[GSoC 2012][svg export filter] outstanding weekly reports (was: Re: [Libreoffice] [GSoC 2012][svg export filter] Weekly Report #9)

[Marco Cecchetti]

On Tue, 07 Aug 2012 19:47:23 +0200, Thorsten Behrens  
<thb at documentfoundation.org> wrote:

> Hi Marco,
>
> you were missing two weekly reports in a row - I hope everything is
> alright your side?

Hi Thorsten,
I'm sorry, I know. The problem is that I'm facing several issues both
personal and technical. So I needed to suspend the log book activity
for concentrating on implementation.
The state of the new svg text exporting feature is the following:

Text is exported with running characters, a <text> element is created
for each text shape, a <tspan> element is created for
1) each text paragraph specifying common font properties
2) each text line or text portion that need to be explicitly
    positioned (e.g. superscript, subscript, bullets)
3) each text portion with specific character properties

The layout is:
<text>
   <tspan class="TextParagraph" ... font properies here> // a paragraph  
with single line
     <tspan class="TextPosition" x="25" y="50" >
       <tspan>Hello </tspan>
       <tspan font-size="800">world</tspan>
       <tspan>!</tspan>
     </tspan>
   </tspan>
   <tspan class="TextParagraph" ... font properies here>
     <tspan class="TextPosition" x="25" y="100" >
       <tspan>This is the first line</tspan>
     </tspan>
       <tspan>This is the second line line</tspan>
     </tspan>
   </tspan>
<text>

To achieve this result I investigated which meta-actions are
produced by a text shape, in such a way I found out that some
interesting meta-comment actions are generated for each end of
line, paragraph, text shape starts and ends. Unfortunately no
specific meta-comment action is generated for list-item and
user inserted line-break. That made more complex the
implementation.
On the other side I have also taken confidence with XText,
XTextRange, XTextContent and all other interfaces needed
to extract, paragraph and text properties from a Text Shape
(XText). I found out how to extract info on list styles,
the same is true for TextFields and specifically for Hyperlinks.

Anyway changing the way of exporting text is a hard job and
several things are still to be fixed.

At present not all bullets based on characters are exported
correctly and bullets based on bitmaps are not exported at all.
Other things that doesn't work anymore are special text such as
shadowed and outlined text, and text decorations such as
underline and strike-through.
Moreover having changed the exported text shape layout, at
present text fields on master pages are not handled correctly
by the JavaScript engine anymore.

Text searching and text selection is working on browsers that
support such features. Unfortunately when the mouse button is
released after performing a selection, the presentation switches
automatically to the next slide. I hope to fix this issue at a
later time by defining some ad-hoc onmouseover event for <text>
elements.

Sorry again for having missed the weekly reports but as you can
see I'm short on time and there are several things to be fixed
in order to have some text exporting features working as it used
to work.

Cheers,
-- Marco


-- 
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/