OAI Identify verb XML parsing error

We have a problem with the configuration of OJS 2.4.8.2 for the magazine Revisiones (http://www.re-visiones.net/ojs) related with the harvest OAI system.

The instalation is in a hosting con Linux with php 5.6 and MySQL 5.5 in a directory /ojs hanging from root directory where the pages for the main domain are, You can access the ojs magazine at www.re-visiones.net/ojs address.

In general everything is working properly, you can set up magazines with differente numbers and articles without problema and the editorial workflow works fine.

The problem arise when you try to use the harvest OAI systeme. For some unknown reason Identify verb is not working properly and give an XML parsing error:

http://www.re-visiones.net/ojs/index.php?journal=RE-VISIONES&page=oai&verb=Identify

The error says:

XML Parsing Error: not well-formed
Location: http://www.re-visiones.net/ojs/index.php?journal=RE-VISIONES&page=oai&verb=Identify
Line Number 11, Column 77: http://www.re-visiones.net/ojs/index.php?journal=RE-VISIONES&page=oai
------------------------------------------------------------------------------------------^

Intead the rest of the verbs of the system work properly:

http://www.re-visiones.net/ojs/index.php?journal=RE-VISIONES&page=oai&verb=ListRecords&metadataPrefix=oai_dc
http://www.re-visiones.net/ojs/index.php?journal=RE-VISIONES&page=oai&verb=ListSets
http://www.re-visiones.net/ojs/index.php?journal=RE-VISIONES&page=oai&verb=ListMetadataFormats
http://www.re-visiones.net/ojs/index.php?journal=RE-VISIONES&page=oai&verb=ListIdentifiers&metadataPrefix=oai_dc

I’ve been trying to figure out what is the problem but I didn’t success in it. Any Clue

Hi @kaluosi,

It looks like you’re operating with path_info_disabled set to true in config.inc.php – this is intended to work around web servers that don’t permit URLs using the CGI PATH_INFO variable, but it has side-effects. URLs are a lot uglier in this mode and are not RESTful, and the OAI interface will operate outside of specification.

I would recommend setting up your server so that you can turn off path_info_disabled if possible.

Regards,
Alec Smecher
Public Knowledge Project Team

Thank you asmecher!!

I have put disable_path_info = Off (there is not a “path_info_disable”
parameter. Is there?) but then the whole ojs give an “404 Not Found”
error in every page. Maybe some other thing is wrong in the configuration.

Hi @kaluosi,

Your web server doesn’t currently support the standard CGI PATH_INFO variable, which is why the work-around is necessary in OJS. I would suggest working with your service provider to get PATH_INFO URLs working in PHP; the details will depend on your server so I’m afraid I can’t give you specific advice.

Regards,
Alec Smecher
Public Knowledge Project Team

Thanks again asmecher. I will check that fact related with CGI PATH_INFO and will tell you about.

Hi Alec,

In the hosting where we have OJS is not possible to set AcceptPathInfo even in .htacess. So I’ve been checking what you say in another server with the same OJS and with AcceptPathInfo activated (I have proved it with a php script) but exactly the same thing continue to happen.

http://www.laurisilva.net/ojs

With disable_path_info = Off we gete a “404 Not Fond” in every page.

With disable_path_info = On every thing works fine but OAI verb Identify still give the same error of XML parsing

http://www.laurisilva.net/ojs?journal=index&page=oai&verb=Identify

I’m completely lost

Hi @kaluosi,

You have two options:

  • Configure your server to accept PATH_INFO URLs, or move to a server that does accept them, or
  • Use disable_path_info = On, but accept that you’ll have ugly URLs and potential trouble with the OAI interface.

Regards,
Alec Smecher
Public Knowledge Project Team