I have now successfully run the test file in the latest version of the Stylesheets/pureodd branch to completion. (pause for applause and outbreak of cheerful emoticons) However, I noticed the following oddity (ha ha) : test34 would fail in the odd2odd process with a message claiming (wrongly) that the input file was not available. Running the exact same test at the command line (i.e. not in the makefile) would work fine. Re-running the Makefile but starting a bit further on (e.g. saying "make test-oddity" rather than "make test") would also work fine until the next time the failure occurred. I hazard a guess that there is some kind of memory leak issue here. Certainly the Makefile looks excessively repetitive and long and is in need of ANT-ifying. This may also explain why Mr Jenkins has been having problems completing the test. I append my command line history in case others can spot a pattern I've missed: 2090 make test 2091 teitornc test34.odd test34.rnc 2092 cd Test 2093 teitornc test34.odd test34.rnc 2094 make test-namespaces 2095 teitohtml --odd --summaryDoc test34.odd test34.odd.html 2096 make test-namespaces 2097 java -Djava.awt.headless=true -jar ../lib/saxon9he.jar test34.odd ../odds/odd2odd.xsl > test34.combined.odd 2098 make test-namespaces 2099 make test-from-html 2100 make test-latex test-fo test-to-docx test-from-docx test-xlsx test-rdf test-text test-odt test-markdown test-cocoa test-epub