Line data    Source code

       1             : /* File: draw_line_breaker.inl; Copyright and License: see below */
       2             : 
       3             : #include "utf8stringbuf/utf8codepointiterator.h"
       4             : #include <assert.h>
       5             : 
       6         256 : static inline void draw_line_breaker_init( draw_line_breaker_t *this_ )
       7             : {
       8         256 : }
       9             : 
      10         256 : static inline void draw_line_breaker_destroy( draw_line_breaker_t *this_ )
      11             : {
      12         256 : }
      13             : 
      14         256 : u8_error_t draw_line_breaker_append( const draw_line_breaker_t *this_,
      15             :                                      const utf8stringview_t *in_text,
      16             :                                      utf8stream_writer_t *out_text )
      17             : {
      18         256 :     u8_error_t err = U8_ERROR_NONE;
      19             :     utf8codepointiterator_t iter;
      20         256 :     utf8codepointiterator_init( &iter, in_text );
      21             : 
      22         256 :     const utf8codepoint_t space0width = utf8codepoint( 0x200B );
      23         256 :     uint32_t last = ' ';
      24        4096 :     for(; utf8codepointiterator_has_next( &iter ); )
      25             :     {
      26        3840 :         utf8codepoint_t codepnt = utf8codepointiterator_next( &iter );
      27        3840 :         uint32_t current = utf8codepoint_get_char ( codepnt );
      28        3840 :         const bool no_space = ( last != ' ' )&&( current != ' ' );
      29        3840 :         if ( no_space )
      30             :         {
      31        3072 :             const bool last_AZ = ( last >= 'A' )&&( last <= 'Z' );
      32        3072 :             const bool last_az = ( last >= 'a' )&&( last <= 'z' );
      33        3072 :             const bool last_09 = ( last >= '0' )&&( last <= '9' );
      34        3072 :             const bool last_other = ( last > '\x7f' );
      35        3072 :             const bool last_stick_right
      36        3072 :                 = ( last == '(' )||( last == '[' )||( last == '{' )||( last == '\x60' );
      37        3072 :             const bool cur_AZ = ( current >= 'A' )&&( current <= 'Z' );
      38        3072 :             const bool cur_az = ( current >= 'a' )&&( current <= 'z' );
      39        3072 :             const bool cur_09 = ( current >= '0' )&&( current <= '9' );
      40        3072 :             const bool cur_other = ( current > '\x7f' );
      41        3072 :             const bool cur_stick_left
      42        3072 :                 = ( last == ')' )||( last == ']' )||( last == '}' )||( last == ',' )||( last == ';' );
      43             : 
      44        3072 :             const bool same_token
      45           0 :                 = ( last_AZ && ( cur_AZ || cur_az || cur_09 ) )
      46        3072 :                 || ( last_az && ( cur_az || cur_09 ) )
      47           0 :                 || ( last_09 && cur_09 )
      48        6144 :                 || ( last_other && cur_other );
      49        3072 :             const bool no_break
      50           0 :                 = ( ( last == '.' ) && cur_09 )  /* do not split numbers with dots */
      51        3072 :                 || ( last_09 && ( current == '.' ) )  /* do not split numbers with dots */
      52        3072 :                 || ( ( last == '-' ) && cur_09 )  /* do not split negative numbers */
      53        3072 :                 || ( ( last == '+' ) && cur_09 )  /* do not split positive numbers */
      54        8960 :                 || ( ( last == current )  /* do not split pairs, e.g. ++, --, ==, &&, ...*/
      55        2816 :                 || ( ( last == '!' ) && ( current == '=' ) )
      56        2816 :                 || last_stick_right
      57        2816 :                 || cur_stick_left );  /* even if not the same token, here no break */
      58             : 
      59        3072 :             if (( ! same_token )&&( ! no_break ))
      60             :             {
      61           0 :                 err |= utf8stream_writer_write_char( out_text, utf8codepoint_get_char( space0width ) );
      62             :             }
      63             :         }
      64        3840 :         err |= utf8stream_writer_write_char( out_text, utf8codepoint_get_char( codepnt ) );
      65        3840 :         last = current;
      66             :     }
      67             : 
      68         256 :     utf8codepointiterator_destroy( &iter );
      69         256 :     return err;
      70             : }
      71             : 
      72             : 
      73             : /*
      74             : Copyright 2024-2024 Andreas Warnke
      75             : 
      76             : Licensed under the Apache License, Version 2.0 (the "License");
      77             : you may not use this file except in compliance with the License.
      78             : You may obtain a copy of the License at
      79             : 
      80             :     http://www.apache.org/licenses/LICENSE-2.0
      81             : 
      82             : Unless required by applicable law or agreed to in writing, software
      83             : distributed under the License is distributed on an "AS IS" BASIS,
      84             : WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
      85             : See the License for the specific language governing permissions and
      86             : limitations under the License.
      87             : */