[Archivesspace_Users_Group] Archivists Toolkit incompatibility with Java 8 (AR-1157)

Wed Feb 3 11:54:51 EST 2016

Hi Mark,
We are considering triggers, but as you say, we'd much prefer fixing it in the application.  We have a contractor looking at this and he thinks it might have something to do with JAXB? Namely that some parameters that AT passes to JAXB may be incorrect or that the JAXB version may not be compatible with java8.  But, we're hoping someone might know if we're barking up the right tree.

Re: what happens in v. 15, java8.  We end up with problems with both EAD imports and EAD exports.

1.       EAD exports add "ns2" namespace prefixes throughout. As in,
<ns2:ead xsi:schemaLocation="urn:isbn:1-931666-22-9 http://www.loc.gov/ead/ead.xsd" xmlns:ns2="urn:isbn:1-931666-22-9" xmlns:ns1="http://www.w3.org/1999/xlink" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
  ...
<ns2:acqinfo id="ref3">
            <ns2:head>Provenance</ns2:head>
            <ns2:p>The Oscar Bluemner papers were donated in several installments ...</ns2:p>
        </ns2:acqinfo>
...

2.       EAD imports end up with extra opening paragraph tags and with (some) ns2 prefixes.

* [Both v 15 and 16 do this] with at the opening of each paragraph.

*         [v 15 only] with "ns2" prefixes on Finding Aid Data tab tags such as num, date, item, language.  (yet, a num within odd/p or scopecontent/p doesn't get the ns2. So perhaps archdesc is ok, but not the eadheader...? Just a guess there... Moot point for us since ATv15 java 7 and ATv16 java 8 don't do this. )

AT then drops those problem tags on export. The paragraphs all export as blank . And, for example:
EAD import
<langusage>Finding aid written in <language langcode="eng">English</language>.</langusage>

AT user interface:
Finding aid written in <ns2:language langcode="eng">English</ns2:language>.

Exports as:
<ns2:langusage>Finding aid written in  .</ns2:langusage>

Nancy

Nancy Kennedy
EAD Coordinator
Smithsonian Institution
kennedyn at si.edu

From: archivesspace_users_group-bounces at lyralists.lyrasis.org [mailto:archivesspace_users_group-bounces at lyralists.lyrasis.org] On Behalf Of Custer, Mark
Sent: Wednesday, February 03, 2016 8:44 AM
To: Archivesspace Users Group <archivesspace_users_group at lyralists.lyrasis.org>
Subject: Re: [Archivesspace_Users_Group] Archivists Toolkit incompatibility with Java 8 (AR-1157)

Nancy,

What happens when you use AT version 15 with java 8, just out of curiosity?

As for the import bug in AT version 16, is this the only issue with the upgrade?  If so, I remember that I used to have a horrible hack to import barcodes into the AT, which just involved running an SQL script after every import (perhaps this could be triggered easily, without changing the core code, too?) that would move the barcodes (imported in the @label attribute) to the barcode field, and then change the label field to "Mixed materials".  Assuming something like that could work, you could try something like this in a test instance of the AT as a temporary workaround:

UPDATE ArchDescriptionRepeatingData
SET noteContent = REPLACE(noteContent, '<p xmlns="urn:isbn:1-931666-22-9" xmlns:ns2="http://www.w3.org/1999/xlink">', '')
WHERE noteContent LIKE '%<p xmlns="urn:isbn:1-931666-22-9" xmlns:ns2="http://www.w3.org/1999/xlink">%';

The AT never stored paragraph tags in the database before, so you'd definitely want to remove those in some way (although fixing it in the application would be best, of course).  I suppose if it had the closing tag it might export okay, but I wouldn't try to add those just for the sake of the exports, since you'll need to migrate the data.

Mark

From: archivesspace_users_group-bounces at lyralists.lyrasis.org<mailto:archivesspace_users_group-bounces at lyralists.lyrasis.org> [mailto:archivesspace_users_group-bounces at lyralists.lyrasis.org] On Behalf Of Kennedy, Nancy
Sent: Tuesday, February 02, 2016 1:02 PM
To: 'archivesspace_users_group at lyralists.lyrasis.org'
Subject: [Archivesspace_Users_Group] Archivists Toolkit incompatibility with Java 8 (AR-1157)

Hello all,
Have any other users implemented the latest AT version 16, for java8 compatibility, per https://archivesspace.atlassian.net/browse/AR-1157<https://urldefense.proofpoint.com/v2/url?u=https-3A__archivesspace.atlassian.net_browse_AR-2D1157&d=AwMFAg&c=-dg2m7zWuuDZ0MUcV7Sdqw&r=s7ciGQfUJeaV_ryx908hbeXDoU9aqDwDN0Z0VbfsJ3Y&m=RKhZyIOCP-rCH87AMz91_Q89CrLKSfiS9bo_ojn3T_4&s=4aOProrU86-3_tlu9ynrV2M5T41DrVQwN6bIr79XWDY&e=>?

While we work towards our ArchivesSpace migration, we need to be able to use the AT with java 8 to address security vulnerabilities. While the latest AT version 16 solves the ns2 namespace export issue, our EAD imports are not working correctly.

Upon EAD import, note fields all end up with extraneous "" at the start of each paragraph. Note that there is no , which seems to result in AT dropping this data during EAD export. (AT exports these notes with empty paragraphs ). This starts to get very messy if notes are multi-paragraph or if any additional <title> and <emph> are involved. Has anyone else seen this?

Importing:
 <acqinfo>
      <head>Provenance</head>
      <p>The Oscar Bluemner papers were donated in several installments from 1970 to 1985 by John
        Davis Hatch, an art historian and close friend of Bluemner.</p>
    </acqinfo>

                                Results in AT data field:
<p xmlns="urn:isbn:1-931666-22-9" xmlns:ns2="http://www.w3.org/1999/xlink<https://urldefense.proofpoint.com/v2/url?u=http-3A__www.w3.org_1999_xlink&d=AwMFAg&c=-dg2m7zWuuDZ0MUcV7Sdqw&r=s7ciGQfUJeaV_ryx908hbeXDoU9aqDwDN0Z0VbfsJ3Y&m=RKhZyIOCP-rCH87AMz91_Q89CrLKSfiS9bo_ojn3T_4&s=bDiPBPnipPu8j-tTV-sktjwQnLZrqdBRUaI42yhJuEs&e=>">The Oscar Bluemner papers were donated in several installments from 1970 to 1985 by John Davis Hatch, an art historian and close friend of Bluemner.

Which then exports as:
<accessrestrict id="ref3">
            <head>Provenance</head>
            <p/>

Thanks for any insights -
Nancy

Nancy Kennedy
EAD Coordinator
Smithsonian Institution
kennedyn at si.edu<mailto:kennedyn at si.edu>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lyralists.lyrasis.org/pipermail/archivesspace_users_group/attachments/20160203/d92864cc/attachment.html>