DataMuseum.dk

top - metrics - download

    Length: 1709 (0x6ad)
    Types: TextFile
    Names: »Xlocale.doc«

└─⟦276d19d6e⟧ Bits:30007243 EUUGD5_I: X11R5
    └─⟦4856bf7e7⟧ »./mit-4/mit-4.00« 
        └─⟦635ff9e7e⟧ 
            └─⟦this⟧ »mit/doc/I18N/Xsi/Xlc/Xlocale.doc« 

TextFile

$Header: Xlocale.doc,v 1.1 90/12/27 20:21:01 morisaki Exp $
$Date: 90/12/27 20:21:01 $


Explain of Data Structure:

1. Conversion 

   The conversion is done by the table-driven method, and also can
be done by procedure-driven method which is registered by application.
The later method is not supported in this implementation.  The value
of cds_cnvproc is always set to NULL pointer.

   The table-driven method is defined with:

    int     cds_cnvindex[MAXCHARSETS]; /* table-driven conversion */
    Range   *cds_cnvlist;              /* conversion list */

The converstion list cds_cnvlist stores all pairs of starting and ending
codepoint of all charsets.  The all of pairs of each charset, say csid,
is indexed from cds_cnvindex[csid] to cds_cnvindex[csid+1]-1.  A pair
is made of 4 elements which specify the range of conversion:
 
    unsigned int  mb_start;    /* starting code point of multibyte*/
    unsigned int  mb_end;      /* ending code point of multibyte  */
    unsigned int  cs_start;    /* starting code point of charset */
    unsigned int  cs_end;      /* ending code point of charset  */

The range (mb_start, mb_end) of multibyte encoding is mapped to the range 
(cs_start, cs_end) of a charset with 1-to-1.

1. State-dependent & State-independent Codeset

   The differency between the 2 codesets is mapping a character to
the correspondent charset id.
   For state-independent codeset:
      use mapping table.
   For state-dependent codeset:
      compare designation sequence.

2. X Locale

    struct {
      ...
      char *  *lc_im_name;
      char *  *lc_im;
      ...
   } XLocale;

   The lc_im will be malloc'ed to store the modifier from
XSetLocaleModifers("im=value");