Re: valgrind still unhappy despite expect changes

On Fri, Sep 02, 2011 at 01:48:29PM -0300, Jorge Arellano Cid wrote:
> On Fri, Sep 02, 2011 at 03:59:17PM +0000, corvid wrote:
> > Jorge wrote:
> > > On Thu, Sep 01, 2011 at 07:47:48PM +0000, corvid wrote:
> > > > Jorge wrote:
> > > > > On Thu, Sep 01, 2011 at 10:56:19AM -0300, Jorge Arellano Cid wrote:
> 
> > > > >   [...]
> > > > >   Any help is appreciatted.
> > > > 
> > > > My guess is that it could be something like
> > > > first, one bw asks for a page, and that becomes conn->bw,
> > > > and then another bw starts also asking for the page, and
> > > > the first bw closes, invalidating the conn->bw, and then
> > > > the second bw closes and the cancel expect code tries to
> > > > use the old bw.
> > > 
> > >   "bw" is a bit a misleading name, bacause each tab is a bw.
> > > 
> > >   OTOH, each tab has its own control panel and thus becomes
> > > a browser window object.
> > > 
> > >   In the case you propose there'd be two different bw.
> > 
> > Right. Why do you feel that they can't be two bws that were
> > both closed from the keyboard?
> 
>   I have no problems with that.
> 
>   What I don't see is two bws with the same memory address being
> closed from the keyboard.
> 
>   But  now  I see what you mean: which is a lingering bw in conn,
> (which is compatible with my analysis), as there's no need for it
> to exist in the UI when the event arrives.
> 
>   So we'de have two close events handled OK by two perfectly sane
> bws, but one call made on an already defunct bw (i.e. not present
> in the UI).
> 
>   And you hit it on the nail, because now I can reproduce it!!!

  Patch committed.

  FWIW, the testcase I used produced:

==17336== ERROR SUMMARY: 115 errors from 76 contexts (suppressed: 13 from 10)

  without the patch, and

==17468== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 13 from 10)

  with it.

  :-)

-- 
  Cheers
  Jorge.-