#129 stylesheets incorrectly handle word break across lines with paragraphos

[Tom Elliott]

Consider this example in which a word is broken across a line with an intervening paragraphos:

<ab>
   <lb n="13"/> <w>τοῖϲ</w> <w>δ'</w> <w>ἀνθρώ
   <milestone rend="paragraphos" unit="undefined"/>
   <lb n="14" break="no"/>ποιϲ</w> 
</ab>

The stylesheets produce:

τοῖϲ δ' ἀνθρώ 
—— -
ποιϲ

Note the appearance of the hyphen after the paragraphos, instead of after the omega at the end of the preceding line where we would expect it.

Even if we remove the intervening whitespace-only text nodes...

<ab>
   <lb n="13"/> <w>τοῖϲ</w> <w>δ'</w> 
   <w>ἀνθρώ<milestone rend="paragraphos" unit="undefined"/><lb 
      n="14" break="no"/>ποιϲ</w> 
</ab>

... we still get this undesirable result.

If we remove the "milestone" element for the paragraphos entirely...

<ab>
   <lb n="13"/> <w>τοῖϲ</w> <w>δ'</w> <w>ἀνθρώ
   <lb n="14" break="no"/>ποιϲ</w> 
</ab>

... the stylesheets behave (NB: even with the intervening whitespace nodes):

τοῖϲ δ' ἀνθρώ-
ποιϲ

Stay tuned for more on this subject as the DCLP project delves into the code ...

[BODARD Gabriel]

There is code in there to deal with the similar problem of intervening <g> and <space> elements, so adding <milestone> to that template ought to be trivial. (Let me know if I can help identify the relevant templates/tests.)

[Tom Elliott]

Here's where this first came up: https://github.com/DCLP/dclpxsltbox/issues/52

[Tom Elliott]

Courtesy of Chirag Gajjar, here's a list of the XSL files in which the strings "break" or "inWord" appear (these are attribute names or values, uniquely diagnostic for P5- and P4-era intra-word line breaks in EpiDoc):

htm-teiab.xsl
htm-teilb.xsl
htm-teimilestone.xsl (no "inWord")
teig.xsl
teilb.xsl
teilgandl.xsl
teispace.xsl (no "inWord")
teisupplied.xsl (no "inWord")
tpl-apparatus.xsl (no "inWord")
txt-teiab.xsl
txt-teilb.xsl
txt-teimilestone.xsl (no "inWord")

[Tom Elliott]

fixed with [r2354]

[Tom Elliott]

actually, not done properly. See https://github.com/DCLP/dclpxsltbox/issues/52

[BODARD Gabriel]

Punting this to next release.

[Tom Elliott]

I unfortunately do not have time to address this issue before 8.22 code freeze.

[Tom Elliott]

A fix for this has been tested and currently resides in /branches/dclp-issue52/example-p5-stylesheets

[Scott Vanderbilt]

Bump -> Future

[Pietro Maria Liuzzo]

I will look at this and check with @paregorios

[BODARD Gabriel]

According the comments above a fix for this ticket has been available since 2017. @pietroliuzzo, @paregorios: was this what you were both planning to try to merge into the Reference Stylesheets? If so, do you think this will be implemented by September?