Re: [Tracker] tracker-miner-fs crash?

On Tue, Nov 10, 2009 at 17:08, Tshepang Lekhonkhobe<tshepang gmail com> wrote:
On Tue, Nov 10, 2009 at 16:54, Philip Van Hoof<spam pvanhoof be> wrote:
On Tue, 2009-11-10 at 15:43 +0200, Tshepang Lekhonkhobe wrote:
On Thu, Nov 5, 2009 at 13:36, Tshepang Lekhonkhobe<tshepang gmail com> wrote:
On Thu, Nov 5, 2009 at 12:45, Martyn Russell<martyn lanedo com> wrote:

Here's what I'm now getting:

Auwch, this is memory corruption in the gslice magazine allocator.

We'll need to you to run valgrind with

export G_SLICE=always-malloc
export G_DEBUG=gc-friendly

And reproduce to find this issue

You can also set G_DEBUG=fatal-criticals so that we can find where the
criticals are coming from. If we are lucky those criticals are related
to the memory corruption. Meaning that we have good assertions ;-)


$ G_SLICE=always-malloc G_DEBUG=fatal-criticals valgrind
--leak-check=full --show-reachable=yes --leak-resolution=high -v
--log-file=valgrind.log /usr/libexec/tracker-miner-fs

produces a quick crash:

Tracker-Message: Starting next miner...
(tracker-miner-fs:23552): Tracker-DEBUG: URN '(null)' (mount point:
urn:nepomuk:datasource:9291a450-1d49-11de-8c30-0800200c9a66) was not
reported to be mounted, but now it is, updating state
(tracker-miner-fs:23552): Tracker-DEBUG: Setting up mount point
Tracker-Message: Crawling recursively directory '/usr/share/applications'

Tracker-CRITICAL **: Could not initialize currently active mount
points: 1.345: syntax error, undefined variable

and now with --track-origins=yes, find attached

my place on the web:

Attachment: valgrind.log.7z
Description: application/7z-compressed

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]