--- In [email protected], "swzoh" <[EMAIL PROTECTED]> wrote:
>
> --- In [email protected], "entropyreduction" <lists@> wrote:

> > In principle, not necessary.  If you mean by "meanwhile",
> > "temporarily", Bruce has done a lot of work to cause intermediate,
> > temporary handles to get destroyed automatically.
>
> I tested it, indirectly by creating another com handle, and noticed
> the temporary handles were not destroyed, something like this with ADO:
> 3000, 3001, 3002 -> 3000, 3003, 3004 -> 3000, 3005, 3006 -> ...

> Maybe it could be hard to decide which one to destroy and which one
> not. There has been already an example where a handle which was
> supposed to be not destroyed was destroyed, when a map handle was an
> element of another map.

The bit you fixed was

for(local i=1;i<=rc;i=i+1)
  clip.set(rs.Fields(0).Value++" <=> "++rs.Fields(1).Value,1)
  rs.MoveNext
endfor

to

for(local i=1;i<=rc;i=i+1)
  wName=rs.Fields(0)
  wMean=rs.Fields(1)
  clip.set(wName.Value++" <=> "++wMean.Value,1)
  wName=wName.release
  wMean=wMean.release
  rs.MoveNext
endfor

That's unavoidable. I don't think there's any way Bruce could
determine that

  rs.Fields(0)

in first script was in a context where it could be thrown away.


> BTW, I tried com.release_all and/or com.unload. Strangely:
> it worked fine with ADO, however, it crashed PP with DAO.
> I have no idea why it should be.

Will check that out, as said.








Attention: PowerPro's Web site has moved: http://www.ppro.org



SPONSORED LINKS
Computer monitoring software Power pro Computer and internet software
Free computer monitoring software


YAHOO! GROUPS LINKS




Reply via email to