Hello all,
I'm new to MonetDB, but I'm not new to databases or coding. I've been building the Apr2011 branch on my home system, a Fedora release 14 (Laughlin) (8x Intel(R) Xeon(R) CPU X3440 @ 2.53GHz, Memory 8187MB) using clang (clang version 2.8 (branches/release_28) Target: x86_64-redhat-linux-gnu Thread model: posix) and having great success - I am thrilled to be able to work with such an amazing piece of software that has such an elegant and well thought out engineering design and structured methodology (nice work!). Everything builds and runs as expected. However when I run "make check" there are a number of things that catch my eye and I want to know from the experts how you'd like me to help with resolving these potential issues.
First, here's how I'm building things:
export JAVA_HOME=/usr/java/jdk1.5.0_22 export PATH=${JAVA_HOME}/bin:$PATH export CC=clang
./configure --enable-testing --enable-developer --enable-debug --enable-assert --enable-instrument --enable-merocontrol --enable-strict --enable-warning --enable-dependency-tracking --enable-bits=64 --enable-jdbc --enable-odbc
make -k --load-average=90 --jobs=`cat /proc/cpuinfo | grep processor | wc -l` && make check
Failures include "major differences" found in the following: tst830.mal in mal vacuum.mal in modules/kernel attach00.mal in modules/mal common.mal in optimizer perl_dbi.MAL.sh in perl/dbi
The attach differences I'm seeing don't seem (at first glance) to match those of http://bugs.monetdb.org/show_bug.cgi?id=2751
My general question is, should I report these as bugs? What kind of information from the test run would you like me to add to the bug record? How would I find out if these are already reported (I've tried a few searches already)? I'm happy to dig in and start working on these issues, contributing patches where/when I can. I hope to follow whatever established patterns there are for developing/fixing MonetDB.
Also, I see a lot of output like "(attach00.test.err.FILTERED) significantly" and "(pithole.test.out.FILTERED) significantly" what does this mean? Are there tests being skipped? Is there some issue/error that's being filtered out (removing noise so as to better see the signal)? Is there a way to say, run all the tests including those that are expected to fail?
Another, semi on-topic, question. I happen to use emacs, has anyone taken the time to write up a .mx mode so that it's a bit easier to develop against the .mx source files? Any debugging hints/tips/tricks hidden on the web anywhere?
best,
-greg