Ticket #291 (closed defect: fixed)

Opened 14 years ago

Last modified 12 years ago

[M] Unescaped ampersands in CEDAR DIF

Reported by: lawrence Owned by: lawrence
Priority: critical Milestone: Replace Metadata Gateway
Component: discovery Version:
Keywords: DiscoveryService Cc:

Description (last modified by selatham) (diff)

Our CEDAR DIF records might contain & in the urls (for A-services ... but maybe elsewhere). These need escaping appropriately (&) ... and corresponding service binding code may also need modification.

Change History

comment:1 Changed 13 years ago by selatham

  • Status changed from new to assigned
  • Description modified (diff)

ampersands are being escaped in descriptions. Need to do this in URLs too.

comment:2 Changed 13 years ago by ko23

In which fields? Note that if you escape ampersands/URLs/etc., then you need to unescape them consistently later, even if one uses the standard functions.

comment:3 Changed 13 years ago by selatham

If they are not escaped then they won't go into exist. So we need to do something.

comment:4 Changed 13 years ago by selatham

  • Keywords XSLT removed
  • Milestone changed from PreBeta to BETA
  • Summary changed from [M] Unescaped ampersands in DIF to [M] Unescaped ampersands in CEDAR DIF

This is linked to service binding issues.

comment:5 Changed 13 years ago by selatham

  • Milestone changed from BETA to BETA+Services

We've changed our service binding to late-binding so may not be such a problem. But could still happen. Need to think it through.

comment:6 Changed 13 years ago by selatham

  • Keywords service_binding added

comment:7 Changed 13 years ago by selatham

  • Milestone changed from PROD Step2 to PROD Final

service binding has gone back to early-binding, so could be a problem.

comment:8 Changed 13 years ago by selatham

  • Owner changed from selatham to lawrence
  • Status changed from assigned to new
  • Component changed from T01_Discovery to DiscoveryService
  • Milestone changed from PROD Final to Replace Metadata Gateway

Ampersands are being escaped in URLs now when creating MOLES/DIF XML. This IS causing a problem following links. e.g. any neodc record  http://www.neodc.rl.ac.uk/?option=displaypage&Itemid=132&op=page&SubMenu=-1

comment:9 Changed 13 years ago by spascoe

Does this issue effect all DIFs ingested into the DiscoveryService? or only those created by BADC?

I think ampersands need to be escaped in element bodies in order to be valid XML, therefore DIF consumers (such as the NERC-Portals portal) will just need to deal with unescaping &

Note the DIF provided by Jon Blower for NERC-Portals already has "&" escaped correctly.

comment:10 Changed 13 years ago by selatham

  • Priority changed from desirable to critical

Actually you can see that Jon's record has escaped ampersands. I was only using the neodc record as an example which I had noticed. I'm assuming escaped things in URLs in general aren't being handled properly by the NDG Discovery/Browse? front-end.i.e. we need to unescape on display. But as NERC Portals are only using the SOAP WS then they will just get the XML records and deal with unescaping themselves. So I think it's really only a problem for our front-end.

comment:11 Changed 13 years ago by lawrence

  • Status changed from new to closed
  • Resolution set to fixed

Well, I've hacked a fix into changeset:3159. Good enough for now.

comment:12 Changed 12 years ago by lawrence

  • Keywords DiscoveryService added; service_binding removed
  • Component changed from DiscoveryService to discovery

Moved from DiscoveryService? component to discovery as part of NDG2 cleanup

Note: See TracTickets for help on using tickets.