ListView and HDN_ENDTRACK

[BOB MECHLER]

From other threads the Header control is a child control of the ListView. So how does one trap the resizing of the header control of a listview column?

In other words, in the WM_NOTIFY event, it seems from MSDN that I should be using a NMHEADER PTR and the NMHDR.hdr.code. That would work for the separate Header control but doesn't seem to work for a header control that is a child control of the ListView control.

Do I need to subclass the ListView to get the Header control message in the WM_NOTIFY event of the ListView? Is there even a WM_NOTIFY message associated with a sub classed Listview?

Any help would be appreciated.

Bob Mechler

[Michael Mattias]

This works for me to expand the last column of a listview to fill available space on resize..

MACRO EQ32(a,b)              = (BITS???(a)=BITS???(b))
' NOTE: THIS MAY BE REPLACED BY BITSE() in 8x or maybe 9x 

UNION LvUnion
   NMHDR  AS NMHDR
   NMLV   AS NMLISTVIEW
   NMIA   AS NMITEMACTIVATE
   LVDI   AS LV_DISPINFO
   LVCD   AS NMLVCUSTOMDRAW
   NMTTDI AS NMTTDISPINFO
END UNION

...

LOCAL hCtrl AS LONG, szText AS ASCIIZ * %MAX_PATH, plvu AS LvUnion PTR, w AS STRING

....
  CASE %WM_NOTIFY
                 plvu   = lparam

                 ' if Endtrack message THIS WORKS
                 IF eq32(@plvu.nmhdr.code, %HDN_ENDTRACK) THEN
                     hCtrl = getDlgItem(hWnd, %ID_LV)
                     IF @plvu.nmhdr.hwndFrom = Listview_GetHeader (hCtrl) THEN
                         Listview_SetColumnWidth  hCtrl, %COL_MAX, %LVSCW_AUTOSIZE_USEHEADER
                          EXIT FUNCTION
                      END IF
                 END IF     '

[BOB MECHLER]

I'm not going to pretend to understand this code fully.

Is the Union useful to trap the HDN_ENDTRACK when normally it can't be trapped by using a select case based on the control id of the listview?

Should it be outside any select case that isolates the listview's control id?

Bob Mechler

[Michael Mattias]

The UNION is immaterial; it's just what I use for handling WM_NOTIFY from lots of different Common controls.

I think I should have gone on to show the rest of the WM_NOTIFY code. That code continues with your 'conventional' test of control ID ..

CASE %WM_NOTIFY
                 plvu   = lparam

                 ' if Endtrack message THIS WORKS
                 IF eq32(@plvu.nmhdr.code, %HDN_ENDTRACK) THEN
                     hCtrl = getDlgItem(hWnd, %ID_LV)
                     IF @plvu.nmhdr.hwndFrom = Listview_GetHeader (hCtrl) THEN
                         Listview_SetColumnWidth  hCtrl, %COL_MAX, %LVSCW_AUTOSIZE_USEHEADER
                          EXIT FUNCTION
                      END IF
                 END IF     '

' ***  'regular'  WM_NOTIFY 'SELECT CASE' block begins here ***
                 SELECT CASE AS LONG @plvu.nmhdr.idfrom

                   CASE %ID_LV

                        SELECT CASE AS LONG @plVU.nmhdr.code

                           CASE  %NM_CUSTOMDRAW  ' returns LVCD (NMLVCUSTOMDRAW)

                              SELECT CASE AS LONG @plvu.LVCD.nmcd.dwDrawStage
                                  ' return:
                                 CASE %CDDS_PREPAINT
           yadda yadda yadda

What I did was FIRST test for "HDN_ENDTRACK from the Listview's Header" ... then go into a 'regular' test for Control, then code.

You have to do it this way because you can't control the "Ctrlid" of the header control created by the listview control (AFAIK) and therefore can't test for it in a 'regular' WM_NOTIFY 'SELECT CASE' block.

MCM

[Michael Mattias]

>Should it be outside any select case that isolates the listview's control id?

I guess I could have just said "yes," huh?

(FWIW I found that code using your Disk Utility program).

[BOB MECHLER]

Thanks for the extra detail. I believe I can proceed from here.

Bob Mechler

[BOB MECHLER]

%HDN_ITEMCHANGED worked better than %HDN_ENDTRACK for some reason but the overall idea worked fine.

In the application I'm tracking up to a possible 10 listviews on different tabs where the width of each column can be unique based on each user of which can be hundreds. Keeping the columnwidths in a random file named with the program name and username.

Bob Mechler