I would like to know which version of the MS XML parser you have installed on your system. It would be of help if you could visit  this link http://www.bayes.co.uk/xml/utils/msxml_sniffer.htm and post the resulting table here.

If this fails in the same way I suggest you download & install MSXML 6.0 from http://www.microsoft.com/downloads/details.aspx?FamilyID=993C0BCF-3BCF-4009-BE21-27E85E1857B1. If however the sample files work correctly then I need to look into what is unusual about your file. Perhaps you could email it to me via GenoPro's 'Send File' facility or using the control panel in this forum.

Harry Potter.gno

Samplegenogram.gno

<Comment>Nom: BARRE - Prénom: jean&#10;Père Prénom: &gt;michel&gt;&#25;.chevalleri&gt;&#25;.</Comment>

Nom: BARRE - Prénom: jean
Père Prénom: >michel>.chevalleri>.
 

I suggest you export your .gno file to .xml format (File/Export/Export to XML Format..) and then edit the xml file replacing all occurences of  with some other character(s)  e.g. ';' say.

Then reopen the edited .xml file with GenoPro and all should be OK.

this is yielding an additionnal question/suggestion: in Gedcom generation (without report generator) you can choose character set (including ANSEL), but the generated gedcom do not carry properly unions.

When using report generator, character set is Unicode, which is most of time not correctly recognized by external genealogy software (but unions are correctly handled), is it possible to have a the gedcom report generator with a skin in ANSEL ??

Thanks

==== NOW OK ! ====

Should GenoPro do anything regarding this problem?

UTF-8 encoding is part of the Gedcom 5.5.1 draft standard and supported by several packages including PAF and Legacy Family Tree. You could always import your Gedcom to one of these and re-export as ANSEL.

However, conversion of Unicode to ANSEL appears to me to be extremely complex, and not something I would like to tackle. So I will bat this question on to Dan - GenoPro has a Unicode to ANSEL routine used in the built-in export to Gedcom, could this be made available in the Report Generator SDK? The SDK already has a method: 

The XML encoding module was not designed to report errors/warning to the user, as encoding of text should not produce any error.  Perhaps GenoPro change the header to XML version 1.1. 

I have never been able to find a table to convert ANSEL to Unicode (or vice versa).  The ANSEL encoding is one of the most cryptic encoding I have ever seen; it has no logic, no documentation, and no sample files.  What I did was installing other genealogy software, typed some text with accents and looked how those characters were encoded in the .ged file.  Worse of all, the ANSEL encoding is an extra [useless] layer on top of the current code page.  As a result, you may have English-ANSEL encoding, Russian-ANSEL encoding, Hebrew-ANSEL encoding, and so on.  When GenoPro imports a Gedcom file, it performs two decoding, one from the current language from the control panel and then the ANSEL decoding.  There are tons of bugs regarding ANSEL encoding because there is no clear specifications.  I would stay away from ANSEL encoding as much as possible.

The method BufferWriteToFile has indeed a parameter for encoding.  At the moment, the only valid value is GZip, however all the other encoding can be accepted.  I looked at the code and the code was explicitly commented out because this feature has not been tested to my satisfaction.  During the next beta phase, I will re-enable the code to support various encoding.  In the meantime, use UTF-8 encoding.

Support for XML 1.1 appears to be a little sparse and is not supported by any version of the MS Parser, therefore changing the header to 1.1 would not be helpful as errors would still occur. But I still think GenoPro should produce valid XML 1.0 so I would be happy if it silently converted invaid codes to some other character. I realise the inverted question mark is used in Spanish so might be a little confusing. Perhaps the currency sign ¤ would be a better choice as a widely available character that is not used much.

If GenoPro converts some characters to ¤ , then how GenoPro will read back those original characters?  Displaying ¤ for an unknown character is a good idea, however storing ¤ in XML will result in data loss. I want XML to be round-tripping, that is writing, reading and writing back to give the same result.

The other question is: how those illegal characters got there? The entity/character  which is the EM (U+0019) should probably be converted to a dash during the Gedcom import, not in the XML export.

I agree that the illegal character conversion should happen during at the Gedcom import.

1. Rename the .gno file to .zip.
2. Double-click on the .zip file and extract the file Data.xml.
3. Double-click on the file Data.xml. Internet Explorer will display the errors.
4. Using Notepad, remove the errors, that is all 
5. Open the file Data.xml using GenoPro.
6. Voilà.

Fixing the Gedcom import module will prevent further errors, however if you are getting errors generating a report, it is too late because the data is already in your .gno file.  The trick then is fixing the .gno file.