OpenXML Filter: merge different font style run

[yan cheng]

In the docx file shown below, the first is a special character that corresponds to the run tag

‌

            <w:r>
                <w:rPr>
                    <w:rFonts w:ascii="Symbol" w:eastAsia="Symbol" w:hAnsi="Symbol" w:cs="Symbol"/>
                    <w:szCs w:val="21"/>
                </w:rPr>
                <w:t></w:t>
            </w:r>
            <w:r>
                <w:rPr>
                    <w:rFonts w:ascii="Arial" w:hAnsi="Arial" w:cs="Arial"/>
                    <w:szCs w:val="21"/>
                </w:rPr>
                <w:t>0.000</w:t>
            </w:r>

Obviously their fonts are different, but during the parsing process, the RunMerger class execution method canMergeWith mistakenly assumes that they have the same format, which eventually causes the XLF file trans-unit format to be parsed as

‌

<trans-unit id="NFDBB2FA9-tu1" xml:space="preserve">
<source xml:lang="en-US"><g id="1">0.000</g></source>
<seg-source><mrk mid="0" mtype="seg"><g id="1">0.000</g></mrk></seg-source>
<target xml:lang="zh-CN"><mrk mid="0" mtype="seg"><g id="1">0.000</g></mrk></target>
</trans-unit>

Ideally, it would look like this

<trans-unit id="NFDBB2FA9-tu1" xml:space="preserve">
<source xml:lang="en-US"><g id="1"></g><g id="2">0.000</g></source>
<seg-source><mrk mid="0" mtype="seg"><g id="1"></g><g id="2">0.000</g></mrk></seg-source>
<target xml:lang="zh-CN"><mrk mid="0" mtype="seg"><g id="1"></g><g id="2">0.000</g></mrk></target>
</trans-unit>

According to the current parsing situation, which is the content in the first trans-unit tag, I modified the XLF file after my translation, and finally exported the translation as follows

‌

            <w:r>
                <w:rPr>
                    <w:rFonts w:ascii="Arial" w:hAnsi="Arial" w:cs="Arial"/>
                    <w:szCs w:val="21"/>
                </w:rPr>
                <w:t>0.000</w:t>
            </w:r>

Then the font of the first special character was also changed and displayed in office as

‌

[Chase Tingley]

@Denis Konovalyenko is this related to some of the cases you’ve already been looking at?

[Denis Konovalyenko]

@Chase Tingley , it seems to me that this is a regression after fonts run property mutation was addressed in the scope of issue #853. I am looking into this.

[Denis Konovalyenko]

The following extraction is also possible if the mechanism, the effective run fonts content categories are detected with, is further improved:

<trans-unit id="NFDBB2FA9-tu1" xml:space="preserve">
<source xml:lang="en">0.000（草坪或地面）</source>
<target xml:lang="fr">0.000（草坪或地面）</target>
</trans-unit>

‌

[Denis Konovalyenko]

A related pull request #369 has been opened.

[Denis Konovalyenko]

The pull request #369 was merged.