Changes between Initial Version and Version 1 of Ticket #3783, comment 9


Ignore:
Timestamp:
03/09/17 00:34:08 (3 years ago)
Author:
egmont
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #3783, comment 9

    initial v1  
    2121}}} 
    2222 
    23 which interprets the lone invalid UTF-8 byte as unsigned 169 and is assigned to the Unicode character, this is semantically the Latin-1 -> UTF-8 conversion. The rest goes on as if this was read as a valid 1-byte character. 
     23which interprets the lone invalid UTF-8 byte as unsigned 169 and is assigned to the Unicode character, this is semantically the Latin-1 -> Unicode conversion. The rest goes on as if this was read as a valid 1-byte character. 
    2424 
    2525The pattern of denoting and carrying invalid UTF-8 bytes as negative numbers is really weird to me. I'm inclined to say that the old design worked "accidentally". At least I don't recall ever thinking about it during my viewer rewrite.