Review Request 109887: Create a stream reader better suited for ODF than QXmlStreamReader

[Jos van den Oever]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://git.reviewboard.kde.org/r/109887/#review31094
-----------------------------------------------------------


The current use is this:
  if (el.namespaceUri() == KoXml::fo && el.name == "border") { ... }
You propose to have code like this:
  if (el.qualifiedName() == "fo:border") { ... }

That is still too error-prone. This would be better:
  if (el === ODFElement::fo::border) { ... }

It has the advantage that no typo in the namespace or local name are possible: the compiler will check that.
In addition, the code behind the operator== and even the type of ODFElement::fo::border can be optimized independently of the code that uses it.

- Jos van den Oever


On April 6, 2013, 3:56 p.m., Inge Wallin wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://git.reviewboard.kde.org/r/109887/
> -----------------------------------------------------------
> 
> (Updated April 6, 2013, 3:56 p.m.)
> 
> 
> Review request for Calligra, Jarosław Staniek and Jos van den Oever.
> 
> 
> Description
> -------
> 
> This patch contains a new XML stream reader based on the QXmlStreamReader that is better suited for ODF.
> 
> Much ODF parsing code in Calligra looks like:
> 
>   if (el.namespaceUri() == KoXml::fo && el.name == "border") { ... }
> 
> The reason for this complicated construction is that the prefix (the "fo" in "fo:border") is not unique but is declared at the beginning of each Xml file. Even though "fo" is the normal prefix there is no guarantee that it is  the same in every document. 
> 
> However, it is a very rare document where it is *not* the normal prefix, so what we want to do is to be able to write code like this:
> 
>   If (el.qualifiedName() == "fo:border") { ... }
> 
> and make the XML stream reader or dom tree rewrite the qualified name in the very rare cases that the prefix does not match what we want.
> 
> This is exactly what the KoXmlStreamReader does. It allows you to write easier and faster code while still be correct in the few cases where the prefixes are not the expected ones. It does this by letting the user enter the expected namespace declarations and then compare those to the actual namespace declarations in the document. If they match everything will be as fast as possible. If they don't, it will be slower but still correct.
> 
> As an extra feature I have allowed the user to declare some extra namespaces (see fixNamespace() in KoXmlReader.cpp). This will let documents created by old versions of OpenOffice.org be read even though they use other namespaces.
> 
> I have code that uses this file but that is not yet ready for review. I wanted to put this up early to get feedback while the rest of the yet unfinished code is maturing.
> 
> 
> Diffs
> -----
> 
>   libs/odf/CMakeLists.txt 3680486 
>   libs/odf/KoXmlStreamReader.h PRE-CREATION 
>   libs/odf/KoXmlStreamReader.cpp PRE-CREATION 
> 
> Diff: http://git.reviewboard.kde.org/r/109887/diff/
> 
> 
> Testing
> -------
> 
> Not much. I will do that when the code that uses this code is ready. This review is for getting feedback on the ideas and implementation details.
> 
> 
> Thanks,
> 
> Inge Wallin
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/calligra-devel/attachments/20130415/e5ed42e5/attachment.htm>