[ros-kernel] comdlg32 build problem part deux

[Ge van Geldorp]

> From: Vizzini
> 
> It's been a while since I've built outside of the device/ 
> tree, and it seems now that I have the same (or similar?) 
> problem that Hartmut reported last week about windres not 
> liking comdlg32's rc files.  I build on Linux, and my windres 
> is 2.14.90 (the Debian testing package of mingw).
> 
> Let me start by saying that i know relatively little about 
> .rc file syntax.  To get this to build for me, I had to 
> change lines like this:
> 
> LISTBOX  "", IDC_SHELLSTATIC,4,20,272,85, LBS_SORT | etc...
> 
> to:
> 
> LISTBOX  IDC_SHELLSTATIC,4,20,272,85, LBS_SORT | etc...
> 
> Once that was done, it built without problems.  I'm not quite 
> sure how to test this, let alone if it breaks anything, so 
> I'm humbly submitting this for review.

Ok, I'm thoroughly confused now....
This all started with a message from Hartmutt on 17 Jan:

> I've problems to compile some of the resource files from comdlg32
> and oledlg. It seems that some controls needs a string at the begin
> of the parameter list. My windres is 2.14.90 20030612. Has anyone
> else the same problem?

Included was a patch to add the empty strings at the begin of the
parameter list. Since it caused no problems for those of us using
windres 2.13.90 20030111 from the MinGW package on Windows I committed
the patch.

Martin Fuchs reported on 18 Jan that he had problems with the new
syntax:

> I just encounderd, my windres version (from binutils CVS)
> has problems with this patch. This is the reported version:
>
> GNU windres 2.14.90 20030807
>
> I switched back to windres of MinGW 3.0.0-1 to be able to
> compile ROS, but seems this is no permament solution.

Checking MSDN, the "official" syntax of LISTBOX is:

LISTBOX id, x, y, width, height [[, style [[, extended-style]]]]

So no string at the beginning. Some EDITTEXT statements were also
changed to include a string at the beginnen, the "official" syntax of
EDITTEXT is:

EDITTEXT id, x, y, width, height [[, style [[, extended-style]]]]

So it seems we have Hartmutt's 2.14.90 version which requires the
strings, Vizzini and Martin's 2.14.90 version which don't allow the
strings, current MinGW 2.13.90 version which doesn't care and MSDN which
says the strings shouldn't be there.

My proposal is to remove the strings again, since that solution seems to
cause problems for the least number of people (only Hartmutt). Hartmutt,
is this acceptable for you?

Ge van Geldorp.