[I added internals@stripped as this topic may be interesting to a
Patrick Crews, 02.09.2009 20:36:
> My questions are:
> 1) Are there any special parameters that you pass to MTR or should it
> run 'out of the box' with the patch?
I added new information to the bug report #46327 yesterday. It shows
more precisely, how I use the tools. But no solution is to be seen so far.
The patch for MTR1 enables it to run a couple of backup tests and
perhaps some more tests. But other backup tests still fail with it
(different paths used by MTR1 and MTR2). Yesterday I also noticed that
MTR1 cannot run the whole test suite at all, even with my patch. Some
environment variables are missing. It tried to add them to the patch,
but gave up after some time. It were too much changes.
> 2) Do you know of any additional 'pressure points' with gcov (ie you
> need to call gcov in the basedir)
No, I don't know. I usually run commands from the tree root (where sql/,
mysys/ etc. live), but for some commands I need to cd into mysql-test.
> 3) Do you have any additional advice / information to share - any
> websites, resources, etc that can help us figure out how to stabilize /
> increase accuracy for gcov numbers?
There is this one, which I use to use: http://forge.mysql.com/wiki/DGCov
And we have mysql-test/mysql-test-lcov.pl
Note that I do not use the --gcov option of MTR, as it produces *.gcov
files, which can be viewed textual, but I'm not interested in this.
DGcov creates them again and extracts the lines I'm interested in, and
lcov (from mysql-test/mysql-test-lcov.pl) creates them again too, to
prepare HTML generation.
[Note that I observe problems with sql_yacc.yy from
mysql-test/mysql-test-lcov.pl. But since others didn't see these, I kept
my patch for this private. Should this happen for someone else too, I'll
Ingo Strüwing, Database Group
Sun Microsystems GmbH, Sonnenallee 1, D-85551 Kirchheim-Heimstetten
Geschäftsführer: Thomas Schröder, Wolfgang Engels, Wolf Frenkel
Vorsitzender des Aufsichtsrates: Martin Häring HRB München 161028