I suggest to add few more convenience functions,

[vadrer]

…which happen to be useful for my scripts,
these are - IsIconic, IsZoomed, OpenIcon, also 'RestoreWindow', which makes window 'restored'

thank you :)

[dk]

Hi, the idea is good, but I'd like to see a more orthogonal set of functions instead of just a single RestoreWindow call. A Get/Set WindowPlacement xs API, and RestoreWindow/MinimizeWindow/MaximizeWindow written in perl or xs, if needed.

[vadrer]

ok, glad to see you approve the idea.
if Get/Set WindowPlacement are ok with you - I'll implement RestoreWindow/MinimizeWindow/MaximizeWindow and then include tests and documentation of those. :)

[dk]

Yes, like those two. I see though numerous problem with referencing a bare scalar as a WindowPlacement structure. I'd propose to send individual elements (either as a hash or list) to a Set function, and return the similar structure from Get function.

If a structure will be a list, then it will be just a list of integers representing flags,showCmd,ptMinPosition.x,ptMinPosition.y etc - easier for the sake of louse syntax. A much better syntax would be a structure cooked up like this:

{
    flags => WPF_SETMINPOSITION,
    showcmd => SW_HIDE,
    ptMinPosition => [ 5, 25 ],
    .. etc ..
}

nicer but more work needed

[vadrer]

those were "raw" low-level functions (Get/Set WindowPlacement ), which then would be used by higher-level functions (RestoreWindow/MinimizeWindow/MaximizeWindow)
Probably "convenience" funcitons "GetWindowPos"/SetWindowPos which are implemented in MFC also easy to implement.

ok, I suggest to rename Get/Set WindowPlacement to _Get/_Set WindowPlacement and provide Get/Set WindowPlacement on top of that, as you've suggested,

[vadrer]

...now should be better :)

[dk]

Yes, almost there! I have some issues more:

• First, tests, I'd like to see it tested on both 32 and 64 bit systems to make sure pack/unpack is done correctly.
• Another one is a more rigorous parameter check to SetWindowPlacement - points and rects have to have lengths 2 and 4, respectively; if left unspecified, the function should either die or use default values if these fields are irrelevant in the context (f.ex. if WPF_SETMINPOSITION is unset). And length should not be a part of input, the function should supply it itself, otherwise memory access/corruption can happen, I think.
• Then, the scalar passed to _GetWindowPlacement is empty, I think one need to pre-allocate the 44 bytes and set the length field correctly.
• Finally, WPF_SETMINPOSITION and all other constants need to be declared to be exportable not by default though

[vadrer]

1,
void main () {
printf("sizeof(long)=%d\n",sizeof(long));
}
outputs 4 on 64 bit windows; (usually I use 32 bit perl and gcc that comes with it, but tried 64 gcc now as well)

2,I haven't introduced "WPF_SETMINPOSITION" yet, so not worried about that yet;

3, same with test - would happily add that too

[dk]

I think it would be enough to die "panic" unless 4 == length pack L => 0 somewhere either in the sub code or in tests

[vadrer]

agree,
I'll make these additions in a few upcoming days, probably at weekend, I think so :)