Re: Duplicate-key-detection failure case found in btree - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Duplicate-key-detection failure case found in btree
Date
Msg-id 4366.1009987774@sss.pgh.pa.us
Whole thread Raw
In response to Re: Duplicate-key-detection failure case found in btree  (Jan Wieck <janwieck@yahoo.com>)
List pgsql-hackers
Jan Wieck <janwieck@yahoo.com> writes:
>     It  always scares me that bugs like this can live untriggered
>     for multiple releases.  What  else  is  lingering  under  the
>     surface ...

FWIW, I don't think that this bug did survive for multiple releases
(though it very nearly made it into 7.2).  The present form of
bt_check_unique and bt_insertonpg was new code in 7.1, replacing
the BTP_CHAIN method of handling identical keys.  That code had a
lot of bugs of its own, but I don't think it had this one.  I will
take the blame for introducing this bug into 7.1 ...
        regards, tom lane


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: problems with new vacuum (??)
Next
From: Tom Lane
Date:
Subject: Re: datetime error?