Mailing-List: contact harmony-dev-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: harmony-dev@incubator.apache.org
Received-SPF: pass (idunn.apache.osuosl.org: domain gmail.com designates
 66.249.82.236 as permitted sender)
DomainKey-Status: good
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
        s=beta; d=gmail.com;
        h=received:message-id:date:from:to:subject:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references;
        b=rCDZFSDsfTol2ueRz3Wh08OgnqRBJYiV2Xrq7I4P+Z7LNKtlnAEo5m2aDXRJpWZPpM9+JAnkWqT610qYO0o/dRv3le+T9k+tP3/elffaxo5NszJJ1n4kL/b4jfGelpldKvw2Hjxdl4XX/Qu2n6RzbJnBMJk/iJVVyuuiETLSbpI=
Message-ID: <e29119980609260551s711a60d6v94518db4ed8813c2@mail.gmail.com>
Date: Tue, 26 Sep 2006 19:51:48 +0700
From: "Evgueni Brevnov" <evgueni.brevnov@gmail.com>
To: harmony-dev@incubator.apache.org
Subject: Re: [drlvm] my latest round of patches broke something
In-Reply-To: <4DD82D70-719E-4522-A68F-91274BE1BCCE@pobox.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
References: <9E8092C8-7E31-485F-8582-F1DA85C0050B@pobox.com>
	 <2842A970-2FF6-471F-A886-8BF8CB7471FF@pobox.com>
	 <23951bd90609250300l75094a4cw7430b85ae7b1b542@mail.gmail.com>
	 <23951bd90609250341u2dbcf837kc600e51a887ffb74@mail.gmail.com>
	 <4518087E.5000408@pobox.com>
	 <23951bd90609252125t1f52f302y78c62936c2ec6353@mail.gmail.com>
	 <fb883e6f0609260021t708da23bp97a56b2f833659b3@mail.gmail.com>
	 <23951bd90609260049p4cfe4fg9d21b213f4e7351b@mail.gmail.com>
	 <23951bd90609260122x76dadac1t7a6f591285f412e3@mail.gmail.com>
	 <4DD82D70-719E-4522-A68F-91274BE1BCCE@pobox.com>

I have some concerns about changes introduced in
build\patches\win\APR\threadproc\win32\thread.c. Namely for
apr_os_thread_get function. The implementation is simple so only one
line makes the difference *thethd = &(thd->td);. It was *thethd =
thd->td; before. I see two problems here:

1) apr_os_thread_get returns an address of the td field of
apr_thread_t structure. The user may access this address even after
corresponding apr_thread_t structure is destroyed. So there will be a
garbage at that adress. I belive this is the main reason why APR
developers implemented this function like we have it now.

2) Our patched APR library become incompatible with original one. So
any external application which uses apr_os_thread_get function may
fail with patched APR.

Thanks
Evgueni

On 9/26/06, Geir Magnusson Jr. <geir@pobox.com> wrote:
> Yes, it would be good if someone approached APR on this.
>
> Does someone wish to volunteer?
>
> geir
>
> On Sep 26, 2006, at 4:22 AM, Vladimir Gorr wrote:
>
> > On 9/26/06, Vladimir Gorr <vvgorr@gmail.com> wrote:
> >>
> >>
> >>
> >> On 9/26/06, Dmitry Durnev <dmitry.a.durnev@gmail.com> wrote:
> >> >
> >> > On 9/26/06, Vladimir Gorr <vvgorr@gmail.com> wrote:
> >> >
> >> > [...]
> >> > > 3) we should ignore the patch you inlined in the message
> >> before this
> >> > one
> >> > >
> >> > >
> >> > > I'm not sure. This patch eliminates a lot of test failures as I
> >> > mentioned
> >> > > before (BTW you also commented on same issue into H-1457).
> >> > > Do we want to live with this? And one more note is our patch
> >> is not
> >> > related
> >> > > with eliminating the ActiveMQ crash. It's another story.
> >> > >
> >> > > Thanks,
> >> > > Vladimir.
> >> >
> >> > For me nothing works under windows (on latest drlvm/debug build)
> >> > without this patch. Even "Hello, world" fails(after printing "Hello
> >> > world!") with
> >> >
> >> > SEH handler: shutdown error
> >> >
> >> > and shows 2 debug messegaboxes(1 is from the launcher).
> >> > And this happens irrespective of my JAVA_HOME setting, I tried both
> >> > to unset it, and to point to Harmony JRE...
> >> > Debugging shows that in apr_thread_ext.c in
> >> apr_thread_set_priority()
> >> > SetThreadPriority(*os_thread,...) call fails with access
> >> violation...
> >> > So I think your patch, Vladimir, is useful :)
> >>
> >>
> >> Dmitry,
> >>
> >> did you run 'build.bat clean' before starting the build process?
> >>
> >
> >
> > I meant *build.bat update*, certainly. In this case all works fine
> > w/o my
> > patch,
> >
> >
> > I'd like to slightly continue the discussion about this patch. No
> > needs to
> > apply it now
> >
> > after Geir committed the thread.c file for *H-1457*. This file
> > contains a
> > fix for the APR bug.
> >
> > (BTW it'd be not bad to mention about this issue on the APR dev
> > list. Sure
> > it makes sense).
> >
> > Our patch is a workaround to avoid this bug.
> >
> >
> >
> > Thanks,
> >
> > Vladimir.
> >
> >
> > Maybe a clue for your issue is here. At least I have same issue
> > right now.
> >>
> >> Thanks,
> >> Vladimir.
> >>
> >> --
> >> >
> >> > Dmitry A. Durnev,
> >> > Intel Middleware Products Division
> >> >
> >> >
> >> ---------------------------------------------------------------------
> >> > Terms of use : http://incubator.apache.org/harmony/mailing.html
> >> > To unsubscribe, e-mail: harmony-dev-
> >> unsubscribe@incubator.apache.org
> >> > For additional commands, e-mail: harmony-dev-
> >> help@incubator.apache.org
> >> >
> >> >
> >>
>
>
> ---------------------------------------------------------------------
> Terms of use : http://incubator.apache.org/harmony/mailing.html
> To unsubscribe, e-mail: harmony-dev-unsubscribe@incubator.apache.org
> For additional commands, e-mail: harmony-dev-help@incubator.apache.org
>
>

---------------------------------------------------------------------
Terms of use : http://incubator.apache.org/harmony/mailing.html
To unsubscribe, e-mail: harmony-dev-unsubscribe@incubator.apache.org
For additional commands, e-mail: harmony-dev-help@incubator.apache.org