Lists: | pgsql-bugs |
---|
From: | Алексей <oktogen(at)mail(dot)ru> |
---|---|
To: | pgsql-bugs(at)postgresql(dot)org |
Subject: | Postgresql 8.4.2, Windows XP - temporary file leak: File still referenced |
Date: | 2009-12-22 07:27:27 |
Message-ID: | E1NMz9T-0003dN-00.oktogen-mail-ru@f291.mail.ru |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Lists: | pgsql-bugs |
windows or linux.
See test script
Attachment | Content-Type | Size |
---|---|---|
2.txt | text/plain | 5.0 KB |
1.txt | text/plain | 1.8 KB |
From: | Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com> |
---|---|
To: | Алексей <oktogen(at)mail(dot)ru> |
Cc: | pgsql-bugs(at)postgresql(dot)org |
Subject: | Re: Postgresql 8.4.2, Windows XP - temporary file leak: File still referenced |
Date: | 2009-12-22 11:40:16 |
Message-ID: | 4B30B020.7030402@enterprisedb.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Lists: | pgsql-bugs |
Алексей wrote:
> windows or linux.
> See test script
Hmm, looks like I broke this with the patch to fix the assertion failure
when a cursor is opeened in a subtransaction and the subtransaction is
then rolled back. Thanks for the report!
The bug is that when we put tuples to the tuplestore that holds the
tuples to be returned from the function, the current resource owner is
the subtransaction resource owner, which is closed at the end of
subtransaction. I'll fix that.
--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com
From: | Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com> |
---|---|
To: | Алексей <oktogen(at)mail(dot)ru> |
Cc: | pgsql-bugs(at)postgresql(dot)org |
Subject: | Re: Postgresql 8.4.2, Windows XP - temporary file leak: File still referenced |
Date: | 2009-12-29 17:44:58 |
Message-ID: | 4B3A401A.2080800@enterprisedb.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Lists: | pgsql-bugs |
Heikki Linnakangas wrote:
> Алексей wrote:
>> windows or linux.
>> See test script
>
> Hmm, looks like I broke this with the patch to fix the assertion failure
> when a cursor is opeened in a subtransaction and the subtransaction is
> then rolled back. Thanks for the report!
>
> The bug is that when we put tuples to the tuplestore that holds the
> tuples to be returned from the function, the current resource owner is
> the subtransaction resource owner, which is closed at the end of
> subtransaction. I'll fix that.
Fix committed now. Thanks for the report!
--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com