database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-09-08 00:47:37 +03:00
Code Activity

Add wchar -> mb conversion routines.

Browse Source 
This is infrastructure for Alexander Korotkov's work on indexing regular
expression searches.

Alexander Korotkov, with a bit of further hackery on the MULE conversion
by me
This commit is contained in:
Robert Haas
2012-07-04 17:10:10 -04:00
parent 248b5fce06
commit 72dd6291f2
3 changed files with 248 additions and 45 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
src/include/mb/pg_wchar.h
View File

@@ -140,6 +140,10 @@ typedef unsigned int pg_wchar;
#define LCPRV1_A 0x9a
#define LCPRV1_B 0x9b
#define IS_LCPRV1(c) ((unsigned char)(c) == LCPRV1_A || (unsigned char)(c) == LCPRV1_B)
#define IS_LCPRV1_A_RANGE(c) \
((unsigned char)(c) >= 0xa0 && (unsigned char)(c) <= 0xdf)
#define IS_LCPRV1_B_RANGE(c) \
((unsigned char)(c) >= 0xe0 && (unsigned char)(c) <= 0xef)
/*
* Postgres-specific prefix bytes for "private" multibyte encodings
@@ -148,6 +152,10 @@ typedef unsigned int pg_wchar;
#define LCPRV2_A 0x9c
#define LCPRV2_B 0x9d
#define IS_LCPRV2(c) ((unsigned char)(c) == LCPRV2_A || (unsigned char)(c) == LCPRV2_B)
#define IS_LCPRV2_A_RANGE(c) \
((unsigned char)(c) >= 0xf0 && (unsigned char)(c) <= 0xf4)
#define IS_LCPRV2_B_RANGE(c) \
((unsigned char)(c) >= 0xf5 && (unsigned char)(c) <= 0xfe)
/*
* Charset IDs for private single byte encodings (0xa0-0xef)
@@ -324,7 +332,11 @@ extern pg_enc2gettext pg_enc2gettext_tbl[];
* pg_wchar stuff
*/
typedef int (*mb2wchar_with_len_converter) (const unsigned char *from,
pg_wchar *to,
pg_wchar *to,
int len);
typedef int (*wchar2mb_with_len_converter) (const pg_wchar *from,
unsigned char *to,
int len);
typedef int (*mblen_converter) (const unsigned char *mbstr);
@@ -337,8 +349,10 @@ typedef int (*mbverifier) (const unsigned char *mbstr, int len);
typedef struct
{
mb2wchar_with_len_converter mb2wchar_with_len; /* convert a multibyte
* string to a wchar */
mb2wchar_with_len_converter mb2wchar_with_len; /* convert a multibyte
* string to a wchar */
wchar2mb_with_len_converter wchar2mb_with_len; /* convert a wchar
* string to a multibyte */
mblen_converter mblen; /* get byte length of a char */
mbdisplaylen_converter dsplen; /* get display width of a char */
mbverifier mbverify; /* verify multibyte sequence */
@@ -419,6 +433,10 @@ extern int pg_mb2wchar(const char *from, pg_wchar *to);
extern int pg_mb2wchar_with_len(const char *from, pg_wchar *to, int len);
extern int pg_encoding_mb2wchar_with_len(int encoding,
const char *from, pg_wchar *to, int len);
extern int pg_wchar2mb(const pg_wchar *from, char *to);
extern int pg_wchar2mb_with_len(const pg_wchar *from, char *to, int len);
extern int pg_encoding_wchar2mb_with_len(int encoding,
const pg_wchar *from, char *to, int len);
extern int pg_char_and_wchar_strcmp(const char *s1, const pg_wchar *s2);
extern int pg_wchar_strncmp(const pg_wchar *s1, const pg_wchar *s2, size_t n);
extern int pg_char_and_wchar_strncmp(const char *s1, const pg_wchar *s2, size_t n);