Incorrect segmentation of TMX file in 2-3 alignments...

[Former user]

Original [issue 277](https://code.google.com/p/okapi/issues/detail?id=277) created by @ysavourel on 2012-09-28T18:56:37.000Z:

I have an srx that segments source text and translated texts correctly, but if I feed Rainbow TMX files and ask for segmentation of both flagging the ´Verify that a target segment matches each source segment when a target content exists´ option, something goes wrong: the target segement is empty if the source segment would be split in 2 and the target in 3 for instance. I would expect that the source and target in such a case would not be touched as the split condition is not present.

[Former user]

Comment [1.](https://code.google.com/p/okapi/issues/detail?id=277#c1) originally posted by @ysavourel on 2012-10-02T15:28:53.000Z:

[Former user]

Comment [2.](https://code.google.com/p/okapi/issues/detail?id=277#c2) originally posted by @ysavourel on 2012-10-04T19:19:27.000Z:

The segmentation output is as expected.

The input: "This sentence should not be split. "

will produce: [This sentence should not be split.][ ]

Assuming no whitespace is trimmed. This produces a usually unwanted additional segment made up of a single whitespace.

The above segmentation was creating a 2-1 alignment by default which was unexpected by the user.