HTML Attachment Translation

adobrin · Post by **adobrin** » Mon Nov 21, 2005 3:27 pm

I set Zimbra to translate attachments to HTML for viewing, and am now recieving this error:
Error occurred during document conversion
* Reason: com.zimbra.cs.convert.ConversionException: Conversion error

when attempting to open a PDF file.
Is the feature not intended for use yet?

14319KevinH · Post by **14319KevinH** » Mon Nov 21, 2005 4:48 pm

It is ready and should work(I assume your using the Network version, since this is not a feature in the OSS version). There should be a convertd.log file that may have more info. Do you see anything in there?

adobrin · Post by **adobrin** » Mon Nov 21, 2005 4:52 pm

2005-11-21 17:43:00,591 DEBUG [Convertd-1] [] TransformationHandler - filename= att971392.jpg

2005-11-21 17:43:00,601 WARN [Convertd-1] [] VerityEngine - Verity filter error

com.verity.api.filter.FilterException: Error loading shared library kvfilter.

at com.verity.api.filter.Filter.getStreamDocFormatInfo(Native Method)

at com.verity.api.filter.Filter.getDocFormatInfo(Unknown Source)

at com.zimbra.cs.convertd.VerityEngine.extract(VerityEngine.java:233)

at com.zimbra.cs.convertd.TransformationHandler.doTextExtraction(TransformationHandler.java:194)

at com.zimbra.cs.convertd.TransformationHandler.processCommand(TransformationHandler.java:102)

at com.zimbra.cs.tcpserver.ProtocolHandler.processConnection(ProtocolHandler.java:230)

at com.zimbra.cs.tcpserver.ProtocolHandler.run(ProtocolHandler.java:197)

at EDU.oswego.cs.dl.util.concurrent.PooledExecutor$Worker.run(Unknown Source)

at java.lang.Thread.run(Thread.java:595)

2005-11-21 17:43:00,603 DEBUG [Convertd-1] [] VerityEngine - deleted /opt/zimbra/convertd/inprocessing/8lylREPq27qMS+tlFxawJ7S0c6c=_3

2005-11-21 17:43:00,607 DEBUG [Convertd-1] [] TransformationHandler - processing command (x) from client at remote port 38961

2005-11-21 17:43:00,608 DEBUG [Convertd-1] [] TransformationHandler - filename= att971393.gif

2005-11-21 17:43:00,608 INFO [Convertd-1] [] VerityEngine - filtering document out-of-process: 8lylREPq27qMS+tlFxawJ7S0c6c=_4

2005-11-21 17:43:16,378 INFO [main] [] TransformationServer/7047 - Convertd is ready

and
2005-11-21 17:42:59,485 DEBUG [Convertd-2] [] TransformationHandler - processing command (x) from

client at remote port 38957

2005-11-21 17:42:59,497 INFO [Convertd-2] [] VerityEngine - found unfinished doc whose hash is: ,Nr9ND9pZ8eBi9+hxpqJbpDquvI=_2

2005-11-21 17:42:59,497 INFO [Convertd-2] [] VerityEngine - found unfinished doc whose hash is: 74twdNPLBSuMRdq9bp1iesW7sRk=_2

2005-11-21 17:42:59,497 INFO [Convertd-2] [] VerityEngine - found unfinished doc whose hash is: o4o6R1bnMI4wtEJJWsaG8jdxeSI=_1.2

2005-11-21 17:42:59,497 INFO [Convertd-2] [] VerityEngine - found unfinished doc whose hash is: HNVTcrN3LybeyCRr4XLitjo2yvo=_1.2

2005-11-21 17:42:59,497 INFO [Convertd-2] [] VerityEngine - found unfinished doc whose hash is: wt8A83mg3wmat+pmHVmuemoj4zA=_2

2005-11-21 17:42:59,498 INFO [Convertd-2] [] VerityEngine - found unfinished doc whose hash is: 8lylREPq27qMS+tlFxawJ7S0c6c=_4

2005-11-21 17:42:59,498 INFO [Convertd-2] [] VerityEngine - found unfinished doc whose hash is: J3W3s,daLPr0x0njU3v1fnKtTFM=_2

2005-11-21 17:42:59,502 DEBUG [Convertd-2] [] TransformationHandler - filename=

2005-11-21 17:42:59,512 WARN [Convertd-2] [] VerityEngine - Verity filter error

com.verity.api.filter.FilterException: Error loading shared library kvfilter.

at com.verity.api.filter.Filter.getStreamDocFormatInfo(Native Method)

at com.verity.api.filter.Filter.getDocFormatInfo(Unknown Source)

at com.zimbra.cs.convertd.VerityEngine.extract(VerityEngine.java:233)

at com.zimbra.cs.convertd.TransformationHandler.doTextExtraction(TransformationHandler.java:194)

at com.zimbra.cs.convertd.TransformationHandler.processCommand(TransformationHandler.java:102)

at com.zimbra.cs.tcpserver.ProtocolHandler.processConnection(ProtocolHandler.java:230)

at com.zimbra.cs.tcpserver.ProtocolHandler.run(ProtocolHandler.java:197)

at EDU.oswego.cs.dl.util.concurrent.PooledExecutor$Worker.run(Unknown Source)

at java.lang.Thread.run(Thread.java:595)

several lines above that there are a bunch of lines like

"Transformation Handler - Processing Command (^H) from client at remote port"
over and over, with a different character (it looks like an escaped control code) in the parenthesis.

14319KevinH · Post by **14319KevinH** » Tue Nov 22, 2005 1:30 am

Think some C++ libs are missing. Can you try installing these pkgs:
compat-libstdc++-296

compat-libstdc++-33
You can use up2date

adobrin · Post by **adobrin** » Tue Nov 22, 2005 10:10 am

Thanks, that solved the problem.

the PDF conversion isn't "perfect" though; and it would be helpful to be able to still download the files. a context menu maybe?
by not perfect, it seems that some font and style changes cause overlapping in the HTML.

6439kchen · Post by **6439kchen** » Tue Nov 22, 2005 12:55 pm

PDF rendering is performed by Verity software from our third party partner. We'll work with them to enhance it in their future releases.

6149Richard · Post by **6149Richard** » Thu Dec 22, 2005 2:42 pm

If you check "translate attachments for viewing", is there any way to retrieve and save the original attachment? We would like to allow people to view the attachments in Zimbra using html, but then save the actual document. A spreadsheet is much less functional as html.

Somewhat related, if you do not check "translate attachments for viewing", are the attachments still indexed? Thanks.
Richard

14319KevinH · Post by **14319KevinH** » Thu Dec 22, 2005 4:31 pm

[quote user="6149Richard"]If you check "translate attachments for viewing", is there any way to retrieve and save the original attachment? We would like to allow people to view the attachments in Zimbra using html, but then save the actual document. A spreadsheet is much less functional as html.

Somewhat related, if you do not check "translate attachments for viewing", are the attachments still indexed? Thanks.
Richard[/QUOTE]
Currently we don't expose the original (although it should still be saved). Seems like we need an additional choice here. Since one use case covered by the original design is to prevent raw attachments from getting to end users.
Yes the attachments are indexed in all cases.

schemers · Post by **schemers** » Thu Dec 22, 2005 6:29 pm

We are planning on adding the ability to view attachments in HTML in the default case. i.e., you could download a word doc or view it as HTML (as long as the admin hasn't turned on view all attachments in HTML).

6149Richard · Post by **6149Richard** » Fri Dec 23, 2005 12:31 am

[quote user="schemers"]We are planning on adding the ability to view attachments in HTML in the default case. i.e., you could download a word doc or view it as HTML (as long as the admin hasn't turned on view all attachments in HTML).[/QUOTE]
Thanks. That would be precisely what we were looking for.
Richard