Re: Reinhard Karcher : user32: Speed improvement for 16bit comm support.

[Francois Gouget]

On Wed, 3 Jan 2007, Alexandre Julliard wrote:
[...]
> Commit: dff43b732b10e603f3aee38db2f684dc7404aa4e

[...]
> user32: Speed improvement for 16bit comm support.

[...]
>      stol = (unsigned char *)COM[cid].unknown + COMM_MSR_OFFSET;
>
> COMM_MSRUpdate( ptr->handle, stol );
> - if (lpStat) {
> -
> lpStat->status = 0;
> -
> -
>   if (lpStat) {
> -
>           lpStat->status = 0;
>
> SleepEx(1,TRUE);
>
> lpStat->cbOutQue = comm_outbuf(ptr);

As far as I can see this patch only contains whitespace changes (and I'm 
not sure they are correct). How can it improve performance? Probably 
something got lost in the way...

Reinhard Karcher : user32: Speed improvement for 16bit comm support.

[Reinhard Karcher]

Francois Gouget schrieb:
> On Wed, 3 Jan 2007, Alexandre Julliard wrote:
> [...]
> > Commit: dff43b732b10e603f3aee38db2f684dc7404aa4e
>
> [...]
> > user32: Speed improvement for 16bit comm support.
>
> [...]
> >      stol = (unsigned char *)COM[cid].unknown + COMM_MSR_OFFSET;
> >
> > COMM_MSRUpdate( ptr->handle, stol );
> > - if (lpStat) {
> > -
> > lpStat->status = 0;
> > -
> > -
> >   if (lpStat) {
> > -
> >           lpStat->status = 0;
> >
> > SleepEx(1,TRUE);
> >
> > lpStat->cbOutQue = comm_outbuf(ptr);
>
> As far as I can see this patch only contains whitespace changes (and I'm 
> not sure they are correct). How can it improve performance? Probably 
> something got lost in the way...

My patch 46c463398cc29d75ec909a53b2d772c5bf41c998 was correctly
committed at 2007-01-03 13:00:46

I can't see the above patch in the git-repository. The line I added
before the line
 SleepEx(1,TRUE);
is missing in the above patch.

Reinhard Karcher : user32: Speed improvement for 16bit comm support.

[Francois Gouget]

On Thu, 4 Jan 2007, Francois Gouget wrote:
> On Wed, 3 Jan 2007, Alexandre Julliard wrote:
> [...]
> > Commit: dff43b732b10e603f3aee38db2f684dc7404aa4e
>
> [...]
> > user32: Speed improvement for 16bit comm support.
>
> [...]
> >      stol = (unsigned char *)COM[cid].unknown + COMM_MSR_OFFSET;
> >
> > COMM_MSRUpdate( ptr->handle, stol );
> > - if (lpStat) {
> > -
> > lpStat->status = 0;
> > -
> > -
> >   if (lpStat) {
> > -
> >           lpStat->status = 0;
> >
> > SleepEx(1,TRUE);
> >
> > lpStat->cbOutQue = comm_outbuf(ptr);
>
> As far as I can see this patch only contains whitespace changes (and I'm 
> not sure they are correct). How can it improve performance? Probably 
> something got lost in the way...

Sorry about that. I see it's been fixed already.