The VulcanJ test ware should be ready for general use now. I ran the sourceforge copy of the test ware last night against our build of Vulcan using S64, Windows XP, MVS, Linux, and R64 (AIX). The test ware worked on all hosts, although not all of the were clean against Vulcan. :) I pushed some corrections this morning, to support FB2.0.
I ran VulcanJ against Firebird 2 on Windows this morning, using the snapshot build, Firebird Version: WI-T184.108.40.20611 Firebird 2.0 Alpha 1. It almost ran clean. There was a problem with CHECK constraints (4 tests) that was reported here earlier and that Dimitry was going to address. The only other problem is an issue where an extra null seems to be appended to error messages when using Jaybird (that I also see with Vulcan, but not FB1.5). I haven't been able to reproduce this problem outside of the Java environment. So, I had 5 failure/errors in total with the FB2.0/Windows.
I have added Adriano's request to include SET NAMES ASCII in some of the ISQL tests for internationalization reasons, and corrected the binary/ASCII problems from the initial push. And fixed Adriano's comment that output directories were not created properly when they did not exist. Thanks, Adriano!!!
I discussed privately with Pavel where to put the test ware, and for the time being, it will stay at firebird/VulcanJ in the source tree where it was initially pushed. Not ideal, but ok for now. Pavel is welcome to move it to where he sees fit later.
Please let me know your experience with the test ware, and if you find any problems.
Product Specialist, Base SAS
Friday, April 15, 2005
Posted by Fikret Hasovic at 4/15/2005 10:30:00 AM