Forum OpenACS Q&A: Translation server questions: Audit trails, character encoding

Collapse
1: Translation server questions: Audit trails, character encoding
Posted by Joel Aufrecht on 10/19/04 10:46 PM
Giancarlo Luxardo asked a few questions which may be of general interest:

<blockquote> - is it normal that we don't get an audit trail for the previous
translations? (is it lost when a new release is cut?)
</blockquote>

The audit trail on the translation server is continuous, but catalog
files do not contain audit trails.  This has several implications.

Developers can add messages to
catalogs without using the translation server, either by editing the
catalogs directly or by exporting from their own servers.  If these
files are committed to CVS, I import them onto the translation server
before the next release.  If this results in new or changed messages
on the translation server, the audit trail shows only my name, not the
original author.  The audit trail from the original server is lost.

Also, when you install a new instance of OpenACS, you get all of the
messages but none of the audit trail.

For example, if we have the string "foo" on the translation server,
the audit trail might be:

"baz" - Joel, changed 7 Feb 2003
"bar" - Peter, changed 6 Feb 2003
"foo" - Joel, original, 1 Jan 2001

Now Rocael installs a new site on 31 Dec 2003.  On Rocael's server, the audit trail is:

"baz" - Rocael, original, 31 Dec 2003

Rocael changes it a few times:

"quux" - Milo, changed 31 March 2004
"qux" - Rocael, changed 29 Feb 2004
"baz" - Rocael, original, 31 Dec 2003

Rocael exports the catalog and commits the file to CVS, and on 1 July 2004 I import catalogs as part of the next release process.  The audit
trail on translate.openacs.org would now read:

"quux" - Joel, changed 1 July 2004
"baz" - Joel, changed 7 Feb 2003
"bar" - Peter, changed 6 Feb 2003
"foo" - Joel, original, 1 Jan 2001

<blockquote> - the XML catalog files for Italian contain quoted text for the special
characters (eg: &agrave; for à ); I don't think this is correct because
this is not the way it is done for the other languages (although it
works if the db is set up correctly); I think the best would be to edit
the XML files directly.

</blockquote>

We would like to push to unicode for all catalogs, rather than HTML
special characters.  I agree that the best fix is to edit the XML and
then commit them.  Let me know when updated catalogs are committed to
CVS, and for which catalogs, and I can update the translation server.

Collapse
2: Re: Translation server questions: Audit trails, character encoding (response to 1)
Posted by Giancarlo Luxardo on 10/20/04 05:47 PM
<blockquote>The audit trail on the translation server is continuous, but catalog
files do not contain audit trails.  This has several implications.

</blockquote>

Thanks for the explanation. I assume that the current audit trails will be kept (unless a new db is installed) but it is not possible to track what happened before this date:
First translated by Joel Aufrecht on 2004-02-06

<blockquote>

then commit them.  Let me know when updated catalogs are committed to
CVS, and for which catalogs, and I can update the translation server.

</blockquote>
... for those who have commit rights to CVS.