Stephen,
If I add the dita.temp.dir property to the ant script and process a transform through ant, my temp files end up in c:\temp where I expect them.
If I change the conductor_template.xml dita.temp.dir property to C:${file.separator}temp and run a transform through ant, my temp files end up in the c:\temp directory where I expect them.
I am still not able to run a transform through the java command line with dita-ot1.2. I can’t get past the “Unable to instantiate specified logger class org.dita.dost.log.DITAOTBuildlogger” even though all of my install directories and environment variables match the 1.2 install docs.
------------------------------------------
Jerry Kimmel – jkimmel at vasont dot com
-----Original Message-----
From: dita-users@yahoogroups.com
[mailto:dita-users@yahoogroups.com] On Behalf
Of Jerry Kimmel
Sent: Friday, March
17, 2006 9:42 AM
To: dita-users@yahoogroups.com
Subject: RE: [dita-users] Re:
DITA-OT1.1.2.1 Temp directory problem
Stephen,
No, the same issue also exists in release 1.2.
Here is my test for DITA-OT1.2:
Change the dita.temp.dir property in conductor_template.xml from the default "${basedir}${file.separator}temp" to C:${file.separator}temp.
Delete the DITA-OT1.2\temp directory
Run an xhtml transform from the commandline.
As the transform is running, watch the DITA-OT1.2 directory – you will see a temp directory be created.
Watch the c:\temp directory - you will see all of the other temp files populate this temp directory.
After the transform is finished, open the DITA-OT1.2\temp directory – the property.temp file will have been created here, rather than in the c:\temp directory.
There is one difference between DITA-OT1.1.2.1 and DITA-OT1.2. For DITA-OT1.2 I must edit the conductor_template.xml, whereas in DITA-OT1.1.2.1 I edit conductor.xml. The conductor.xml is rewritten each time a transform is run in DITA-OT1.2.
FYI – I was unable to actually complete a transform in DITA-OT1.2. I added the DITA-OT1.2 dost.jar to my CLASSPATH, but did not install Sun JDK 1.5. (I get the “Unable to instantiate specified logger class org.dita.dost.log.DITAOTBuildlogger” error.)
Will DITA-OT1.2 *only* run in Sun JDK 1.5.x? If so, this needs to be spelled out more clearly in the release notes.
Even though my transform failed, the DITA-OT1.2\temp directory is created and populated with the new property.temp file in DITA-OT1.2.
------------------------------------------
Jerry Kimmel - jkimmel at vasont dot com
-----Original Message-----
From: dita-users@yahoogroups.com
[mailto:dita-users@yahoogroups.com] On Behalf
Of imagiczhang
Sent: Thursday, March 16, 2006
8:40 PM
To: dita-users@yahoogroups.com
Subject: [dita-users] Re: DITA-OT1.1.2.1
Temp directory problem
Jerry,
If possible, can you have a try on the latest
release of DITA-OT. I
tested on release 1.2 and believe that this bug is
fixed on release
1.2 . Thanks.
Best Regards,
Stephen Zhang
--- In dita-users@yahoogroups.com, "Jerry
Kimmel" <jkimmel@...> wrote:
>
> We ran into an issue when we changed the
default dita.temp.dir property
> (in conductor.xml) from the default
"${basedir}${file.separator}temp" to
> a directory on a mapped drive, and deleted
the temp directory from the
> DITA-OT1.1.2.1 directory tree.
>
> FYI - the mapped drive is different for each
user.
>
> Everything was fine except that the
"property.temp" file was still being
> written to the 'default' DITA-OT/temp
directory (and the temp directory
> was being created), rather than to the
directory on the mapped drive.
>
> This created a problem from a permissions
point of view in that a user
> who couldn't write to the 'default'
DITA-OT/temp directory couldn't run
> any transforms - even though he had permission
to the new temp directory
> on the mapped drive, he had no write
permissions to the 'default'
> directory where DITA-OT was installed. He now
has to have permission to
> write here.
>
> As a temporary fix, I make sure to delete the
'default' DITA-OT/temp
> directory after every transform - but this
still presents a problem
> because different users could potentially
overwrite (or delete) each
> other's property.temp file because that file
always get written to the
> same place.
>
> ------------------------------------------
> Jerry Kimmel - jkimmel at vasont dot
com
>