Writer2ePub bug reports

[eBookLuke]

Originally Posted by clarity

Hi Luke,
1) heading link ids get changed to 'headering_id_xx'.

2) bookmark links get removed.

3) images in frames don't appear when viewing the epub, but are exported.

4) in my book source file two tables that were directly one after another caused the plugin to crash.

1) can you explain better this issue?

2) I still work on the internal links, external links are working

3) and 4) I know that issues, I still work on it. The crash in the tables was create by a bug, and will be solved in the next release.

I don't know anything about openoffice plugins or BASIC, but I'm an experienced programmer.

Thanks Joe, do you have experience in XSLT? I wish implement a totally different approach to the conversion, but I have a lot of problems with XSL transformation.
If you can help me, W2E can dramatically improve the performance

Luke

[clarity]

Hi Luke,

Thanks for your quick reply. I've been looking through the code in the plugin archive. I guess Traverse.xba and Functions.xba are the main files for my issues. There's a lot of rather distracting &quot; &comma; &lt; etc in the files though. Does the file get passed into another form? could just writing the " ' < etc not work?

Anyway, I don't have XSLT experience. Most of my programming has been in C++, Java (which I notice in the plugin), LAMP stack, Perl and others. But reading through http://www.w3schools.com/xsl/xsl_intro.asp I can see why you might want to use it. If you go down that road, I might be able to offer some assistance with debugging.

re your questions:

when you create headers (Header 1, Header 2 etc.) for chapter, section, sub-sections in writer, writer 'tags' those sections so you can reference them with cross-references (Insert -> Cross-reference...). The tags/ids are also used by the table of contents that writer can generate. I also use cross-references to Bookmarks (Insert -> Bookmark) for the book index and for the 'see page xx' references in the main book text.

The epub source from your plugin for the table of contents writer generates for my book looks like this:

<p><b>Contents</b></p>
<p><a href="#__RefHeading__52796_1662616202"><i>Acknowle dgements vi</i></a></p>
<p><a href="#__RefHeading__3271_928939564"><i>Introducti on vii</i></a></p>
<p><a href="#__RefHeading__55283_791455818"><i>How to use this book ix</i></a></p>

<p><a href="#__RefHeading__87745_29593047"><b>Obstacles and new perspectives</b></a><a href="#__RefHeading__87745_29593047"><b> </b></a><a href="#__RefHeading__87745_29593047"><b>1</b></a></p>

So I'm guessing writer is generating its own tags for headers (and I guess other internal links), but your plugin is replacing the link destination ids as previously described with 'heading_id_xx', hence the links break.

If I need to I'm prepared to fix the table of contents manually, but the other internal links, mainly bookmarks, would take a heck of a lot of labour to fix that way and I have the feeling that if I knew your code better there's probably a couple of lines that would allow the internal links ('see page xx' or, in the book index 'honesty xx-xx,xx,xx...') to be preserved.

From the looks of the plugin code I think it will take me longer than I can spend at the moment to figure out for myself, but if you would like me to test or try out specific changes to the code, I'm happy to do that.

cheers,
Joe

[Michael_ZZ]

Originally Posted by eBookLuke

Thanks, I will check for it.
Can you send me some .odt examples to writer2epub[at]gmail.com?
Thanks

*Hrmpf* Forget about my lame talking ... Today everything works as expected
W2E correctly fills out all the defined metadata fields -- if defined correctly.

However, understanding that one took me a while.
My ODT is saved as "Lorem Ipsum - Lady Mondegreen - 2011.odt"

If in the W2E extension I define "T - A - Y" the fields contain the expected texts: Title="Lorem Ipsum", Author="Lady Mondegreen" and Year="2011".
If I define just "T" everything's just left blank.
To get the file name completely into the title field I have to define the metadata prefs as "T - -".
(It was my imagination that with such a definition Title would only result in "Lorem Ipsum", cutting off Author and Year. Defining "T - A -" does exactly that: cutting off the year.)

And the second issue was a misunderstanding as well. I thought Writer2EPUB would store all metadata somewhere into the source ODT. So whenever I open it another time and click the W2E metadata button all information would be there again ... which isn't the case obviously

[EDIT]
Hm, the behaviour is still quite puzzling. Defining "T - - Y" now also leads to the complete file name resting in the title field...
[/EDIT]

[eBookLuke]

Originally Posted by Michael_ZZ

[EDIT]
Hm, the behaviour is still quite puzzling. Defining "T - - Y" now also leads to the complete file name resting in the title field...
[/EDIT]

Thanks for your feedback… As soon as I have some spare time, I will investigate about this issue.

About the metadata storage:
all the fields filled in the metadata window will be stored in the original .odt file. You can find the fields when open File>Properties>Custom Properties.

I will check also about missing storage.

Thanks

Luke

[Michael_ZZ]

Originally Posted by eBookLuke

About the metadata storage:
all the fields filled in the metadata window will be stored in the original .odt file. You can find the fields when open File>Properties>Custom Properties.

Wow. Great!

Thanks for leading me there

ZZ

P.S.: I just see that there are no w2e_... metadata names present in my (German) LO properties' names list. Is there a list with possible names anywhere?

[varelov]

Trying to convert an .odt file into an epub and pressed on the first of three buttons. I got slapped with an error message: BASIC runtime error. Object variable not set.
OS is Windows XP Service Pack 3. OpenOffice version 3.3.0

P.S. This .odt document is copy-paste from a web page, if it changes anything...

[eBookLuke]

Originally Posted by varelov

Trying to convert an .odt file into an epub and pressed on the first of three buttons. I got slapped with an error message: BASIC runtime error. Object variable not set.
OS is Windows XP Service Pack 3. OpenOffice version 3.3.0

P.S. This .odt document is copy-paste from a web page, if it changes anything...

Hi,

can you send to me the original .odt file to writer2epub[at]gmail.com?
Probably there are some objects in the document that crash the macro

Thanks

Luke

[BobC]

Originally Posted by varelov

Trying to convert an .odt file into an epub and pressed on the first of three buttons. I got slapped with an error message: BASIC runtime error. Object variable not set.
OS is Windows XP Service Pack 3. OpenOffice version 3.3.0

Just an observation but I use OOOFBTools to build FB2 books, quite similar to W2E. When I updated to OOO 3.3.0 I discovered a number of similar Basic run-time errors when using the macros. It appears they have made changes to BASIC in 3.3.0 which may clash with existing code.

BobC

[net-buoy]

Is writer2epub still not fully functional for LibreOffice? I do not get a response to preference button and the macro drops all mathematical formulae.

[eBookLuke]

Originally Posted by net-buoy

Is writer2epub still not fully functional for LibreOffice? I do not get a response to preference button and the macro drops all mathematical formulae.

W2E is devoleped with LibreOffice!
About the response, please wait a bit after pressing the preference button. It need some time to build the font list. If you have a lot of fonts, the time will be longer. If you press again this button while it's worging, the time restarts. If you press another button, the process quits.

About the formulae, there is no process at the moment. Epub doesn't supports natively formulas, but I'studiyng a method to render them.

Luke

[net-buoy]

Ahhhh, I am now informed that this is not a function of writer2epub, but of the failure of epub readers to read mathml islands.....

[DaleDe]

Originally Posted by net-buoy

Ahhhh, I am now informed that this is not a function of writer2epub, but of the failure of epub readers to read mathml islands.....

You can convert MathML to SVG which ePUB does support. MathML will be in the 3.0 ePUB version.

http://sourceforge.net/projects/svgmath/

DAle

[PeterMzizi]

I have just been playing around with some documents, and notice that document metadata is not copied into the metadata dialogue in writer2epub. Looking at some of the previous comments, in this thread, it implies that this should happen.
Title is present in both the odt and epub so this would be useful if it would copy over.
Author (as you have previously mentioned) in odt may not reflect the real author but how about owner? In LibreOffice this is an option under custom properties (as is publisher). Could creator be used as author with the option to use owner instead if wanted (or if not empty)?
Tags (epub) and keywords (odt) are more or less the same and could be useful if copied
Comments (odt) and description (epub) would also be potentially useful if copied across.

Thanks

[eBookLuke]

Originally Posted by PeterMzizi

I have just been playing around with some documents, and notice that document metadata is not copied into the metadata dialogue in writer2epub…

Hello Peter,

I wish to improve the metadata handling in W2E, but I still need to understand what is better.

We can get metadata from multiple source:
1) filename: the filename can contain author/title/other infos.
2) native document metadata: many document formats (.ost, rtf, .doc…) can store some metadata. The document title, the creator and other data can be available in such fles.
3) special W2E metadata fields: to better handle the metadata, W2E adds many user fields to the original documents. So, the document can embed some data difficult to fit in the field of naive documents (like original title, first publishing date, isbn and so on)

Now I ask to all:
what must I do when one or more of the data sorce are available? Must I replace 3) with 2) if 2) is present? Or 3) must overwrite 2)?
Other suggestions?

thanks

Luke

[PeterMzizi]

Originally Posted by eBookLuke

Hello Peter,

I wish to improve the metadata handling in W2E, but I still need to understand what is better.

We can get metadata from multiple source:
1) filename: the filename can contain author/title/other infos.
2) native document metadata: many document formats (.ost, rtf, .doc…) can store some metadata. The document title, the creator and other data can be available in such fles.
3) special W2E metadata fields: to better handle the metadata, W2E adds many user fields to the original documents. So, the document can embed some data difficult to fit in the field of naive documents (like original title, first publishing date, isbn and so on)

Now I ask to all:
what must I do when one or more of the data sorce are available? Must I replace 3) with 2) if 2) is present? Or 3) must overwrite 2)?
Other suggestions?

thanks

Luke

Hi Luke
If it is a new .odt file then there will be no W2E metadata so use what is there. At this point a user may then add additional metadata, and if this document is now saved you could assume that is what the user would want next time.
so use 2 if no 3 but 3 if 2 and 3 present.
I guess the main area of conflict would be over owner/author and so for this 3 must be the right answer but other things like creation date may reflect when the work was written and not when typed into a computer. Certainly would be great to have the W2E metadata stored in the odt for future reference/use

Peter