OK, it does look there is a bug, related to handling of default paragraph
directionality across delimited text run boundaries. If we reformat your
input FO a little to show nesting:

<fo:block font-family="FreeSans">
  <fo:block/>[1] RTL insert קום
  <fo:block/>[2] (parens wrong).
  <fo:block/>[3] Back to normal because of this RTL insert: בּ .
</fo:block>

I see you are using <fo:block/> in an unusual fashion, effectively as a
paragraph separator. As a consequence, the outer block effectively has 6
child blocks, where the three text runs constitute three anonymous child
blocks. It appears that the bidi context of the first of these "[1] ..." is
being fed to the second of these "[2] ..." erroneously.

I've opened an issue at [1].

[1] https://issues.apache.org/jira/browse/FOP-2290

There is a simple work around for this problem until a fix is effected.
Namely, wrap the three texts as separate blocks rather than mapping your
input <br/> to <fo:block/>.

<fo:block font-family="FreeSans">
  <fo:block>[1] RTL insert קום<fo:block>
  <fo:block>[2] (parens wrong).<fo:block>
  <fo:block>[3] Back to normal because of this RTL insert: בּ .<fo:block>
</fo:block>

On Tue, Aug 6, 2013 at 12:34 PM, Raphael Finkel <raph...@cs.uky.edu> wrote:

> Thanks, Glenn, for the suggestion.
>
> Here is the FO file, which I call mini.fo.xml, without any of the XML/XSL
> encrustation:
>
>         <?xml version="1.0" encoding="UTF-8"?><fo:root xmlns:fo="
> http://www.w3.org/1999/XSL/Format";><fo:layout-master-set><fo:simple-page-master
> master-name="myPage"><fo:region-body/></fo:simple-page-master></fo:layout-master-set><fo:page-sequence
> master-reference="myPage"><fo:flow flow-name="xsl-region-body"><fo:block
> font-family="FreeSans"><fo:block/>[1] RTL insert קום <fo:block/>[2] (parens
> wrong).<fo:block/>[3] Back to normal because of this RTL insert: בּ
> .</fo:block></fo:flow></fo:page-sequence></fo:root>
>
> I convert to PDF using fop-1.1 like this:
>
>         /tmp/fop-1.1/fop -c mini-fo.cfg -fo mini.fo.xml mini.pdf
>
> The cfg file is just to define a Hebrew-aware font, FreeSans.
>
> As before, the pdftotext version of the PDF output looks like
>
>         [1] RTL insert ‫קום‬
>         ]2[ )parens wrong(.
>         [3] Back to normal because of this RTL insert: ּ‫. ב‬
>
> showing the bug.
>
> Raphael
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: fop-users-unsubscr...@xmlgraphics.apache.org
> For additional commands, e-mail: fop-users-h...@xmlgraphics.apache.org
>
>

Reply via email to