Re: bug: fuzzystrmatch levenshtein is wrong

On Mon, Dec 7, 2009 at 8:33 AM, marcin mank <marcin(dot)mank(at)gmail(dot)com> wrote:
> The current behavior of levenshtein(text,text,int,int,int) is wrong. Consider:
>
> leki_dev=# select levenshtein('','a',2,4,5);
>  levenshtein
> -------------
>           1
> (1 row)
>
>
> leki_dev=# select levenshtein('a','',2,4,5);
>  levenshtein
> -------------
>           1
> (1 row)
>
>
> leki_dev=# select levenshtein('aa','a',2,4,5);
>  levenshtein
> -------------
>           1
> (1 row)
>
>
> leki_dev=# select levenshtein('a','aa',2,4,5);
>  levenshtein
> -------------
>           1
> (1 row)
>
> versus (after patch)
>
> postgres=# select levenshtein('','a',2,4,5);
>  levenshtein
> -------------
>           2
> (1 row)
>
> postgres=# select levenshtein('a','',2,4,5);
>  levenshtein
> -------------
>           4
> (1 row)
>
> postgres=# select levenshtein('aa','a',2,4,5);
>  levenshtein
> -------------
>           4
> (1 row)
>
> postgres=# select levenshtein('a','aa',2,4,5);
>  levenshtein
> -------------
>           2
> (1 row)
>
> patch attached.

I cannot get this patch to apply for anything. All 4 hunks fail, both
on HEAD and on the the pre-8.4-pgindent version of fuzzystrmatch.c.
Either I'm doing something wrong here, or there's something wrong with
this patch file.

...Robert