You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by "Hyrum K. Wright" <hy...@mail.utexas.edu> on 2009/04/06 17:11:54 UTC
Re: svn commit: r37026 - branches/1.6.x
On Apr 6, 2009, at 11:59 AM, Arfrever Frehtes Taifersar Arahesis wrote:
> Author: arfrever
> Date: Mon Apr 6 09:59:35 2009
> New Revision: 37026
>
> Log:
> On the '1.6.x' branch:
>
> * STATUS: Nominate r37009 and vote for some revisions.
> ...
> + * r37009
> + Recommend SQLite 3.6.12.
> + Votes:
> + +1: arfrever
Note that even if this gets merged, it's doubtful I'll use the updated
script when generating the deps tarball. Unless there is a critical
bug found in 3.6.11, I plan on pegging that version with the 1.6.x
branch. There's just no compelling reason to change it. And we
probably ought to ship the recommended version in our deps tarball, so
it doesn't make much sense to update the recommended version on
configure.
(Also note that this is just for our deps tarball. Chances are that
most people will be running Subversion from a distribution or other
third-party tool which will link against the sqlite library of their
choice.)
-Hyrum
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1563209
Re: svn commit: r37026 - branches/1.6.x
Posted by "Hyrum K. Wright" <hy...@mail.utexas.edu>.
On Apr 6, 2009, at 12:22 PM, Arfrever Frehtes Taifersar Arahesis wrote:
> 2009-04-06 19:11:54 Hyrum K. Wright napisał(a):
>> On Apr 6, 2009, at 11:59 AM, Arfrever Frehtes Taifersar Arahesis
>> wrote:
>>
>>> Author: arfrever
>>> Date: Mon Apr 6 09:59:35 2009
>>> New Revision: 37026
>>>
>>> Log:
>>> On the '1.6.x' branch:
>>>
>>> * STATUS: Nominate r37009 and vote for some revisions.
>>> ...
>>> + * r37009
>>> + Recommend SQLite 3.6.12.
>>> + Votes:
>>> + +1: arfrever
>>
>> Note that even if this gets merged, it's doubtful I'll use the
>> updated
>> script when generating the deps tarball. Unless there is a critical
>> bug found in 3.6.11, I plan on pegging that version with the 1.6.x
>> branch.
>
> 2009-03-31 15:41:30 D. Richard Hipp napisał(a):
>> ...
I read the release announcement last week, no need to parrot it back
on this list.
We don't use *either* of the failure modes listed in that
announcement, in-memory databases or a shared memory cache, so neither
of these bugs affect us. Therefore, an upgrade isn't required. QED.
Also, let me repeat: the number of people which build our packages
from source is miniscule compared the to the number of people who use
a precompiled binary. And the people building those binaries control
which versions of SQLite they use, regardless of what we ship.
But this isn't worth my time or effort arguing about.
-Hyrum
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1563439
Re: svn commit: r37026 - branches/1.6.x
Posted by Arfrever Frehtes Taifersar Arahesis <Ar...@GMail.Com>.
2009-04-06 19:11:54 Hyrum K. Wright napisał(a):
> On Apr 6, 2009, at 11:59 AM, Arfrever Frehtes Taifersar Arahesis wrote:
>
> > Author: arfrever
> > Date: Mon Apr 6 09:59:35 2009
> > New Revision: 37026
> >
> > Log:
> > On the '1.6.x' branch:
> >
> > * STATUS: Nominate r37009 and vote for some revisions.
> > ...
> > + * r37009
> > + Recommend SQLite 3.6.12.
> > + Votes:
> > + +1: arfrever
>
> Note that even if this gets merged, it's doubtful I'll use the updated
> script when generating the deps tarball. Unless there is a critical
> bug found in 3.6.11, I plan on pegging that version with the 1.6.x
> branch.
2009-03-31 15:41:30 D. Richard Hipp napisał(a):
> SQLite version 3.6.12 is now available on the SQLite website
>
> http://www.sqlite.org/
>
> During release testing for 3.6.12, we stumbled over a critical bug
> that has existed in SQLite since version 3.4.0. This bug will cause
> corruption of in-memory databases if an incremental vacuum is rolled
> back. The corruption is virtually guaranteed if you rollback an
> incremental vacuum in an in-memory database. Presumably not many
> people are doing that or we would have seen this sooner. As far as we
> are aware, incremental vacuum works correctly for on-disk databases.
> We will continue to study this issue over the next weeks and release
> additional patches if we find new problems.
>
> A manual code review uncovered another bug in the lookaside memory
> allocator which can result in a segfault when shared cache is
> enabled. At the very least, users of shared cache should disable
> lookaside memory. A better solution is to update to version 3.6.12.
>
> New features added to version 3.6.12 include the
> sqlite3_unlock_notify() interface and the reverse_unordered_selects
> PRAGMA. Additional information about both is available on the website.
>
> Because of the incremental vacuum and lookaside memory allocator
> issues identified above, we recommend that everyone update to SQLite
> version 3.6.12.
>
> D. Richard Hipp
> drh@hwaci.com
>
>
>
> _______________________________________________
> sqlite-announce mailing list
> sqlite-announce@sqlite.org
> http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-announce
--
Arfrever Frehtes Taifersar Arahesis