Nathan Wertman - 09/Apr/08 12:37 AM This is actually a pretty significant issue for us. It is hitting a build which tests a large portion of shared code in the ocunit.pp. Turning off the ocunit pp is pointless because the whole point of this build is to run tests on this shared code (it doesn't produce any artifacts)

jason - 09/Apr/08 10:55 AM This issue is a bit more complicated than face value: we are escaping XML special characters in our test data. However, there are some characters (notably most ASCII control characters) that cannot be represented in XML at all. The usual escaped form (i.e. &#x<code>;) is not even allowed for these characters in XML 1.0. The only way to preserve them in our test data is to add our own extra layer of encoding on top. As this is an exceedingly rare case, we also don't want to encode all test data all the time. Further, I have some concern that the reason this character is showing up could be due to a character encoding issue. If not, it is an odd character to spit out as part of test output. I guess the test could be exercising code against special characters and some assertion is just spitting them out as diagnostics.

jason - 09/Apr/08 12:12 PM Fixed in change 4283.

jason - 09/Apr/08 12:14 PM Merged to trunk in change 4284.