Re: Output a new line after the XML declaration using indent="yes"

[Jörg Hohwiller <jo...@j-hohwiller.de>]

Hi there,


Brian Minchau wrote:
> 
> Hi Stanimir.
> 
> The Xalan
> serializer doesn't know about whether the serialized XML will
> be used in the the future as an external general parsed
> entity and included in yet another XML file.
> 
> It is possible that the XML will be included next to a text node that
> is not all whitespace and the extra whitespace that we inject after the
> XML header would be included next to non-whitespace
> text and become part of that text node, modifying it.
> 
> Extra whitespace added for indentation is done in ignorable locations,
> but this particular one (just after the header) might not be ignored.
> 
> Added indentation or extra whitespace before the document element
> is not always correct, so Xalan doesn't do it.
> 
> There is no Xalan specific option to control this behavior.
> 
> - Brian
> 

1. Could you please give an example or a link to the
XML-specification to point out why a newline after
XML-declation or at the end of the file should make the 
XML illegal. I am NOT talking about your problems
to implement this properly, but why
<?xml ....?>^n
<root>...</root>^n
should be illegal!?!?

2. Having a newline after XML-declaration as well
as at the end of the file is a common need that
JAXP users have. I can NOT accept your point
saying that this is generally NOT supported because
it could cause some trouble you may have implementing this.
This has a large impact after everybody has to eat this
when just using a plain JDK. It is a lot worse if people
start adding newlines manually to OutputStream,
because this will cause real trouble with encodings, especially
if the encoding is NOT 8-bit-wise (e.g. UTF-16) and a newlines
get broken because hackers of such workaround do NOT think
of such problems.

Please provide a solution here.

Thanks
  Jörg


-- 
View this message in context: http://www.nabble.com/Output-a-new-line-after-the-XML-declaration-using-indent%3D%22yes%22-tp15040090p24017219.html
Sent from the Xalan - J - Users mailing list archive at Nabble.com.

Re: Output a new line after the XML declaration using indent="yes"

[Jörg Hohwiller <jo...@j-hohwiller.de>]

Hi Dave,

thanks for your response...

> > 
> > 2. Having a newline after XML-declaration as well
> > as at the end of the file is a common need that
> > JAXP users have. I can NOT accept your point
> > saying that this is generally NOT supported because
> > it could cause some trouble you may have implementing this.
> I find the tone of your post to be offensive.  No one on this list owes 
> you anything in particular, and I suggest you be more polite in future 
> posts.

Sorry for that. Of course nobody owes me anything.
I just ran into this problem and read this thread.
This gave me the impression that a user demand
-that is obvious to me- is not really seen here (I am
talking about my impression). 
Besides English is not my native language so I do
not know if there is something very offensive in my words.
I just wanted to make my personal position very clear.
Besides I tended to always write "NOT" capitalized,
because I often read over this word and get the wrong 
message. A friend explained me that this is no good style
here because it overemphasizes what I wanted to say...

> > This has a large impact after everybody has to eat this
> > when just using a plain JDK. It is a lot worse if people
> > start adding newlines manually to OutputStream,
> > because this will cause real trouble with encodings, especially
> > if the encoding is NOT 8-bit-wise (e.g. UTF-16) and a newlines
> > get broken because hackers of such workaround do NOT think
> > of such problems.
> Any XML parser that requires a newline between the XML declaration and 
> the root element, or after the root element, is non-conforming and 
> should be fixed.

Absolutely true. But XML is also read by humans.
I wrote an open-source tool (maven-plugin) that modifies 
XML that can be handwritten. Users tell me that there is a
bug in my product because newlines get lost.
I think they are right but my problem is that
I see the main problem in xalan-j.

> > 
> > Please provide a solution here.
> You can always post-process the document and 
> add the new line characters.

Of course I can do that. I already wrote about this option
before and the pitfalls one can run into with encodings, etc.
I was choosing Xalan-J because I used JAXP
and did not want to add thirdparty libs (like XOM, JDOM or 
Dom4j). But if I need such workaround I'd better use one
of those external libs instead.

Besides Xalan-J also stips a newline between XML-Comments
and the root-tag. I can not simply workaround this problem without
reimplementing an XML-Writer or just using an other product.

> Dave

Regards
  Jörg


-- 
View this message in context: http://www.nabble.com/Output-a-new-line-after-the-XML-declaration-using-indent%3D%22yes%22-tp15040090p24129270.html
Sent from the Xalan - J - Users mailing list archive at Nabble.com.

Re: Output a new line after the XML declaration using indent="yes"

[David Bertoni <db...@apache.org>]

Jörg Hohwiller wrote:
> Hi there,
> 
> 
> Brian Minchau wrote:
>> Hi Stanimir.
>>
>> The Xalan
>> serializer doesn't know about whether the serialized XML will
>> be used in the the future as an external general parsed
>> entity and included in yet another XML file.
>>
>> It is possible that the XML will be included next to a text node that
>> is not all whitespace and the extra whitespace that we inject after the
>> XML header would be included next to non-whitespace
>> text and become part of that text node, modifying it.
>>
>> Extra whitespace added for indentation is done in ignorable locations,
>> but this particular one (just after the header) might not be ignored.
>>
>> Added indentation or extra whitespace before the document element
>> is not always correct, so Xalan doesn't do it.
>>
>> There is no Xalan specific option to control this behavior.
>>
>> - Brian
>>
> 
> 1. Could you please give an example or a link to the
> XML-specification to point out why a newline after
> XML-declation or at the end of the file should make the 
> XML illegal. I am NOT talking about your problems
> to implement this properly, but why
> <?xml ....?>^n
> <root>...</root>^n
> should be illegal!?!?
Brian didn't say it would make the result "illegal."  He simply said 
that it would modify the content of the result inappropriately.  The 
processor generates a external general parsed entity, and a newline 
between the XML declaration would introduce whitespace into the content 
of the entity.

> 
> 2. Having a newline after XML-declaration as well
> as at the end of the file is a common need that
> JAXP users have. I can NOT accept your point
> saying that this is generally NOT supported because
> it could cause some trouble you may have implementing this.
I find the tone of your post to be offensive.  No one on this list owes 
you anything in particular, and I suggest you be more polite in future 
posts.

> This has a large impact after everybody has to eat this
> when just using a plain JDK. It is a lot worse if people
> start adding newlines manually to OutputStream,
> because this will cause real trouble with encodings, especially
> if the encoding is NOT 8-bit-wise (e.g. UTF-16) and a newlines
> get broken because hackers of such workaround do NOT think
> of such problems.
Any XML parser that requires a newline between the XML declaration and 
the root element, or after the root element, is non-conforming and 
should be fixed.

> 
> Please provide a solution here.
You can always post-process the document and add the new line characters.

Dave