Re: unstable openoffice metadata

["Andreas J. Guelzow" <aguelzow pyrshep ca>]

On Wed, 2009-11-18 at 13:09 +0800, bill lam wrote:
> I found that if a workbook has been saved from openoffice. It add some
> metadata into it and gnumeric retains this metadata even if the
> workbook was converted into gnumeric's xml format.  However the 
> metadata changes from version to version of gnumeric even nothing
> actually changed and that workbook was not accessed by openoffice
> again.  As an example, the diff of a workbook 
>
> @@ -1,6 +1,6 @@
>  <?xml version="1.0" encoding="UTF-8"?>
>  <gnm:Workbook xmlns:gnm="http://www.gnumeric.org/v10.dtd"; 
> xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"; xsi:schemaLocation="http://www.gnumeric.org/v9.xsd";>
> -  <gnm:Version Epoch="1" Major="9" Minor="15" Full="1.9.15"/>
> +  <gnm:Version Epoch="1" Major="9" Minor="16" Full="1.9.16"/>
>    <gnm:Attributes>
>      <gnm:Attribute>
>        <gnm:type>4</gnm:type>
> @@ -30,12 +30,12 @@
>    </gnm:Attributes>
>    <office:document-meta xmlns:office="urn:oasis:names:tc:opendocument:xmlns:office:1.0" 
> xmlns:xlink="http://www.w3.org/1999/xlink"; xmlns:dc="http://purl.org/dc/elements/1.1/"; 
> xmlns:meta="urn:oasis:names:tc:opendocument:xmlns:meta:1.0" xmlns:ooo="http://openoffice.org/2004/office"; 
> office:version="1.1">
>      <office:meta>
> -      <meta:creation-date>2009-04-22T19:24:48Z</meta:creation-date>
> +      <meta:user-defined meta:name="msole:codepage" meta:type="string">-535</meta:user-defined>
>        <meta:user-defined meta:name="gsf:thumbnail" meta:type="string"></meta:user-defined>
>        <meta:editing-cycles>0</meta:editing-cycles>
> -      <dc:date>2009-04-22T19:24:48Z</dc:date>
> -      <meta:user-defined meta:name="msole:codepage" meta:type="string">-535</meta:user-defined>
>        <meta:editing-duration>2009-04-22T19:24:48Z</meta:editing-duration>
> +      <meta:creation-date>2009-04-22T19:24:48Z</meta:creation-date>
> +      <dc:date>2009-04-22T19:24:48Z</dc:date>
>        <meta:print-date>2009-04-22T19:24:48Z</meta:print-date>
>      </office:meta>
>    </office:document-meta>
>
> specifically the position for the lines of creation-date and codepage
> oscillating for nearly every release of gnumeric.  Is this related to
> gnumeric itself or other xml library?  Is there any option in
> ssconvert to strip off all foreign metadata for gnumeric xml target?

The meta data information is read and written by libgsf. I wouldn't be
surprised if you would see complete reversion of the metadata, but that
doesn't seem to be the case.

Of course the order of the metadata info shouldn't really matter.

The metadata is gnumeric's metadata (you can edit it in
file->properties) so I wouldn't consider it "foreign metadata". Ww just
happen the use the same format as for our metadata as ODF.

Andreas


-- 
Andreas J. Guelzow <aguelzow pyrshep ca>