[
https://issues.apache.org/jira/browse/CONNECTORS-1521?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16571557#comment-16571557
]
James Thomas commented on CONNECTORS-1521:
------------------------------------------
(I am not a developer but ...) I have explored the Javadoc for
com.documentum.fc.common, Interface IDfTime in DFC 16.4
There are ~50 patterns, here's a couple of examples:
{noformat}
DF_TIME_PATTERN8 Represents time pattern string "month dd, yyyy".
DF_TIME_PATTERN34 Represents time pattern string "mon-yyyy hh:mi:ss".
{noformat}
As far as I can see none of the patterns gives anything related to time zone or
locale, or mentions UTC, DST etc.
This is consistent with the blog post that you found.
Googling around a little, I came across these:
[https://stackoverflow.com/a/8315831] which looks related to your comment about
asString().
[https://msroth.wordpress.com/2011/05/15/the-content-server-and-time-zones|https://msroth.wordpress.com/2011/05/15/the-content-server-and-time-zones,m]
and [https://blog.documentum.pro/2015/02/04/time-in-documentum] which are in
this general area.
> Documentum Connector users ManifoldCF's local time in queries constraints
> against the Documentum server without reference to time zones
> ---------------------------------------------------------------------------------------------------------------------------------------
>
> Key: CONNECTORS-1521
> URL: https://issues.apache.org/jira/browse/CONNECTORS-1521
> Project: ManifoldCF
> Issue Type: Bug
> Components: Documentum connector
> Affects Versions: ManifoldCF 2.10
> Reporter: James Thomas
> Assignee: Karl Wright
> Priority: Major
>
> I find that the time/date constraints in queries to the Documentum server are
> based on the "raw" local time of the ManifoldCF server but appear to take no
> account of the time zones of the two servers.
> This can lead to recently modified files not being transferred to the output
> repository when you would naturally expect them to be. I'd like the times to
> be aligned, perhaps by including time zone in the query. In particular, is
> there a way to use UTC perhaps?
> Here's an example ...
> * create a folder in Documentum
> * set up a job to point at the folder and output to the file system
> * put two documents into a folder in Documentum
> * Select them, right click and export as CSV (to show the timestamps):
> {noformat}
> 1.png,48489.0,Portable Network Graphics,8/7/2018 9:04 AM,
> 2.png,28620.0,Portable Network Graphics,8/7/2018 9:04 AM,,{noformat}
> Check the local time on the ManifoldCF server machine. Observe that it's
> reporting consistent time with the DM server:
> {noformat}
> [james@manifold]$ date
> Tue Aug 7 09:07:25 BST 2018{noformat}
> Start the job and look for the query to Documentum in the manifoldcf.log file
> (line break added for readability):
> {noformat}
> DEBUG 2018-08-07T08:07:47.297Z (Startup thread) - DCTM: About to execute
> query= (select for READ distinct i_chronicle_id from dm_document where
> r_modify_date >= date('01/01/1970 00:00:00','mm/dd/yyyy hh:mi:ss') and
> r_modify_date<=date('08/07/2018 08:07:34','mm/dd/yyyy hh:mi:ss')
> AND (i_is_deleted=TRUE Or (i_is_deleted=FALSE AND a_full_text=TRUE AND
> r_content_size>0)) AND ( Folder('/Administrator/james', DESCEND) ))
> ^C{noformat}
> Notice that the latest date asked for is *before* the modification date of
> the files added to DM. (And is an hour out, see footnote.)
>
> See whether anything has been output by the File System connector. It hasn't:
> {noformat}
> [james@manifold]$ ls /bigdisc/source/PDFs/timezones/
> [james@manifold]$
> {noformat}
> Now:
> * change the timezone on the ManifoldCF server machine
> * restart the ManifoldCF server and the Documentum processes
> * reseed the job
> Check the local time on the ManifoldCF server machine; it has changed:
> {noformat}
> [james@manifold]$ date
> Tue Aug 7 10:10:29 CEST 2018{noformat}
> Start the job again and notice that the query has changed by an hour, plus
> the few minutes it took to change the date etc (and is still an hour out, see
> footnote):
> {noformat}
> r_modify_date<=date('08/07/2018 09:11:02','mm/dd/yyyy hh:mi:ss')
> {noformat}
> Observe that the range of dates now covers the timestamps on the DM data, and
> also that some data has now been transferred by the File System connector:
> {noformat}
> [james@manifold]$ ls
> /bigdisc/source/PDFs/timezones/http/mfserver\:8080/da/component/
> drl?versionLabel=CURRENT&objectId=090000018000e515
> drl?versionLabel=CURRENT&objectId=090000018000e516
> {noformat}
>
>
> [Footnote] It appears that something is trying to take account of Daylight
> Saving Time too.
> If I set the server date to a time outside of DST, the query is aligned with
> the current time:
> {noformat}
> [i2e@i2ehost manifold]$ date
> Mon Oct 29 00:01:13 CET 2018
> r_modify_date<=date('10/29/2018 00:01:39','mm/dd/yyyy hh:mi:ss')
> {noformat}
> But if I set the time inside DST, the time is an hour before:
> {noformat}
> [i2e@i2ehost manifold]$ date
> Sat Oct 27 00:00:06 CEST 2018
> r_modify_date<=date('10/26/2018 23:00:26','mm/dd/yyyy hh:mi:ss')
> {noformat}
> This is perhaps a Java issue rather than a logic issue in the connector? See
> e.g. [https://stackoverflow.com/questions/6392/java-time-zone-is-messed-up]
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
