Re: Timestamps and PostgreSQL
On 05/12/2010, at 8:08 PM, Paul Hoadley wrote: > EOF calls PreparedStatement.setTimestamp(int, Timestamp) (implemented in > AbstractJdbc2Statement), which in turn calls > PreparedStatement.setTimestamp(int, Timestamp, Calendar) with a null Calendar. Just for completeness, that's not quite true. Ramsey pointed out that JDBCColumn.takeInputValue() calls setObject() on that class. This is the top of the stack trace in setTimestamp(int, Timestamp, Calendar): at org.postgresql.jdbc2.AbstractJdbc2Statement.setTimestamp(AbstractJdbc2Statement.java:3049) at org.postgresql.jdbc2.AbstractJdbc2Statement.setTimestamp(AbstractJdbc2Statement.java:1334) at org.postgresql.jdbc2.AbstractJdbc2Statement.setObject(AbstractJdbc2Statement.java:1649) at org.postgresql.jdbc3.AbstractJdbc3Statement.setObject(AbstractJdbc3Statement.java:1438) at org.postgresql.jdbc2.AbstractJdbc2Statement.setObject(AbstractJdbc2Statement.java:1700) at com.webobjects.jdbcadaptor.JDBCColumn.takeInputValue(JDBCColumn.java:569) at com.webobjects.jdbcadaptor.JDBCChannel._bindInputVariablesWithBindingsAndExecute(JDBCChannel.java:260) at com.webobjects.jdbcadaptor.JDBCChannel._evaluateExpression(JDBCChannel.java:337) at com.webobjects.jdbcadaptor.JDBCChannel.evaluateExpression(JDBCChannel.java:296) at com.webobjects.jdbcadaptor.JDBCChannel.selectAttributes(JDBCChannel.java:220) at com.webobjects.eoaccess.EODatabaseChannel._selectWithFetchSpecificationEditingContext(EODatabaseChannel.java:897) at com.webobjects.eoaccess.EODatabaseChannel.selectObjectsWithFetchSpecification(EODatabaseChannel.java:234) at com.webobjects.eoaccess.EODatabaseContext._objectsWithFetchSpecificationEditingContext(EODatabaseContext.java:3055) at com.webobjects.eoaccess.EODatabaseContext.objectsWithFetchSpecification(EODatabaseContext.java:3195) at com.webobjects.eocontrol.EOObjectStoreCoordinator.objectsWithFetchSpecification(EOObjectStoreCoordinator.java:488) setObject() is called, and then eventually setTimestamp(int, Timestamp) then setTimestamp(int, Timestamp, null). So I was wrong to implicate EOF even peripherally: looks like it's entirely the JDBC driver. -- Paul. http://logicsquad.net/ ___ Do not post admin requests to the list. They will be ignored. Webobjects-dev mailing list (Webobjects-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/webobjects-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Re: datepicker on AMD problem
This ADP is inline, however, i get the same result from a component action. Ted --- On Sun, 12/5/10, Simon wrote: From: Simon Subject: Re: datepicker on AMD problem To: "Theodore Petrosky" Cc: webobjects-dev@lists.apple.com Date: Sunday, December 5, 2010, 12:33 PM are you rendering the content inline ? we use ADP in AMD's quite a bit and not seen this issue, but we render content from a component action. i'm wondering if rendering inline is somehow causing it to get confused about the location of the text field. i'll also double check with our UI chap at work tomorrow because we've got various patches (read: hacks) associated with AMD and ADP. simon On 5 December 2010 17:05, Theodore Petrosky wrote: I have a little problem with a datepicker on an AMD. Here is an example app. http://173.54.121.129/cgi-bin/WebObjects/DatePicker this is simulating if I had a table of data that extends past the bottom of the display so i have to scroll down to click the link that brings in the AMD. Is there something more I am supposed to be doing to make the DatePicker attach to the AMD and not the page that is bringing in the AMD? Am I not supposed to use a ADP on an AMD? Ted ___ Do not post admin requests to the list. They will be ignored. Webobjects-dev mailing list (Webobjects-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/webobjects-dev/simon%40potwells.co.uk This email sent to si...@potwells.co.uk ___ Do not post admin requests to the list. They will be ignored. Webobjects-dev mailing list (Webobjects-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/webobjects-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
[ANN] Safari Extension 0.95
Hi @, A new version of WebObject Safari Extension is available. Download : http://www.webobjects.me/safari/woExtension/index.html What's new? * Blog is now on one extern Page. * Golipse download and explanation is now on one extern Page. * Safari Extension explanation is new * Add GoDeploy Icon. Go to an extern Page for Deploy Scripts. (under Construction) Thank you K's ROOM (ISHIMOTO Ken) [E-Mail] [iChat:] [HP] http://www.ksroom.com/ _ This e-mail has not been scanned for viruses because it was written on an Mac, and there are NO Viruses on an Apple Computer. For further information visit http://www.apple.com ___ Do not post admin requests to the list. They will be ignored. Webobjects-dev mailing list (Webobjects-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/webobjects-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Re: datepicker on AMD problem
are you rendering the content inline ? we use ADP in AMD's quite a bit and not seen this issue, but we render content from a component action. i'm wondering if rendering inline is somehow causing it to get confused about the location of the text field. i'll also double check with our UI chap at work tomorrow because we've got various patches (read: hacks) associated with AMD and ADP. simon On 5 December 2010 17:05, Theodore Petrosky wrote: > I have a little problem with a datepicker on an AMD. Here is an example > app. > > http://173.54.121.129/cgi-bin/WebObjects/DatePicker > > > this is simulating if I had a table of data that extends past the bottom of > the display so i have to scroll down to click the link that brings in the > AMD. > > Is there something more I am supposed to be doing to make the DatePicker > attach to the AMD and not the page that is bringing in the AMD? > > Am I not supposed to use a ADP on an AMD? > > Ted > > > > ___ > Do not post admin requests to the list. They will be ignored. > Webobjects-dev mailing list (Webobjects-dev@lists.apple.com) > Help/Unsubscribe/Update your Subscription: > > http://lists.apple.com/mailman/options/webobjects-dev/simon%40potwells.co.uk > > This email sent to si...@potwells.co.uk > ___ Do not post admin requests to the list. They will be ignored. Webobjects-dev mailing list (Webobjects-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/webobjects-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
datepicker on AMD problem
I have a little problem with a datepicker on an AMD. Here is an example app. http://173.54.121.129/cgi-bin/WebObjects/DatePicker this is simulating if I had a table of data that extends past the bottom of the display so i have to scroll down to click the link that brings in the AMD. Is there something more I am supposed to be doing to make the DatePicker attach to the AMD and not the page that is bringing in the AMD? Am I not supposed to use a ADP on an AMD? Ted ___ Do not post admin requests to the list. They will be ignored. Webobjects-dev mailing list (Webobjects-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/webobjects-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Re: Timestamps and PostgreSQL
I just couldn't let this go. Anyone not using PostgreSQL can tune out.
On 04/12/2010, at 9:47 AM, Paul Hoadley wrote:
>>> Displaying those times _without_ a formatter in the app shows the first
>>> timestamp as '2010-12-03 04:51:55 Etc/GMT' which is exactly as expected,
>>> but the second (that was entered from the console behind the app's back) as
>>> '2010-12-02 18:21:55 Etc/GMT'.
>>>
>>> Is it obvious to any other PostgreSQL users what is going on? It looks
>>> like the database server's machine-local timezone is being used to adjust
>>> the timestamp somewhere between WebObjects and the database.
>>
>> I agree.
>
> After some furious Googling, it turns out that the JDBC driver is affected by
> the 'user.timezone' system property. Launching the app with
> '-Duser.timezone=GMT' on my laptop (which otherwise gets set to
> 'Australia/Adelaide') results in the behaviour I was expecting. I assume
> this was the missing link in the "UTC from top to bottom" chain, as I can see
> from the logs that both app servers were picking up their respective local
> values for this property. Just to be clear, this doesn't actually matter too
> much if the app is just running indefinitely on the same server: the
> timestamps seem to be adjusted on the way in and back out of the database, so
> that everything Just Works from within the app. It was only when I dumped
> and restored the database somewhere else that it became a problem.
The JDBC driver is, indeed, adjusting the timestamp, and apparently this is a
feature. EOF calls PreparedStatement.setTimestamp(int, Timestamp) (implemented
in AbstractJdbc2Statement), which in turn calls
PreparedStatement.setTimestamp(int, Timestamp, Calendar) with a null Calendar.
To format the timestamp, TimestampUtils.toString(Calendar, Timestamp) is
called, again with a null Calendar. The null Calendar is replaced by a new
GregorianCalendar(), which is used to format the string sent to the database.
All of this is fine, _except_ that the GregorianCalendar's timezone turns out
to be the machine-local timezone, despite calling
TimeZone.setDefault(TimeZone.getTimeZone("GMT")) in the application's
constructor. (It's not obvious to me why this would be the case—does
TimeZone.setDefault() not apply JVM-wide?) In any case, setting the
user.timezone system property (to UTC) _is_ sufficient to cause that
GregorianCalendar to be constructed with its timezone set to UTC, and then the
JDBC driver doesn't perform any adjustments to timestamps on the way in or out.
The PostgreSQL JDBC driver's behaviour strikes me as somewhat
counter-intuitive, to say the least. I'm pretty sure it nails the issue of
keeping timestamps in UTC with Postgres, though.
--
Paul.
http://logicsquad.net/
___
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (Webobjects-dev@lists.apple.com)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/webobjects-dev/archive%40mail-archive.com
This email sent to arch...@mail-archive.com
