[sword-devel] reference parser - yet another bug?

David Haslam dfhmch at googlemail.com
Fri Feb 13 01:42:00 MST 2015

In the Romanian USFM files, book abbreviations in cross-references end in a
full-stop for 49 of the 66 books.
They match exactly what is defined in \toc3

There are no instances of "Jude." or "Jude"  There are 419 instances of

These full-stops are retained by the OSIS converter usfm2osis.py, 
but the Perl script xreffix.pl removes these full-stops before parsing.
It leaves them removed after it's done its job.
IMHO, this is wrong, because it's a change from the authentic source text.

When displaying a cross-reference, the module should show the reference item
exactly as it was in the source text, albeit as a linked item facilitated by
the OSIS stuff normally hidden to the reader.

I emailed Peter to start a discussion about this yesterday.

What would happen if the locale file contained in the [Book Abbrevs] lines
like this?


and the full-stop removal was disabled?


View this message in context: http://sword-dev.350566.n4.nabble.com/reference-parser-yet-another-bug-tp4654610p4654615.html
Sent from the SWORD Dev mailing list archive at Nabble.com.

More information about the sword-devel mailing list