Bug in serendipity_track_url()

[ads]

Just tracked done a bug in the serendipity_track_url() function:

In this function, a table entry is updated which the following conditions: scheme, host, port, path, query, day, entry_id. If affected_rows() is 0, then a new entry is inserted into the statistic table. This table has an unique constraint about three of the seven columns: host, day and entry_id.

So if i have 2 URLs in a blog entry, which have the same host but different pathnames, i get a SQL error about unique key violation, because it won't match the first update but will match the unique constraint.

Bug ID on SF is: 1604918

[ads]

I just realized: this bug is still open.

Either the primary key must be changed to include the "path" column or the select must be changed to not include the "path" column.

[garvinhicking]

Hi!

In which table? For the "exits"-Table the current primary ID is set to include "path" as well, maybe an DB update wasn't executed on your DB or maybe it was only added for mysql and not postgresql properly...?

Regards,
Garvin

[ads]

For the "exits" table, yes. Dunno what happened, probably too long ago.
Where can i find the database change, so i can try it manually?