Problem

The mid stored to the segment is used as identifier for re-importing and pivot translation matching.

The mid contains currently the trans-unit id and just a counter for the whole file, so that the mid is unique per translate5 file.
The reason for just adding a counter were:

• translate5 has no influence on how the the trans-unit id may be generated outside, so there might be duplicated IDs although that makes technically no sense
• a xlf file may contain several <file> tags where the trans-unit ID may be duplicated between the different <file> tags
• since the trans-unit id may contain several sub-segments in the form of mrk tags, the trans-unit id is not enough

That counter is now the problem, that the mid may not match on re-import / pivot translation if the segmentation changes. Due changed segmentation the trans-unit ID might be still matching but the counter will have different values, therefore the whole mid will not match anymore.

solution

The new mid should be created out of this data:

translate5 file fileId + xliff>file[id] + $transunitId + [mrkid or mrk counter] 

TO BE EDITED BY THOMAS

        \editor_Models_Import_FileParser_Xlf::parseSegmentAttributes
        $fileId + xliff>file + $transunitId → hash → in bestehendes transunitId Feld
        bestehendes TransunitIdFeld umbenennen in transunit hash
        segmentMid beinhaltet zukünftig: transunitHash + + [mrkid | mrk counter]

In addition, for each translate5 file add a new flag isReimportable.
It should be set to false on import

• if the file format does not support re-import → this is currently implemented somehow in the UI
• when for the mrk no mids were set (mids are not mandatory there)

For re-import only this flag is then evaluated in the UI