Monday's valgrind results: no new errors

[Dan Kegel]

http://kegel.com/wine/valgrind/logs-2007-12-10-summary.txt
seems to show lots of valgrind regressions in mshtml, but
http://kegel.com/wine/valgrind/logs-2007-12-10/vg-mshtml_dom-diff.txt
shows the difference is that I happen to have installed Gecko.
So they aren't really regressions (or if there are any, they're
buried in the noise), and today's git is probably about
as Valgrind-clean as Saturday's.
- Dan

[Alistair Leslie-Hughes]

Dan Kegel wrote:
> http://kegel.com/wine/valgrind/logs-2007-12-10-summary.txt
> seems to show lots of valgrind regressions in mshtml, but
> http://kegel.com/wine/valgrind/logs-2007-12-10/vg-mshtml_dom-diff.txt
> shows the difference is that I happen to have installed Gecko.
> So they aren't really regressions (or if there are any, they're
> buried in the noise), and today's git is probably about
> as Valgrind-clean as Saturday's.
> - Dan

Hi Dan,

From todays list there isn't a memory leak in vg-shlwapi_string.txt 
from but on the 8th there was.  Just a little concerned that something 
has been missed. I have a patch locally that I haven't submitted yet.

Also vg-msxml3_domdoc.txt has 2 "Conditional jump or move depends on 
uninitialised value" statements that don't appear in todays results.

Did something change?

Best Regards
  Alistair Leslie-Hughes

[Alistair Leslie-Hughes]

Alistair Leslie-Hughes wrote:
> Dan Kegel wrote:
> > http://kegel.com/wine/valgrind/logs-2007-12-10-summary.txt
> > seems to show lots of valgrind regressions in mshtml, but
> > http://kegel.com/wine/valgrind/logs-2007-12-10/vg-mshtml_dom-diff.txt
> > shows the difference is that I happen to have installed Gecko.
> > So they aren't really regressions (or if there are any, they're
> > buried in the noise), and today's git is probably about
> > as Valgrind-clean as Saturday's.
> > - Dan
>
> Hi Dan,
>
> From todays list there isn't a memory leak in vg-shlwapi_string.txt 
> from but on the 8th there was.  Just a little concerned that something 
> has been missed. I have a patch locally that I haven't submitted yet.
>
> Also vg-msxml3_domdoc.txt has 2 "Conditional jump or move depends on 
> uninitialised value" statements that don't appear in todays results.

Just ignore the last domdoc comment.  Its the noise of 
err:msxml:create_node shouldn't be here!.  That I am not seeing.

Best

[Dan Kegel]

Alistair Leslie-Hughes wrote:
> Dan Kegel  wrote:
> > http://kegel.com/wine/valgrind/logs-2007-12-10-summary.txt
> > ... today's git is probably about as Valgrind-clean as Saturday's.
>
> From todays list there isn't a memory leak in vg-shlwapi_string.txt
> from but on the 8th there was.  Just a little concerned that something
> has been missed. I have a patch locally that I haven't submitted yet.

I haven't been paying any attention to memory leaks myself.
In fact, I only show files that include pointer and uninitialized access errors,
so if a module doesn't have any non-memory-leak
errors, I don't even show the file.  That may be why the
file is missing on the 10th.

I guess my thinking was "let's fix all the serious errors, then start
fixing memory leaks".  But I guess I should put some effort into
making the logs more useful for people looking for memory leaks.

Of course, if you can run valgrind locally, that's the best way to
investigate things in detail.
- Dan