[Solved+Issue] Writer - Exporting to XHTML with formatting?

[JonsBlog]

I've just updated to V.3.4.0, and suddenly the option to export to XHTML with page formatting seems to have vanished.

Previously with earlier versions, I created a document within Writer from scratch, chose "File > Export", then in "Save as type", I chose XHTML format, and Writer beautifully created an HTML document with all the formatting I used in my text document. But now, the only option there (yes, really, a dropdown select box with only one option!) is PDF.

So instead I try "File > Save As", then choose "Save as type: HTML Document (OpenOffice.org Writer)", and it does indeed create an HTML document, but all my page formatting is lost.

Is there some sort of new setup option I have to configure in order to save a document as HTML with page formatting? If so, where?

I've come to rely on this, using it several times each week, so I'd really like to be able to do what it did before.

Thanks in advance.
Jon

[acknak]

Strange. I have AOO 3.4, and if I do File > Export, XHTML is one of the available formats.

[rudolfo]

That's the same for me as for acknak: The XHTML export is just another option next to the PDF export. And I did not install any additional extension for this.
But I think there is one big difference between PDF and XHTML export: XHTML works as a XML filter (Tools -> XML Filter Settings ...)

XML Filter Settings

and is therefor using the Java/Xalan based XSL Transformation when exporting, while PDF export is a C++ or UNO object function and will always work out of the box. So you might check if your Apache OpenOffice can find an appropriate Java Runtime. Note that even on a 64bit Windows this has to be the 32bit version, and preferably 1.6.x, because OpenOffice is a 32bit application.

[floris v]

Export to xhtml is only available if you select support for, IIRC, legacy formats. It's available in LibreOffice.

[JonsBlog]

>> "Strange. I have AOO 3.4, and if I do File > Export, XHTML is one of the available formats."

Strange indeed! I just checked again, and it's definitely not there on my copy of the very same version. Anybody else experiencing this? Might there be some sort of Writer add-on I need to install, or an external program this version of Writer expects to find on my computer (external to itself) in order to provide this?

BTW, I'm running the 64-bit version of Windows Professional. Might this present some sort of problem here?

[JonsBlog]

>> "Export to xhtml is only available if you select support for, IIRC, legacy formats."

Sorry, floris v, but I don't find such a setting in Writer. Where, exactly, is that?

[JonsBlog]

OK, rudolfo, in my "XML Filter Settings" window, I find only "MS Excel 2003 XML" and "MS Word 2003 XML". How do I add those other options you show?

Also, you say "So you might check if your Apache OpenOffice can find an appropriate Java Runtime." How does one do that? I've looked under "Options" in Writer, but find no mention there anywhere of Java. Is that configured via some part of OO other than Writer?

[acknak]

Tools > Options > OO.org > Java

The Java settings are shared among all the OOo applications.

[JonsBlog]

Thanks acknak for directing me to that Java config section. The "Use Java runtime environment" box is checked, and it says it's using "Sun Microsystems Inc. 1.6.0_31". Is that correct?

[JonsBlog]

An update ... I've just now downloaded and installed Java 32-bit, then rebooted.

Re-invoked Writer, but still no option to save as HTML in "Export", and the option under "Save As" still loses all my document formatting.

Any more suggestions?

[acknak]

I'd take that as indicating that Java is installed and available to OOo.

I don't know if this export filter is packaged in some kind of additional package or component, or maybe as an extension? That's just my guess; I have it without installing anything special, but the Linux package may be different.

Did you do a complete install--no manual component selection?

[JonsBlog]

>> "Did you do a complete install--no manual component selection?"

Yes. Default "out of the box". No changes.

BTW ... after downloading/installing that 32-bit Java, I rechecked the Java config mentioned earlier, and now found two different versions, both the one I mentioned earlier and a new one "Oracle Corporation 1.7.0_04" I selected this new one, closed Writer and re-invoked it. Still no change to the problem.

I've just now checked the OO extensions web site and found something called "Writer2xhtml". I'll try installing that and see if it gets around this difficulty.

[JonsBlog]

heheheh ... this is just getting crazier and crazier ...

I just downloaded and installed the "Writer2xhtml" extension. It now gives me several "Export" options, and they all produce a document which, when viewed in either Internet Explorer or Mozilla Firefox, displays a totally blank page! This is from loading a previous .ODT file which was successfully exported to XHTML with an earlier version of Writer.

Really weird thing: When I use the "view source" option in Firefox, all the header and body data is there, and the text color directives are correct. Yet NOTHING appears in the browser. Amazing ...

[rudolfo]

If you see some of the XML filter settings this is usually a sign that the Java base XSLT transformation is working. And as far as I remember if you see some Java Runtime versions in the OpenOffice Options Dialog this indicates that you have a 32bit Java Runtime, because I am quite sure that the 64bit versions are not shown by OpenOffice (at least on my system with both JRE 32bit and JRE 64bit, only the 32bit in "C:\Program Files(x86)" are shown.

You can install the writer2latex extension to get around your problem, that you don't have XHTML export. Though you should have this XHTML export option without any need to install something additional, the XHTML export in writer2latex is somehow superior(improved indentation, more options for finetuning,...) and it is always worth installing, even if you never plan to use OpenOffice for working with TeX/LaTeX.

[Villeroy]

XHTML export works for me with AOO 3.4 under Linux (32 bit) and Java 1.7.

[JonsBlog]

OK guys ... first of all, thanks so much for all the help you've offered. Unfortunately, none of it has made Word usable for me as previous versions were. Over the past few years I've very much enjoyed using Word to create and convert with "one click" simple Word text pages into individual text web pages. But alas, so far as I'm concerned, and in my Windows 64-bit environment, this present version is simply broken, and none of the suggestions you've offered (much as I really do appreciate them) has fixed it.

I've now uninstalled and reinstalled this version of OO several times, with and without all the options. I've installed "Writer2xhtml" and "writer2latex" several times both separately and together. And yet still, any attempt to save a Word text page as html results in an html page with no margin settings and a default text font. The only thing which gets properly converted from .odt to .html is the font color. Maybe there are configuration settings within Word which can make this conversion process work, but thus far I haven't discovered any.

Obviously this version of Word has not been sufficiently debugged in a Windows 64-bit environment, and after fussing with this for several days now, I'm just out of time. I have deadlines to meet, so regrettably, I'll have to move on to some other word processor. Sometimes "If It Ain't Broke, Don't Fix It" really is the best policy.

Perhaps I'll check back in a few months for a later version in which this has been adequately addressed.

P.S. ... do any of you know if and where previous versions are available somewhere? I'd happily revert to an earlier version which works for me if I could find one.

Thanks again
Jon

[lgusaas]

P.S. ... do any of you know if and where previous versions are available somewhere? I'd happily revert to an earlier version which works for me if I could find one.

http://www.openoffice.org/download/legacy/index.html

[JonsBlog]

Thanks Lgusaas. Sounds like a plan. I'll go take a look ...

[JonsBlog]

Nope. Downloaded and installed OO 3.3.0, and still no usable HTML export. Guess I was using a version earlier than 3.3. Anybody know where I can find yet earlier versions? (God, I REALLY don't want to install Microsoft Office!)

Thanks
Jon

[Hagar Delest]

Have you tried to reset your OOo user profile?

[lgusaas]

Nope. Downloaded and installed OO 3.3.0, and still no usable HTML export. Guess I was using a version earlier than 3.3. Anybody know where I can find yet earlier versions? (God, I REALLY don't want to install Microsoft Office!)

Then the problem is with your setup since HTML export works for others on this list. Have you reset your user profile as Hager suggested?

If you want to try an older version go to the link I gave earlier. If you carefully read the whole page you would see under "Additional Resources" a link to "Archived legacy releases' which gives links to the archives mirrors with all older versions.

[JonsBlog]

Thanks, Hagar Delest ... just tried your suggestion.

No joy. Same problem.

Methinks the dead giveaway to the problem is the fact that previously, when I clicked within Word on "File > Export" it offered me the options of exporting the document either as .pdf or .html. Now, weirdly, that dropdown option box offers only .pdf and nothing else (regardless of which or how many extensions I install). I can still choose "Save As" then choose an .html output, but the resulting document loses all page formatting.

BTW ... I just now found a web site which will translate .odf files to .html on the fly:

http://www.convertfiles.com/convert/doc ... -HTML.html

... and it works perfectly. Just now converted two .odt files with it, and they're flawless.

I'll continue looking around ... maybe there's a good stand-alone converter I can use without too much hassle. But I sure wish Word could still do it ...

[Hagar Delest]

Weird.

You can try to create another Windows user on your machine and check with this new configuration.

[JonsBlog]

"XHTML export works for me with AOO 3.4 under Linux (32 bit) and Java 1.7."

I'm sure that's true, Villeroy, but there are two major differences between our versions of OO: operating system and bit level. You're running Linux, I'm running Windows. Your environment is 32-bit, mine is 64. Thus I've suspected all along that the error surfaced in porting the source code to a Windows executable.

[Villeroy]

Right now I exported a few lines from Writer to XHTML using LibreOffice 3.5 and AOO3.4 on Win7(32). I won't install anything experimental on our 64bit server.

[JonsBlog]

"Then the problem is with your setup since HTML export works for others on this list."

Thanks, and fair enough. But are those others running it specifically in a Windows 64-bit environment? Methinks this release may well need further debugging on that specific platform.

Note that during the five-minute interval during which I deleted the fully working version of OO, and installed this latest update, nothing else changed on my PC. The environment remained identical.

[Villeroy]

There is no special 64 bit version for Windows. Windows is supposed to run any 32bit application as is. I doubt that Microsoft will fix the issue if it affects this application only.
Since Java is involved: Are you sure that your Java is a 32 bit version?

[JonsBlog]

Yes, I know there is no 64-bit version of OO, and that Windows is SUPPOSED to run 32-bit applications just fine. But while I have never before encountered such an incompatibility issue myself, and my previous version of OO ran perfectly on this 64-bit Windows platform, I have heard of other 32-bit programs which do indeed have problems on the Windows 64-bit platform. Thus I suspect that assuming OO will run in the Windows 64-bit environment without actually testing it on that specific platform may be the issue here.

Fact is, I rarely install new software at all ... if it works, I keep using it, and I only attempted to update OO (which I now deeply regret) because Secunia PSI alerted me to a possible security issue with that older version of OO that I was running, and provided a direct link to this current version.

You ask if my Java is a 32-bit version. I asked here how to determine that earlier, but nobody described a way to do that. So just now I have gone to the Oracle/Java web site and attempted a download. It tells me "You have the recommended Java installed (Version 7 Update 4)". I assume that is the 64-bit version. So I dig a bit there, and find the current 32-bit version. I download and install that.

Next, I delete and reinstall OpenOffice from scratch. I open it, do "Tools > Options > Java" and find three versions of Java there:

Sun Microsystems Inc. 1.6.0_31
Sun Microsystems Inc. 1.6.0_32
Oracle Corporation 1.7.0_04

Since I've just downloaded Java from the Oracle web site, I assume the correct version to choose is Oracle 1.7.0_04, and I do.

I now create a one-line text document and attempt to export it as HTML. I still find only the .PDF option under "Export" (no HTML option as was available in the earlier version). So I use the "Save As" option, and select HTML. Same problem. The resulting document has no margins at all. The text sits at the far left edge of the document.

So I go back and select each of the other two Java versions. No change. Same problem with all three.

[Villeroy]

I just had to give a try: http://user.services.openoffice.org/en/ ... 16#p236816
Windows(64) requires Office(32) + Java6(32)
Windows(32) works with Office(32) + Java7(32) too.
XHTML is disabled until you enable support for legacy file formats (re-run the installer).

[JonsBlog]

Thanks, Villeroy. I am, in fact using a 32-bit version of Java. I just installed it, as mentioned earlier.

At your suggestion, I have now reinstalled OO with the "Optional Components/Legacy Filters" option enabled.

No improvement.

Would you please do this test for me:

Create a simple one- or two-line text file in Write, then attempt to export it as HTML.

When you select "Export", do you see TWO options in the drop-down list, one for PDF and one for HTML? If so, choose HTML, export the document, and then open it in a browser. Did the export function retain your default Word left margin setting, or is all the text bunched up at the left side of the page?

As I mentioned, I have never once seen an HTML export option under "Export" in this latest version, so the only remaining option is "Save As", and that destroys margin settings (and some other page formatting). So if you don't see that option under "Export" either, try the one under "Save As", and let me know whether it preserves document margins.

I just now opened up the source code for an HTML page correctly exported from an older version of OO, and compared it to these newly failing ones. The difference is obvious ... the older ones contain a "table" directive (necessary for preserving the appearance of page margins in HTML), while these created by this present version do not.