Moved the generation of the release metadata entry for the MD5 zip outside of
the parallel section, so it can't co-incide with the zipping of the binaries.
Ensured that any errors generated when merging log files are not just hidden
by putting them in the output file (which will render it not well-formed XML).
<?xml version="1.0" encoding="utf-8"?>
<diamonds-build>
<schema>10</schema>
<#assign raptor_errors=0/>
<#assign raptor_warnings=0/>
<#list raptor_summary as raptor_item>
<#if raptor_item.category = "raptor_error">
<#assign raptor_errors=raptor_errors+1 />
<#elseif raptor_item.category = "raptor_warning">
<#assign raptor_warnings=raptor_warnings+1 />
</#if>
</#list>
<faults>
<total severity="Raptor Errors">${raptor_errors}</total>
<total severity="Raptor Warnings">${raptor_warnings}</total>
<total severity="Recipes Failed">${ant['sf.job.totalyarperrors']}</total>
</faults>
<files>
<file>
<type>log</type>
<name>BRAG Summary</name>
<url>file:///${ant['sf.spec.publish.networkdrive']}\${ant['sf.spec.job.name']}\builds\${ant['sf.spec.job.codeline']}\${ant['build.id']}\summary\_brag.xml</url>
</file>
<file>
<type>log</type>
<name>Raptor summary</name>
<url>file:///${ant['sf.spec.publish.networkdrive']}\${ant['sf.spec.job.name']}\builds\${ant['sf.spec.job.codeline']}\${ant['build.id']}\logs\raptorbits\summary.csv</url>
</file>
<#list files as f>
<file>
<type>log</type>
<name>${f.name}</name>
<url>file:///${f.path}</url>
</file>
</#list>
</files>
</diamonds-build>