On Feb 24, 2011, at 1:56 PM, Christopher Cambly wrote:

> That's too bad, it worked for us on AIX 5.3. Are you using a vacpp > built bjam or a gcc bjam?

vacpp built bjam. I can try a gcc built one and see if the results are any different.

Yeah, just curious.

> I could be mistaken but I think we've dumped our AIX machines so I'm > no longer able to debug this for you. I'll check around and see if > we've still got access to an AIX box. I do wish you had mentioned > this a few years ago when I first did this work.. >
Here is an example using bimap -typeof test. We are running the test for bimap/typeof but the output has date_time:

So to be clear, the actual test log (bjam.log) is mixed up and it's not a problem that the log file is okay but process_jam_log is hosing the xml results generation?

Test output: IBM_aix - bimap - typeof / vacpp
Rev 69148 / Tue, 22 Feb 2011 16:31:14 +0000
Report Time: Wed, 23 Feb 2011 14:51:27 +0000
Compile [2011-02-22 20:36:14 UTC]: fail
echo passed > "/nfs/sparky/home/montana/boost_NEW/boost-CVS/dev.aix/results/boost/bin.v2/libs/date_time/test/testdate_facet_new.test/vacpp/debug/testdate_facet_new.test"

I can also try the -p0 option as well to see if that helps.

What OS is this on? Any chance you can send me a full bjam.log off list?

-- Noel