[
https://issues.apache.org/jira/browse/MAILBOX-228?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Kohei Nozaki updated MAILBOX-228:
---------------------------------
Attachment: MAILBOX-228.patch
The problem is that StoreMessageResultIterator assumes underlying dataset is
always ordered by the uid, while RDBMS doesn't guarantee it (at least with
PostgreSQL).
This patch adds {{ORDER BY message.uid ASC}} to all of JPQLs that for
retrieving Message entity.
> StoreMessageResultIterator returns only a part of messages with JPA backend
> on some databases
> ---------------------------------------------------------------------------------------------
>
> Key: MAILBOX-228
> URL: https://issues.apache.org/jira/browse/MAILBOX-228
> Project: James Mailbox
> Issue Type: Bug
> Components: jpa
> Affects Versions: 0.5
> Environment: James3 which acquired from SVN trunk on Feburuary 2015
> PostgreSQL 9.2.6
> Thunderbird 31.6.0
> Reporter: Kohei Nozaki
> Attachments: MAILBOX-228.patch
>
>
> Consider you have following dataset.
> Query:
> {noformat}
> SELECT t0.mailbox_id, t0.mail_uid, t0.mail_is_answered,
> t0.mail_body_start_octet, t0.mail_content_octets_count, t0.mail_is_deleted,
> t0.mail_is_draft, t0.mail_is_flagged, t0.mail_date, t0.mail_mime_type,
> t0.mail_modseq, t0.mail_is_recent, t0.mail_is_seen, t0.mail_mime_subtype,
> t0.mail_textual_line_count FROM public.james_mail t0 WHERE (t0.mailbox_id = 1
> AND t0.mail_uid >= 1 AND t0.mail_uid <= 18) LIMIT 200
> {noformat}
> Result against the preceding query:
> {noformat}
> mailbox_id mail_uid mail_is_answered mail_body_start_octet
> mail_content_octets_count mail_is_deleted mail_is_draft
> mail_is_flagged mail_date mail_mime_type mail_modseq
> mail_is_recent mail_is_seen mail_mime_subtype
> mail_textual_line_count
> 1 1 false 1450 1464 false false false "2015-04-25
> 08:28:40.920000" text 4 false true plain 1
> 1 2 false 3536 4237 false false false "2015-04-25
> 09:56:13.350000" text 8 false true plain 23
> 1 3 false 4574 9174 false false false "2015-04-25
> 17:26:48.450000" multipart 12 false true alternative
> NULL
> 1 4 false 2901 3646 false false false "2015-04-26
> 06:28:19.710000" multipart 16 false true alternative
> NULL
> 1 6 false 1603 2645 false false false "2015-04-26
> 12:06:53.350000" text 22 false true plain 31
> 1 5 false 1603 2264 false false false "2015-04-26
> 12:06:53.350000" text 21 false true plain 22
> 1 17 false 1612 2509 false false false "2015-04-27
> 03:52:02.870000" text 64 false true plain 21
> 1 18 false 1605 2200 false false false "2015-04-27
> 04:14:57.180000" text 62 false true plain 20
> 1 7 false 1646 2572 false false false "2015-04-26
> 13:52:56.720000" text 41 false true plain 28
> 1 8 false 1645 2958 false false false "2015-04-26
> 13:58:01.320000" text 42 false true plain 34
> 1 9 false 1645 3093 false false false "2015-04-26
> 14:07:01.740000" text 43 false true plain 36
> 1 10 false 1603 2428 false false false "2015-04-26
> 14:25:03.150000" text 44 false true plain 27
> 1 11 false 1603 2483 false false false "2015-04-26
> 14:25:51.300000" text 45 false true plain 31
> 1 12 false 1603 2428 false false false "2015-04-26
> 14:31:57.130000" text 46 false true plain 27
> 1 13 false 1603 2459 false false false "2015-04-26
> 14:32:51.110000" text 47 false true plain 29
> 1 14 false 1604 2471 false false false "2015-04-26
> 14:49:51.700000" text 52 false true plain 29
> 1 15 false 1604 2490 false false false "2015-04-26
> 14:49:51.730000" text 53 false true plain 31
> 1 16 false 1646 3082 false false false "2015-04-26
> 18:24:56.440000" text 57 false true plain 38
> {noformat}
> Note that the dataset doesn't ordered by mail_uid. with that dataset, the
> IMAP server returns following response:
> {noformat}
> 3 uid fetch 1:* (FLAGS)
> * 1 FETCH (FLAGS (\Seen NonJunk) UID 1)
> * 2 FETCH (FLAGS (\Seen NonJunk) UID 2)
> * 3 FETCH (FLAGS (\Seen NonJunk) UID 3)
> * 4 FETCH (FLAGS (\Seen NonJunk) UID 4)
> * 5 FETCH (FLAGS (\Seen NonJunk) UID 6)
> * 6 FETCH (FLAGS (\Seen NonJunk) UID 5)
> * 7 FETCH (FLAGS (\Seen) UID 17)
> * 8 FETCH (FLAGS (\Seen NonJunk) UID 18)
> 3 OK FETCH completed.
> {noformat}
> I believe the server should return all of dataset, but records that has
> mail_uid larger than 18 were skipped.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: server-dev-unsubscr...@james.apache.org
For additional commands, e-mail: server-dev-h...@james.apache.org
