libxmlplusplus-general

[libxml++] ignore empty text nodes

[Andrej v. d. Z. <and...@gm...>]

Hi,

I am a happy user of libxmlpp for some time now. One thing I could not
find out yet. How can I ignore empty text nodes? The problem is that I
construct a dom document by reading two XML files and merging them
together with import_node(). I also remove some nodes manually with
remove_child(). Finally I do a doc->write_to_stream_formatted(cout,
"UTF-8") and end up with ugly output with holes in them like this:

<?xml version="1.0" encoding="UTF-8"?>
<session xmlns="XXX" version="1.0" clientName="Belastingdienst"
projectName="GBV" phase="1" iteration="1" sessionName="M2 - AUTO
TEST3">
  <hosts>
    <host name="apmvsq1" cluster="apmvsp1+2" type="mainframe"
smtEnabled="0" numCpus="666" opsys="ToBeOverwittenByParser"
hardware="ToBeOverWrittenByParser"
serviceLevel="ToBeOverwittenByParser" cpuSpeed="1"
cpuSpeedBenchmark="MIPS" cacheHitPerc="95" cacheHitTime="0.0015"
cacheMissTime="0.005" monitoringTool="ascb" timeDiff="67"/>







  </hosts>
<msrs><msr xmlns=XXX" type="SU" guiName="SU1"
ignoreEteValidator="0"><parser-logs><ete
filename="SU1/20100304-GBV-M2-SU1.ete"/><resource
filename="SU1/20100304-GBV-M2-SU1.mf"
hostname="apmvsq1"/></parser-logs></msr></msrs></session>


I understand that these are empty text-nodes, but I wish to ignore
them. How can I do that without writing my own version of
write_to_stream_formatted()? I just wish to ignore them altogether.

Thank you,
Andrej

Re: [libxml++] ignore empty text nodes

[Murray C. <mu...@mu...>]

On Fri, 2010-03-26 at 19:47 +0900, Andrej van der Zee wrote:
> Hi,
> 
> I am a happy user of libxmlpp for some time now. One thing I could not
> find out yet. How can I ignore empty text nodes?

By checking for them in your application when you read the XML document.
It's entirely up to your application to decide whether white space is
interesting.

>  The problem is that I
> construct a dom document by reading two XML files and merging them
> together with import_node(). I also remove some nodes manually with
> remove_child(). Finally I do a doc->write_to_stream_formatted(cout,
> "UTF-8") and end up with ugly output with holes in them like this:
> 
> <?xml version="1.0" encoding="UTF-8"?>
> <session xmlns="XXX" version="1.0" clientName="Belastingdienst"
> projectName="GBV" phase="1" iteration="1" sessionName="M2 - AUTO
> TEST3">
>   <hosts>
>     <host name="apmvsq1" cluster="apmvsp1+2" type="mainframe"
> smtEnabled="0" numCpus="666" opsys="ToBeOverwittenByParser"
> hardware="ToBeOverWrittenByParser"
> serviceLevel="ToBeOverwittenByParser" cpuSpeed="1"
> cpuSpeedBenchmark="MIPS" cacheHitPerc="95" cacheHitTime="0.0015"
> cacheMissTime="0.005" monitoringTool="ascb" timeDiff="67"/>
> 
> 
> 
> 
> 
> 
> 
>   </hosts>
> <msrs><msr xmlns=XXX" type="SU" guiName="SU1"
> ignoreEteValidator="0"><parser-logs><ete
> filename="SU1/20100304-GBV-M2-SU1.ete"/><resource
> filename="SU1/20100304-GBV-M2-SU1.mf"
> hostname="apmvsq1"/></parser-logs></msr></msrs></session>
> 
> 
> I understand that these are empty text-nodes, but I wish to ignore
> them. How can I do that without writing my own version of
> write_to_stream_formatted()? I just wish to ignore them altogether.
> 
> Thank you,
> Andrej
> 
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Libxmlplusplus-general mailing list
> Lib...@li...
> https://lists.sourceforge.net/lists/listinfo/libxmlplusplus-general


-- 
mu...@mu...
www.murrayc.com
www.openismus.com

Re: [libxml++] ignore empty text nodes

[Andrej v. d. Z. <and...@gm...>]

Hi,

>
> Yes. I guess you are doing that somehow anyway during your "merge" of
> two documents.

Actually I am using import_node().


> I guess I'd accept a patch that adds an
> Element::remove_empty_text_nodes() method. The you could call that on
> get_root_node().

I will have a look when I have time.

Though, I am doubting the correctness of the
"doc->write_to_stream_formatted()" method now.  I expect the output to
be formatted, but if you look at the last few nodes, they are not
formatted at all. See below. Am I misunderstanding something?

Thank you,
Andrej


-------------- output of write_to_stream_formatted(), is it really
formatted? ----------------


<?xml version="1.0" encoding="UTF-8"?>
<session xmlns="http://www.contentional.eu" version="1.0"
clientName="Belastingdienst" projectName="GBV" phase="1" iteration="1"
sessionName="M2 - AUTO TEST3">
  <hosts>
    <host name="apmvsq1" cluster="apmvsp1+2" type="mainframe"
smtEnabled="0" numCpus="666" opsys="ToBeOverwittenByParser"
hardware="ToBeOverWrittenByParser"
serviceLevel="ToBeOverwittenByParser" cpuSpeed="1"
cpuSpeedBenchmark="MIPS" cacheHitPerc="95" cacheHitTime="0.0015"
cacheMissTime="0.005" monitoringTool="ascb" timeDiff="67"/>







  </hosts>
<msrs><msr xmlns="http://www.contentional.eu" type="SU" guiName="SU1"
ignoreEteValidator="0"><parser-logs><ete
filename="SU1/20100304-GBV-M2-SU1.ete"/><resource
filename="SU1/20100304-GBV-M2-SU1.mf"
hostname="apmvsq1"/></parser-logs></msr></msrs></session>

Re: [libxml++] ignore empty text nodes

[Murray C. <mu...@mu...>]

On Fri, 2010-03-26 at 20:07 +0900, Andrej van der Zee wrote:
> Though, I am doubting the correctness of the
> "doc->write_to_stream_formatted()" method now.  I expect the output to
> be formatted, but if you look at the last few nodes, they are not
> formatted at all. See below. Am I misunderstanding something?

I stopped using it in Bakery (now in Glom) too. I think it just gives up
when it finds child text nodes, because it can't generically know if
they should be indented with white space.

I did this:
http://git.gnome.org/browse/bakery/commit/?id=54d85442f58228609a147f461934b75387fa8d7d
 

-- 
mu...@mu...
www.murrayc.com
www.openismus.com

Re: [libxml++] ignore empty text nodes

[Knut A. R. <kn...@if...>]

* Andrej van der Zee

> I am a happy user of libxmlpp for some time now. One thing I could not
> find out yet. How can I ignore empty text nodes?
[...]

Hi.

I have used the following XPath-based technique to remove such nodes in 
order to "normalize" XML-documents that are stored in pretty-printed 
format. The reason I had them stored pretty-printed, was to improve 
readability in a test suite. I would not use pretty-printing on XML that 
is going directly between machines.

void remove_whitespace_nodes(xmlpp::Document & doc)
{
     xmlpp::NodeSet whitespace_nodes = 
doc.get_root_node()->find("//text()[normalize-space()='']");
     for (xmlpp::NodeSet::const_iterator it = whitespace_nodes.begin();
          it != whitespace_nodes.end();
          ++it)
     {
         (*it)->get_parent()->remove_child(*it);
     }
}

Make sure the XPath above does not identify nodes that carries meaning 
to your application, otherwise you would have to refine it somehow.

-- 
Sincerely,
Knut Aksel Røysland

Re: [libxml++] ignore empty text nodes

[Andrej v. d. Z. <and...@gm...>]

Hi,

>
> void remove_whitespace_nodes(xmlpp::Document & doc)
> {
>    xmlpp::NodeSet whitespace_nodes =
> doc.get_root_node()->find("//text()[normalize-space()='']");
>    for (xmlpp::NodeSet::const_iterator it = whitespace_nodes.begin();
>         it != whitespace_nodes.end();
>         ++it)
>    {
>        (*it)->get_parent()->remove_child(*it);
>    }
> }
>

Thank you, that's a nice one!

Greets,
Andrej