
///
class translation {
public:
wchar_t chr; ///< composed character
std::wstring str; ///< decomposed string
wchar_t chr; ///< Composed character
std::wstring str; ///< Decomposed string
};
public:

26
lib/libZRCola/include/zrcola/translate.h
View File

@ -45,16 +45,6 @@ namespace ZRCola {
unsigned __int32 start; ///< Composed character offset
unsigned __int32 end; ///< Decomposed string end offset
///
/// Returns translation string start offset
///
inline unsigned __int32 GetChrStart() const { return start; }
///
/// Returns translation string start offset
///
inline unsigned __int32 GetStrStart() const { return start + 1; }
///
/// Returns translation string length
///
@ -62,9 +52,23 @@ namespace ZRCola {
};
#pragma pack(pop)
#pragma pack(push)
#pragma pack(2)
#pragma warning(push)
#pragma warning(disable: 4200)
///
/// Translation data
///
struct translation {
wchar_t chr; ///< Composed character
wchar_t str[]; ///< Decomposed string
};
#pragma warning(pop)
#pragma pack(pop)
std::vector<index> idxComp; ///< Composition index
std::vector<index> idxDecomp; ///< Decomposition index
std::vector<wchar_t> data; ///< Transformation data
std::vector<wchar_t> data; ///< Transformation data
public:
///

14
lib/libZRCola/src/translate.cpp
View File

@ -49,7 +49,7 @@ void ZRCola::translation_db::Compose(_In_z_count_(inputMax) const wchar_t* input
// Get the j-th character of the composition.
// All compositions that get short on characters are lexically ordered before.
// Thus the j-th character is considered 0.
wchar_t s = j < idxComp[m].GetStrLength() ? data[idxComp[m].GetStrStart() + j] : 0;
wchar_t s = j < idxComp[m].GetStrLength() ? ((translation*)&data[idxComp[m].start])->str[j] : 0;
// Do the bisection test.
if (c < s) r = m;
@ -60,14 +60,14 @@ void ZRCola::translation_db::Compose(_In_z_count_(inputMax) const wchar_t* input
// Narrow the search area on the left to start at the first composition in the run.
for (size_t rr = m; l < rr;) {
size_t m = (l + rr) / 2;
wchar_t s = j < idxComp[m].GetStrLength() ? data[idxComp[m].GetStrStart() + j] : 0;
wchar_t s = j < idxComp[m].GetStrLength() ? ((translation*)&data[idxComp[m].start])->str[j] : 0;
if (c <= s) rr = m; else l = m + 1;
}
// Narrow the search area on the right to end at the first composition not in the run.
for (size_t ll = m + 1; ll < r;) {
size_t m = (ll + r) / 2;
wchar_t s = j < idxComp[m].GetStrLength() ? data[idxComp[m].GetStrStart() + j] : 0;
wchar_t s = j < idxComp[m].GetStrLength() ? ((translation*)&data[idxComp[m].start])->str[j] : 0;
if (s <= c) ll = m + 1; else r = m;
}
@ -79,7 +79,7 @@ void ZRCola::translation_db::Compose(_In_z_count_(inputMax) const wchar_t* input
// The search area is empty.
if (j && l_prev < compositionsCount && j == idxComp[l_prev].GetStrLength()) {
// The first composition of the previous run was a match.
output += data[idxComp[l_prev].GetChrStart()];
output += ((translation*)&data[idxComp[l_prev].start])->chr;
i = ii;
if (j > 1 && map) {
// Mapping changed.
@ -97,7 +97,7 @@ void ZRCola::translation_db::Compose(_In_z_count_(inputMax) const wchar_t* input
if (l < compositionsCount && j == idxComp[l].GetStrLength()) {
// The first composition of the previous run was a match.
output += data[idxComp[l].GetChrStart()];
output += ((translation*)&data[idxComp[l].start])->chr;
i = ii;
if (j > 1 && map) {
// Mapping changed.
@ -138,12 +138,12 @@ void ZRCOLA_API ZRCola::translation_db::Decompose(_In_z_count_(inputMax) const w
for (size_t l = 0, r = decompositionsCount;; ) {
if (l < r) {
size_t m = (l + r) / 2;
wchar_t decompSrc = data[idxDecomp[m].GetChrStart()];
wchar_t decompSrc = ((translation*)&data[idxDecomp[m].start])->chr;
if (c < decompSrc) r = m;
else if (decompSrc < c) l = m + 1;
else {
// Character found.
output.append(&data[idxDecomp[m].GetStrStart()], idxDecomp[m].GetStrLength());
output.append(((translation*)&data[idxDecomp[m].start])->str, idxDecomp[m].GetStrLength());
i++;
if (map) {
// Mapping changed.