Re: Datetime picker updown support

[Rob Shearman]

Duane Clark wrote:
> This adds support for the updown control in the datetime picker 
> (previously it existed but did nothing).

...
> @@ -874,13 +1008,23 @@ DATETIME_Size (DATETIME_INFO *infoPtr, W
>     TRACE("Height=%ld, Width=%ld\n", infoPtr->rcClient.bottom, infoPtr->rcClient.right);
> infoPtr->rcDraw = infoPtr->rcClient;
> -
> - /* set the size of the button that drops the calendar down */
> - /* FIXME: account for style that allows button on left side */
> - infoPtr->calbutton.top   = infoPtr->rcDraw.top;
> - infoPtr->calbutton.bottom= infoPtr->rcDraw.bottom;
> - infoPtr->calbutton.left  = infoPtr->rcDraw.right-15;
> - infoPtr->calbutton.right = infoPtr->rcDraw.right;
> -
> -
>
> if (infoPtr->dwStyle & DTS_UPDOWN) {
> -
>    /* the updown control seems to ignore SetWindowPos messages */
> -
>
> DestroyWindow(infoPtr->hUpdown);
> -
>
> /* hmmm... the updown control seems to ignore the width parameter */
> -
>
> infoPtr->hUpdown = CreateUpDownControl (WS_CHILD | WS_BORDER | WS_VISIBLE,
> -
> 				infoPtr->rcClient.right-14, 0, 20, infoPtr->rcClient.bottom, 
> -
> 				infoPtr->hwndSelf, 1, 0, 0, UD_MAXVAL, UD_MINVAL, 0);
> -
>
> }
> -
>
> else {
> -
>    /* set the size of the button that drops the calendar down */
> -
>    /* FIXME: account for style that allows button on left side */
> -
>    infoPtr->calbutton.top   = infoPtr->rcDraw.top;
> -
>    infoPtr->calbutton.bottom= infoPtr->rcDraw.bottom;
> -
>    infoPtr->calbutton.left  = infoPtr->rcDraw.right-15;
> -
>    infoPtr->calbutton.right = infoPtr->rcDraw.right;
> -
>
> }
>
> /* set enable/disable button size for show none style being enabled */
> /* FIXME: these dimensions are completely incorrect */

This seems really wrong. You are destroying and creating the updown 
control on every size event?!?! That's really going to slow resizes 
down. Is there any evidence that the native version does it like this?

Rob

Datetime picker updown support

[Duane Clark]

Rob Shearman wrote:
> This seems really wrong. You are destroying and creating the updown 
> control on every size event?!?! That's really going to slow resizes 
> down. Is there any evidence that the native version does it like this?

No, I don't have evidence of that. On the other hand, it is unlikely the 
control will ever be resized, except once after creation. Looking 
through the messages that the updown control responds to, there are none 
that allow resizing that I can see. And while I have not verified this 
with Microsoft docs, it supposedly implements all documented functions 
(according to the header).

Datetime picker updown support

[Robert Shearman]

Duane Clark wrote:
> Rob Shearman wrote:
> > This seems really wrong. You are destroying and creating the updown 
> > control on every size event?!?! That's really going to slow resizes 
> > down. Is there any evidence that the native version does it like this?
>
> No, I don't have evidence of that. On the other hand, it is unlikely 
> the control will ever be resized, except once after creation. Looking 
> through the messages that the updown control responds to, there are 
> none that allow resizing that I can see. And while I have not verified 
> this with Microsoft docs, it supposedly implements all documented 
> functions (according to the header).

I have tested using ControlSpy & Spy++ and indeed it isn't destroyed 
after each resize. Using SetWindowPos works for me. Can you make sure 
the app still works with the attached patch?

Rob

Datetime picker updown support

[Duane Clark]

Robert Shearman wrote:
> I have tested using ControlSpy & Spy++ and indeed it isn't destroyed 
> after each resize. Using SetWindowPos works for me. Can you make sure 
> the app still works with the attached patch?

Yep, that works fine.