Bug in construction of POS list

[mthouv]

The Table panel of the 'FST-text' window contains an 'Export all text as POS list' button that converts the text automaton into a 'POS list' format. This format is a regular expression equivalent to the automaton of the sentence. However there seems to be some inconsistencies in the way the POS list file is created when it comes to compound words.

What steps will reproduce the problem ?

Open the attached French text and click the "Export all text as POS list" button.
80jours.zip

Go check line number 217 in the resulting file and search for the POS sequences corresponding to à cause de.

What is the expected output ?

We should see "à/à/PREP/z1" followed by "(cause/cause/N/z1:fs|cause/causer/V/z1:P1s:P3s:S1s:S3s:Y2s)"

What do you see instead ?

On line 217, we can see the sequences are displayed as followed:

"à/à/PREP/z1 à cause/à cause/ADV/PCDN/z1 (à cause de/à cause de/PREP/EPCPQ/z1|à cause de/à cause de/PREP/PCDN/z1|à cause de/à cause de/PREP/PCDN1/z1|à cause de/à cause/PREP/Prépconjs/1) (cause/cause/N/z1:fs|cause/causer/V/z1:P1s:P3s:S1s:S3s:Y2s)"

This regular expression corresponds to the following text: à à cause à cause de cause.