No hyphenation in hyphenated block

Question

Is there a possibility to disable hyphenation for a text snippet inside a fo:block? My problem is That it’s not possible to set hyphenate="false" on fo:inline because it’s not a property. And fo:block inside fo:block creates a new line...

Fo-Example:

<fo:block hyphenate="true">
This text should be hyphenated <fo:inline hyphenate="false">This text shouldn’t be hyphenated</fo:inline>
</fo:block>

UPDATE:

I tried the soulution Posted by @DanielNorberg because none of all solutions worked properly. It seems to be kind of a workaround but still not delivering the output i would like to get.

I use this template:

<xsl:template match="text()[ancestor::span[@class='nobreak']]">
    <xsl:param name="text"><xsl:value-of select="." /></xsl:param>
    <fo:inline hyphenate="false" color="red">
        <xsl:for-each select="tokenize(replace(replace($text,'(.)','$1\\n'),'\\n$',''),'\\n')">
        <fo:inline keep-with-next.within-line="always">
            <xsl:value-of select="."/>
            <fo:character font-size="0pt">
                <xsl:value-of select="' '" />
            </fo:character>
        </fo:inline>
        </xsl:for-each>
    </fo:inline>
</xsl:template>

The fo part looks like this

<fo:block xmlns:fn="http://www.w3.org/2005/xpath-functions" space-after="14pt">
    <fo:block text-align-last="left" font-size="10pt" color="black" text-align="justify"
        font-family="ITCFranklinGothicStd-Book" line-height="14pt" wrap-option="wrap">
        <fo:block hyphenate="true" xml:lang="de">
            <fo:block>Die Entwicklung der folgenden Jahre bestätigte unsere Auffassung. Nicht nur erwiesen
                sich die <fo:inline hyphenate="false" color="red">
                    <fo:inline keep-with-next.within-line="always">T<fo:character font-size="0pt"> </fo:character></fo:inline>
                    <fo:inline keep-with-next.within-line="always">r<fo:character font-size="0pt"> </fo:character></fo:inline>
                    <fo:inline keep-with-next.within-line="always">e<fo:character font-size="0pt"> </fo:character></fo:inline>
                    <fo:inline keep-with-next.within-line="always">i<fo:character font-size="0pt"> </fo:character></fo:inline>
                    <fo:inline keep-with-next.within-line="always">b<fo:character font-size="0pt"> </fo:character></fo:inline>
                    <fo:inline keep-with-next.within-line="always">e<fo:character font-size="0pt"> </fo:character></fo:inline>
                    <fo:inline keep-with-next.within-line="always">r<fo:character font-size="0pt"> </fo:character></fo:inline>
                </fo:inline>
            </fo:block>
        </fo:block>
    </fo:block>
</fo:block>

So the word "Treiber" should not be hyphenated. But the PDF output looks the following:

SOLUTION UPDATE: The final workaround that worked for me was similar to the template above but with adding a non-breaking space (⁠) between each character.

<xsl:template match="text()[ancestor::span[@class='nobreak']]">
    <xsl:param name="text"><xsl:value-of select="replace(., ' ', '&#160;')" /></xsl:param>
    <fo:inline>
        <xsl:for-each select="tokenize(replace(replace($text,'(.)','$1\\n'),'\\n$',''),'\\n')">
        <fo:inline>
            <xsl:value-of select="."/>
            <!-- non-breaking invisible space after each character-->
            <fo:inline>&#8288;</fo:inline>
        </fo:inline>
        </xsl:for-each>
    </fo:inline>
</xsl:template>

Much thanks to @DanielNorberg

Daniel Norberg · Accepted Answer · 2020-02-05T10:53:40.353

3

I found a solution for this (verified for FOP 2.2). The code "disables" hyphenation and keeps the inline content on the same line. It's of course a hack with no guarantee to work for later releases but it works for my purpose, maybe it can help you as well.

Sample XML:

<block>Lorem ipsum <inline class='nobreak>This is neither hyphenated or line broken</inline>.</block>

XSLT 2.0 solution:

Create a parent inline element (XSLT 1.0 and 2.0):

<xsl:template match="inline[@class='nobreak"]">
   <fo:inline>
      <xsl:apply-templates />
   </fo:inline>
</xsl:template>

"Disable" hyphenation and make the text content to be kept on the same line (XSLT 2.0):

<xsl:template match="text()[ancestor::inline[@class='nobreak']]">
    <xsl:param name="text"><xsl:value-of select="." /></xsl:param>
    <xsl:for-each select="tokenize(replace(replace($text,'(.)','$1\\n'),'\\n$',''),'\\n')">
       <fo:inline keep-with-next.within-line="always">
          <xsl:value-of select="."/>
          <fo:character font-size="0pt">
             <xsl:value-of select="' '" />
          </fo:character>
       </fo:inline>
    </xsl:for-each>
</xsl:template>

UPDATE XSLT 2.0, this seems to be better:

<xsl:template match="text()[ancestor::phrase[@class='nobreak']]">
    <xsl:param name="text"><xsl:value-of select="replace(., ' ', '&#160;')" /></xsl:param>
    <xsl:for-each select="tokenize(replace(replace($text,'(.)','$1\\n'),'\\n$',''),'\\n')">
       <fo:inline>
          <xsl:value-of select="."/>
          <fo:inline font-size="0pt">
             <xsl:value-of select="'$'" />
          </fo:inline>
       </fo:inline>
    </xsl:for-each>
</xsl:template>

If you want an XSLT 1.0 solution for this you can easily find code on the web that parses the text one character at the time.

edited Feb 05 '20 at 10:53

answered Feb 03 '20 at 11:11

Daniel Norberg

61
5

Wow if that works that’s gonna be great. I will try this as soon as possible. Thanks! – CrazyEight Feb 04 '20 at 10:08
Ok i tried now as you said. It sadly doesnt really fix the problem for me. my .fo output is the following: ```xml Die Entwicklung der folgenden Jahre bestätigte unsere Auffassung. Nicht nur erwiesen sich die T r von gestern... ``` – CrazyEight Feb 04 '20 at 22:02
1

I'll check your fo code as soon as possible. Hopefully I can find something obvious. – Daniel Norberg Feb 05 '20 at 05:43
I can unfortunately not post my whole fo code because of the character limitation... I will update my Post so u can see my fo and pdf output – CrazyEight Feb 05 '20 at 08:56
i updated my Original Post. Hopefully u have all necessary information – CrazyEight Feb 05 '20 at 09:32
1

I updated my comment with new XSLT. See "UPDATE XSLT 2.0". It seems to work better. Can you try? – Daniel Norberg Feb 05 '20 at 10:54
1

@CrazyEight Did you see that I updated the XSLT? I am new on Stack Overflow and do not have enough rep to add comments to your original question but you can see the new XSLT in the updated answer. – Daniel Norberg Feb 06 '20 at 08:24
I got a meesage that u updated your answer. And WOW it works!!! Thank you very much. Ive been struggeling with this so long! – CrazyEight Feb 06 '20 at 09:07
Glad to hear that :-) I am happy too since I needed this badly myself. I searched around for solutions first but found none. Now there is at least a workaround for this on the web if others have the same issue. – Daniel Norberg Feb 06 '20 at 11:17
I had a little different solution with adding a non-breaking space after each character. I updated my Question if u want to see the final solution that worked for me. As i tried out your solution i recognised that words connected by a "-" were still hyphenated. With non breaking space i could not find any hyphenations anymore. Thanks for your big help! – CrazyEight Feb 07 '20 at 08:27
Thanks! I will investigate that with my own stylesheet as soon as possible. – Daniel Norberg Feb 07 '20 at 11:00

score 1 · Answer 2 · answered Jan 21 '20 at 16:42

1

The feature that should do the trick is:

<fo:block hyphenate="true">
  This text should be hyphenated 
  <fo:inline keep-together.within-line="always">This text shouldn’t be hyphenated</fo:inline>
</fo:block>

but was reported to be buggy some years ago. So, if you have a version of fop that is recent enough, you could try it.

answered Jan 21 '20 at 16:42

Pierre François

5,850
1
17
38

But if the line is only wide enough for "This text should be hyphenated This"? – Tony Graham Jan 21 '20 at 17:24
Also tried that. I am using fop2.3. I think that’s the most recent version? – CrazyEight Jan 21 '20 at 17:58
My version is now fop 2.1 and it doesn't work for me either. Sorry. – Pierre François Jan 22 '20 at 15:39
The last version of fop is 2.4. – Pierre François Jan 22 '20 at 16:01

Kevin Brown · Answer 3 · 2020-01-22T20:02:06.720

I would think the only solution would be XSL FO markup that would be like this:

<fo:block hyphenate="true">
<fo:inline>I am some text that can be hyphenated. </fo:inline> 
<fo:inline keep-together.within-line="always">I</fo:inline> <fo:inline keep-together.within-line="always">am</fo:inline> <fo:inline keep-together.within-line="always">text</fo:inline> <fo:inline keep-together.within-line="always">that</fo:inline> <fo:inline keep-together.within-line="always">isn't</fo:inline> <fo:inline keep-together.within-line="always">hyphenated</fo:inline>.
</fo:block>

In theory, set hyphenation on the overall block and any child fo:inline that must be kept together in the line (word by word) specified. Now, how you get there with some template is likely not straight-forward considering what maybe are word boundaries and what are punctuation and such.

This would require some testing. You might also need to put breaking-space characters or additional inline elements inbetween the words with the space character.

Update 1

I used the above and created a simple FO. Formatting with RenderX I get what I would expect, even changing spome margins I can get hyphenation in the areas where it is not marked and none in the area where keeps are set.

RenderX output:

Using the exact same template (even adjusting margins several times because each formatter is different when it comes to line ending treatment), I cannot get any hyphenation at all in FOP.

FOP output:

Conclusion is that there seems to be an issue in using FOP and attempting to do what I said as a workaround.

Update 2

With RenderX as well as AHF as stated below, setting hyphenation on fo:inline works. Here is output from RenderX using this:

 <fo:block hyphenate="true">
    <fo:inline>I am some text that can be hyphenated. I am some text that can be hyphenated. I am some text that can be hyphenated. </fo:inline> 
    <fo:inline hyphenate="false">I am some text that cannot be hyphenated. I am some text that cannot be hyphenated. I am some text that cannot be hyphenated. </fo:inline>         
 </fo:block>

The problem is that as soon as fop-hyphenation decides to hypenate a word it doesn't matter if there is `keep-together.within-line="always"`. I think the only solution is to set a `fo:block` twhere u can set `hyphenate="false"` that behaves like a `fo:inline` — CrazyEight, Jan 22 '20 at 08:00
Thanks for your trouble... is there any possibility to use the ability from RenderX or AHF of setting hyphenation property on a `fo:inline` with Apache fop2.3 ? — CrazyEight, Jan 23 '20 at 08:46
Sure there is an option, but not "use the ability" as FOP is FOP and the other products are closed-source commercial. Apache FOP is open source. You could download the code or hire folks to do it or submit a request to the FOP team do it. — Kevin Brown, Jan 23 '20 at 18:34

score 1 · Answer 4 · answered Feb 03 '20 at 05:35

One trick is to set a "bogus" language for the inner <inline>:

Here, I've set the inner block to Ukrainian, so the formatter attempts to use Ukrainian hyphenation rules, which obviously are not applicable.

In fact, I've found this trick because I needed Ukrainian hyphenation for one of my documents and it did not apply unless you explicitly specify the language.

<fo:block hyphenate="true">
This text should be hyphenated This text should be hyphenated This text should be
hyphenated <fo:inline  color="green" language="uk">This text shouldn’t be
hyphenated This text shouldn’t be hyphenated This text shouldn’t be
hyphenated </fo:inline>
</fo:block>

Result using RenderX XEP:

left: as in OP's question; right: with "bogus" language set

Tried this but still hyphenating... but i am also using `wrap-option="wrap"` and `text-align="justify"` on the outer `fo:block` but still hyphenating with inner ``... — CrazyEight, Feb 04 '20 at 22:45

Tony Graham · Answer 5 · 2020-01-25T09:17:22.120

0

It should be possible to set hyphenate="false" on the fo:inline. hyphenate is an inherited property, and it applies to both fo:block and fo:character. (See https://www.w3.org/TR/xsl11/#fo_character and https://www.w3.org/TR/xsl11/#hyphenate.) Every character in the text of your XSL-FO document is interpreted as an fo:character, so it should just work.

For this XSL-FO:

<!-- Without fo:inline/@hyphenate: -->
<fo:block hyphenate="true" color="red">
This text should be hyphenated This text shouldn’t be hyphenated</fo:block>
<!-- With fo:inline/@hyphenate: -->
<fo:block hyphenate="true">
This text should be hyphenated <fo:inline hyphenate="false">This text shouldn’t be hyphenated</fo:inline>
</fo:block>

I get this result with AH Formatter V6.6:

edited Jan 25 '20 at 09:17

answered Jan 21 '20 at 17:32

Tony Graham

7,306
13
20

I will try this for a different word in my text. Maybe it’s just buggy with that one word. I tried both: `hyphenate="false"` on the `fo:inline`, but also `keep-together.within-line`... none of them worked – CrazyEight Jan 21 '20 at 17:56
I do not think hyphenate is a valid option on fo:inline, it applies to block level elements only. – Kevin Brown Jan 21 '20 at 23:53
It does not apply to `fo:inline`, but it should be inherited by the nominal `fo:character` for each character in the text content of the `fo:inline`. – Tony Graham Jan 22 '20 at 09:44
ok thanks... with apache fop 2.3 and version 2.4 it doesnt work – CrazyEight Jan 22 '20 at 15:42
@TonyGraham, you may want to edit your sample for future records, someone looking at the result may be confused. Your first line which *can* be hyphenated has the sentence "This text shouldn't be ...". I knew what you meant but it could be confusing. – Kevin Brown Jan 23 '20 at 18:37
Thanks. I added comments in the XSL-FO instead. My first test for this had different text in each `fo:block`, but that made it hard to be sure that the difference wasn't the result of the ordinary line-breaking. This way, the only difference is the `hyphenate` property value. – Tony Graham Jan 25 '20 at 11:15

No hyphenation in hyphenated block

5 Answers5

Update 1

Update 2