[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: "chunk is already free" bug

From: Ben Collins-Sussman <sussman_at_newton.ch.collab.net>
Date: 2001-04-05 20:16:00 CEST

RADICS Peter <mitch@lbcons.net> writes:

> I think I found the "chunk is already free" bug
> (seen in "fs-test 24" for example on FreeBSD at least).
>
> svn_fs__track_dbt() was called for a DBT that was not an
> svn_fs__result_dbt, but a DBT whose data was allocated in a pool. The
> attached patch fixes this, but I'm not sure if this is the "rigth"
> solution to this problem, 'cos I only had this afternoon to look into
> this and I don't see clearly whether it's OK to use this
> pool-allocated DBT here since the data is changed in the
> last_key_before() call.

Wow! Mike Pilato and I were on this exact same train of reasoning
last night... we realized that the __result_dbt() calls were paired
with the __track_dbt() calls, and we hypothesized that the problem was
exactly what you describe.

We just never actually *found* the erroneous __track_dbt. Thank you
so much!
Received on Sat Oct 21 14:36:28 2006

This is an archived mail posted to the Subversion Dev mailing list.