2009-01-25 07:23:35 Daniel Shahaf napisał(a):
> > @@ -1074,6 +1075,14 @@ output_unified_diff_modified(void *baton
> > {
> > output_baton->hunk_extra_context[--p] = '\0';
> > }
> > + const char *invalid_character =
> > + svn_utf__last_valid(output_baton->hunk_extra_context,
> > + SVN_DIFF__EXTRA_CONTEXT_LENGTH - 1);
>
> s/SVN_DIFF__EXTRA_CONTEXT_LENGTH - 1/SVN_DIFF__EXTRA_CONTEXT_LENGTH + 1/ ?
15 lines earlier there is:
/* Save the extra context for later use.
* Note that the last byte of the hunk_extra_context array is never
* touched after it is zero-initialized, so the array is always
* 0-terminated. */
strncpy(output_baton->hunk_extra_context,
output_baton->extra_context->data,
SVN_DIFF__EXTRA_CONTEXT_LENGTH);
svn_utf__last_valid()'s doc string contains:
> /* Return a pointer to the first character after the last valid UTF-8
> * multi-byte character in the string SRC of length LEN. If SRC is a valid
> * UTF-8 the return value will point to the byte after SRC+LEN, otherwise
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> * it will point to the start of the first invalid multi-byte character.
> * In either case all the characters between SRC and the return pointer are
> * valid UTF-8.
So IMHO current code is correct.
--
Arfrever Frehtes Taifersar Arahesis
Received on 2009-01-25 16:31:17 CET