Hi Michael, thanks for your reply.
My apologies for the mail, I actually did not thought about different
mail ready system. But fortunately your guessing skills helps :-) !
I had -strip:none to the command line and it does the trick : I get
exactly the same indentation result as when commenting the DOCTYPE on
the source file !
I'm using Saxon-HE 9.2.1.1J
I've just tried with the brand new "Saxon-HE 9.4.0.1J", and it is the
same behaviour.
I guess -strip:none is not the default option.
Thanks a lot anyway, this is what I was looking for :-) !!
Just one more question about the output :
A <meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />
element is added to the <head> element, this is the only residual
unindented element in the file. I'm not sure why saxon add this ? It
doesn't seems to depend on the xsl:output settings : even when
xsl:output is absent I get the <meta> added, with no attribute
@encoding="UTF-8" on the xsl:output I also get the meta....
Regards,
Matthieu
Le 02/02/2012 12:09, Michael Kay a écrit :
On 02/02/2012 10:48, Matthieu Ricaud-Dussarget wrote:
Hi all,
In my project I concatenate multiple xhtml files in one xml files.
This aggregate file has to be edited by hand, that means indentation
is important here for convenience.
Before I discovered XML Catalog, I used to delete all DOCTYPE
declarations within source XHTML file with a perl script (which also
remplace named entities with UTF-8 ones). This worked fine : the
concatenated files were indented exactly like the XHTML sources.
But this was a bit dangerous in case I didn't match a special entity
to replace with perl. And this was not a really good XML practice.
Now that I'm using a local XML Catalog and run my tranformation with
Saxon in command line with this options :
-r:org.apache.xml.resolver.tools.CatalogResolver
-x:org.apache.xml.resolver.tools.ResolvingXMLReader
-y:org.apache.xml.resolver.tools.ResolvingXMLReader
I can't see exactly what's happening here because your mail client and
mine have conspired to ignore the whitespace which was critical to
understanding your message.
Generally, if you validate against a DTD, then whitespace in elements
whose content model is defined as element-only (for example head and
body) will be treated as ignorable, which means it's liable to be lost
in a copy operation. Perhaps this is what is happening.
Try the option -strip:none on the command line to prevent this
behaviour. The documentation says this is the default, but I'm not
convinced it is correct: I seem to remember it changing some time ago
in response to a W3C change.
Michael Kay
Saxonica
--~------------------------------------------------------------------
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
To unsubscribe, go to: http://lists.mulberrytech.com/xsl-list/
or e-mail: <mailto:xsl-list-unsubscribe(_at_)lists(_dot_)mulberrytech(_dot_)com>
--~--
--
Matthieu Ricaud
05 45 37 08 90
NeoLibris
--~------------------------------------------------------------------
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
To unsubscribe, go to: http://lists.mulberrytech.com/xsl-list/
or e-mail: <mailto:xsl-list-unsubscribe(_at_)lists(_dot_)mulberrytech(_dot_)com>
--~--