OpenXML Filter: improve Word styles optimisation for the reiterated properties on the directly applied formatting level

[Denis Konovalyenko]

Please consider the following case. There is a paragraph with 2 runs, the second of which reiterates the rFonts property, which is present under the default paragraph style applied for that run already on the paragraph level.

<w:p w:rsidR="00D577F2" w:rsidRPr="0076395F" w:rsidRDefault="0076395F">
    <w:pPr>
        <w:rPr>
            <w:lang w:val="en-US"/>
        </w:rPr>
    </w:pPr>
    <w:r w:rsidRPr="00D00CAC">
        <w:rPr>
            <w:sz w:val="24"/>
            <w:szCs w:val="24"/>
            <w:lang w:val="en-US"/>
        </w:rPr>
        <w:t>Run 1.</w:t>
    </w:r>
    <w:r w:rsidR="004629C4" w:rsidRPr="00D00CAC">
        <w:rPr>
            <w:rFonts w:ascii="Arial" w:hAnsi="Arial" w:cs="Arial"/>
            <w:lang w:val="en-US"/>
        </w:rPr>
        <w:t>Run 2.</w:t>
    </w:r>

<w:style w:type="paragraph" w:default="1" w:styleId="Normal">
    <w:name w:val="Normal"/>
    <w:qFormat/>
    <w:rPr>
        <w:rFonts w:ascii="Arial" w:hAnsi="Arial" w:cs="Arial"/>
    </w:rPr>
</w:style>

Currently, this is extracted as

<trans-unit id="NFDBB2FA9-tu1" xml:space="preserve">
    <source xml:lang="en"><g id="1">Run 1.</g><g id="2">Run 2.</g></source>
    <target xml:lang="fr"><g id="1">Run 1.</g><g id="2">Run 2.</g></target>
</trans-unit>

It would be nice to have the following extraction:

<trans-unit id="NFDBB2FA9-tu1" xml:space="preserve">
    <source xml:lang="en"><g id="1">Run 1.</g>Run 2.</source>
    <target xml:lang="fr"><g id="1">Run 1.</g>Run 2.</target>
</trans-unit>

For more information please refer to the attached document.

[Chase Tingley]

With the fix to #887, this example now extracts as

<trans-unit id="NFDBB2FA9-tu1" xml:space="preserve">
<source xml:lang="en"><g id="1">Run 1.</g>Run 2.</source>
<target xml:lang="fr"><g id="1">Run 1.</g>Run 2.</target>
</trans-unit>

Does the fix for #887 serve as a more general form of this improvement, or are there still cases of this sort that would require their own handling?

[Denis Konovalyenko]

@Chase Tingley , you re right, the solution for issue #887 provides a more general improvement and this is how runs are extracted with it applied:

<trans-unit id="NFDBB2FA9-tu1" xml:space="preserve">
<source xml:lang="en">Run 1.<g id="1">Run 2.</g></source>
<target xml:lang="fr">Run 1.<g id="1">Run 2.</g></target>
</trans-unit>

There are 2 equal sets of run properties, so, the first run properties are considered as a base for others.

With this issue improvement, we would get:

<trans-unit id="NFDBB2FA9-tu1" xml:space="preserve">
<source xml:lang="en"><g id="1">Run 1.</g>Run 2.</source>
<target xml:lang="fr"><g id="1">Run 1.</g>Run 2.</target>
</trans-unit>

Where the second run is going to get empty properties, thus, they will be chosen as a base for others.

[Chase Tingley]

@Denis Konovalyenko Based on the previous comments, can we close this?

[Denis Konovalyenko]

@Chase Tingley , another round of styles optimisations (do not remember sharply which one) has led to fixing this particular issue:

<source xml:lang="en"><g id="1" ctype="x-empty" equiv-text="&lt;run1>">Run 1.</g>Run 2.</source>

Thanks for keeping an eye on this!

[Denis Konovalyenko]

The issue has been resolved in the scope of other ones related to the styles optimisation process.