Oxygen XML Forum

I installed XEP RenderX and it was working for PDF output from DITA, but every other time or so no PDF would appear. Now, I'm getting no PDF output each time, even though "Build Successful" appears and "exit code 0".

I see this message, can anyone tell me what it is or at least where to investigate?
Is this referring to my bookmap or to a configuration file? The map and topics I'm trying to generate show no errors when I validate in Oxygen.

I also have this which seems less likely to be the culprit (lacking the word "fatal" in it for one thing) :
I investigated the file flag-rules.xsl and found no variable without content with the opening and closing variable tags, at least nothing obvious like a <variable> </variable> with nothing inside it.

Thanks,

Hello,

H_Too_Oh wrote:I installed XEP RenderX and it was working for PDF output from DITA, but every other time or so no PDF would appear. Now, I'm getting no PDF output each time, even though "Build Successful" appears and "exit code 0".

I see this message, can anyone tell me what it is or at least where to investigate?

Code: Select all

transform.fo2pdf.xep:

     [java] [Fatal Error] xep.xml:169:4: The markup in the document following the root element must be well-formed.

     [java] Could not create XEP formatter: com.renderx.xep.lib.ConfigurationException: Structure error in configuration file: org.xml.sax.SAXParseException: The markup in the document following the root element must be well-formed.

The intermediary XSL-FO file from which the final PDF output must be created is not well-formed. Please send us a zip archive with a working sample (a simple DITA map, the topics from the map, any other necessary resources for the PDF transformation like image files, etc.) for reproducing the error.

H_Too_Oh wrote:I also have this which seems less likely to be the culprit (lacking the word "fatal" in it for one thing) :

Code: Select all

transform.topic2fo.flagging.filter:

     [xslt] Processing /spf-doc/spf-doc (trunk)/temp/stage1.xml to /spf-doc/spf-doc (trunk)/temp/stage1a.xml

     [xslt] Loading stylesheet /Applications/oxygen/frameworks/dita/DITA-OT/demo/fo/xsl/fo/flagging-preprocess.xsl

     [xslt] /Applications/oxygen/frameworks/dita/DITA-OT/demo/fo/xsl/fo/flag-rules.xsl:198:30: Warning! A variable with no following sibling instructions has no effect

I investigated the file flag-rules.xsl and found no variable without content with the opening and closing variable tags, at least nothing obvious like a <variable> </variable> with nothing inside it.

That XSLT warning has nothing to do with the well-formedness of the XSL-FO file created by the XSLT stylesheets. Please send some sample files for reproducing the error.


Regards,
Sorin

Thanks for the reply,

I changed to using the Apache FO for now and it's actually doing everything I need, so because of time constraints will just stay with that. I'll look into using Render X again if I have time later. If I have a pause I'll assemble a small sample doc and reproduce the error using Render X if I can and send it to you, the only one I have now is the whole doc which is proprietary.

Thanks again,

Hi,

The error obtained when using RenderX XEP:
seems to indicate that the xep.xml configuration file used by RenderX is not wellformed. Did you make any manual changes to it?
You could try to open it in Oxygen and fix it (it is usually located in the XEP installation directory).

Regards,
Radu

Just getting back to this, thanks for the reply.

I didn't edit that xep.xml file, no, the only one on my computer has this at line 169:

<fonts xmlns="http://www.renderx.com/XEP/config"
xml:base="fonts/"
default-family="Helvetica">

Does that look right? The URL leads to a "file not found" page at XEP, and I can't sort out at their site where it seemed to want to be pointing. Just an outdated link perhaps?

I'll make a simple test map and send it to you at Oxygen in the meantime.

Hi,

Yes, having some samples would be great. Please also attach the xep.xml to the email.
If you open the xep.xml in Oxygen, does Oxygen report any validation problem?

Regards,
Radu

I just realized that I could open the xml file in Oxygen and validate it (I was using a text editor to look at it) and finally found the problem, after chasing around. This line at the top was commented out:
which led to that <font> tag being pointed to I guess, but the last tag in the file was </config> so clearly that needed to be the opening tag at the top.

I didn't comment it out, maybe it was just a way of deprecating it or something because that's the same URL that doesn't point to anything.

Anyway, now it all works.

Thanks very much,

This seems to be acting up again, or at least the PDF build process is, it builds successfully about every third time but only after finishing several times with a successful build code but no PDF opening in the Web Browser, and if I open the PDF directly it gives a "can't be opened, damaged" message.

I'll submit the output log file to see if you can spot something? I don't see any errors, only one warning that seems unrelated partly because it was there before when it built a PDF every time.

Thanks,

Hi Jeffrey,

Thanks for the logging information.

The cause of the problem as reported in the console seems to be this one:

transform.fo2pdf.xep:
[java] Rendering failed: Can not close compressed stream

I have no idea what it means, somehow XEP fails to produce the PDF because an internal error occurs.

I found a very similar post here:

http://lists.renderx.com/pipermail/xep- ... 06954.html

What I can do further would be to get in touch with some contacts we have on the RenderX XEP side, share this logging information with them and then possibly obtain from them more details, maybe a XEP installation which contains additional debug information.

You can contact them also directly by using the "support at renderx.com" email address.

So just tell me your decision.

Regards,
Radu

Hi Radu,

Okay thanks, I had discounted that because I have another map that seems to display the same error but produces a PDF every time-- I'll have to double check if it actually gives that error each time.

It's very spotty/buggy in any case, the map I'm having trouble producing sometimes does, sometimes doesn't, and if I remove all of the content except the front and back matter, then at first it works, then putting the content back works... then it doesn't again.

Right now for instance the PDF is produced only in the Web browser, the PDF file itself doesn't appear in the output directory, so if I save the PDF from the browser, I get the PDF. That's new today.

It actually sounds to me like the error you get if the old PDF is still open so it can't replace it with the new one, but checking the processes on my Mac I don't see Preview open and I've tried closing Chrome to be sure the browser UI isn't still hanging onto the old PDF somehow.

I'll contact XEP and see what they say, thanks very much.

PS: There isn't some other transformation that you know of that produces an index, is there? I'm a little wary of Render X since it seems so off and on, right now is okay but once I'm sending out these docs, suddenly not being able to produce PDFs would really be a problem. What Apache produces looks fine to me, the lack of index is the only problem.

Just an update: I tried just reinstalling XEP, so far so good. Now everything works. There was an existing installation and a new one over it so maybe things got mixed, I just wiped it off and started over. Thx-

Well, reinstalling RenderX hadn't solved it after all, as you know from our email exchange when I sent the sample.

I decided that it just had to be something wrong with the map, since my other ditamap builds fine no matter what. I just kept experimenting with the problem map, tossing parts of it out and trying it that way.

Then the "results too large" part of the error message got me wondering. One difference between the maps is that this one had only one <part>, with all of the chapters inside the one part. This is just for design reasons, it's sort of a reference doc and I didn't want or need a lot of separation.

So I separated out the map into three or four parts instead, and now it builds PDFs using XEP again.

My fingers are crossed but that seemed to have been the problem, something about the XEP engine seems to prefer more than one part per map; the Apache FOP conversion was fine with it.

Hi Jeffrey,

The console output which you sent to our support email address did not contain any message like results too large so this might be a different problem than the initial one.
Apache FOP and XEP are not applied directly on the DITA Map, they are applied on an intermediary output in the XSL-FO format.
I cannot give you any details, maybe you could send me the console output when the results too large is encountered.

Regards,
Radu

Actually the "results too large" was in the error message that you pointed me to that was similar to mine, I should have been more clear about that. I had seen that on the Web also actually, and it was just a clue, since the error was so similar to the one I was getting.

I had a response from XEP, which I appreciated, that talked about another error message I had sent them:

> transform.fo2pdf.xep:
> [java] Could not create XEP formatter: org.xml.sax.SAXParseException:
> The processing instruction target matching "[xX][mM][lL]" is not allowed.


and saying that it was the result of not well-formed XML, i.e. possibly the opening tag had a space in front of it. As I responded to them though, ever since I broke the bookmap into more than one part, everything has worked, without changing any opening tags, and the above error message is gone now also.

As I say I had another bookmap that has never had any problems generating a PDF, and the only difference I could see was that it had mutiple parts instead of one long one. So that might not have been the problem but everything seems to work for now in any case.

Thanks again,

I had a similar issue with a map building successfully but the PDF not actually appearing. I narrowed down the issue to a specific table in one topic. I replaced the colspec lines with another set from a table that did work and the PDF build and displayed. The log was full of warning/error messages about:

• content overflowing the viewport

• table-layour="auto" as a feature not yet implemented by Apache FOP

• org.apache.fop.fo.properties.PropertyMaker - forcing opt to min in LengthRange

Don't know if these are related to your issue, but I thought it might help with the debugging.
Karen