XLIFF Joiner does not properly merge <file> elements that were split up

[Chase Tingley]

• In Rainbow, add the attached XLIFF to a project. Note that this XLIFF contains only a single <file> element. Create a pipeline consisting of the "XLIFF Word-Count Splitter" step. Set the max word count param to 100. Execute the pipeline.
• This will generate a large set of PART XLIFF files. (Also attached as a zip, for reference.)
• Clear (or restart) rainbow, and add the PARTs as input files. Create a pipeline containing "XLIFF Joiner". Execute the pipeline.

Examine the resulting CONCAT XLIFF file. It contains multiple <file> elements, each with original="400.html".

This is XLIFF corruption. If the original XLIFF was produced by extraction from another source format (as in this case), the CONCAT XLIFF will not merge correctly.

[Chase Tingley]

Fix #724 - Correctly join XLIFF parts split by word count

Better handling of cases where a single <file> element in the
original XLIFF was split across multiple `PART` files by the
splitter.  Improved unittest coverage.

→ <<cset 3ee9dceb5087>>

[Chase Tingley]

Reopening because the code isn't merged yet

[Chase Tingley]

Fix #724 - Correctly join XLIFF parts split by word count

Better handling of cases where a single <file> element in the
original XLIFF was split across multiple `PART` files by the
splitter.  Improved unittest coverage.

→ <<cset 833562f3a719>>

[Chase Tingley]

Merged in issue724 (pull request #266)

Fix #724 - Correctly join XLIFF parts split by word count

Approved-by: YvesS [e-mail]

→ <<cset 77315b9e08a4>>