Hi Pino:
I've suppressed some OCaml and libosinfo valgrind errors in virt-v2v.
The remaining valgrind errors are here:
http://oirase.annexia.org/tmp/v2vvg/
They all seem to be basically the same. But I couldn't work out if
these are expected leaks in the libosinfo code (in which case we
should suppress them), or if they are actual bugs because we are
missing a true destructor here:
https://github.com/libguestfs/virt-v2v/blob/8e870da79b5a61513f568b0b81c77...
Perhaps there's a reason why we cannot have a destructor, for example
that the C database is supposed to hold references to the OS objects?
Unfortunately we never free the database. It could be that to express
this properly we'd need to expose (db, os) tuples to the OCaml garbage
collector.
If it's all too hard to fix correctly, then adding suppressions is
fine, but it'd be nice to add a comment about what the problems are so
we can work on them in future.
BTW the informational string given here seems to be wrong - copy and
paste error?
https://github.com/libguestfs/virt-v2v/blob/8e870da79b5a61513f568b0b81c77...
Rich.
--
Richard Jones, Virtualization Group, Red Hat
http://people.redhat.com/~rjones
Read my programming and virtualization blog:
http://rwmj.wordpress.com
Fedora Windows cross-compiler. Compile Windows programs, test, and
build Windows installers. Over 100 libraries supported.
http://fedoraproject.org/wiki/MinGW