Proposed EM_Fontrange Patch

[James McKenzie]

All:

I would like to see the problems with EM_Fontrange be fixed.  The 
attached patch is for comment/review and I would like SPECIFIC guidance 
on how to implement this patch as it relates to a Wine 1.0 bug 
(bug/issue 6254) and fixes the problem specified there.  The conformance 
tests are already in tests/editor.c and all they involve is removing 
some todo_wine blocks related to this issue.  The entire patch will do 
this.  Sorry that the patch is in UNIX diff format, I could not get git 
to do what I wanted.  The actual patch will be git format and use the 
daily git.

Also, I would like to extend thanks to Tony Rollo, Lei Zhang and Eric 
Pouech for their guidance and prior fixes.  I would like to thank Erich 
Pouech for moving some of the functionality that the original patch 
introduced into OLE handling.

Please be BRUTAL as I feel we have a limited time period to get this in 
and other related issues that the patch will expose fixed before Wine 1.0.

Thank you.

James McKenzie

[Eric Pouech]

James McKenzie a écrit :
> All:
>
> I would like to see the problems with EM_Fontrange be fixed.  The 
> attached patch is for comment/review and I would like SPECIFIC 
> guidance on how to implement this patch as it relates to a Wine 1.0 
> bug (bug/issue 6254) and fixes the problem specified there.  The 
> conformance tests are already in tests/editor.c and all they involve 
> is removing some todo_wine blocks related to this issue.  The entire 
> patch will do this.  Sorry that the patch is in UNIX diff format, I 
> could not get git to do what I wanted.  The actual patch will be git 
> format and use the daily git.

I think that a couple of things should be improved in the patch:
- there are 3 hDC stored in the context, while only 2 are of use 
(rendering and measure)
- there is some interdependency with the EM_SETTARGETDEVICE message 
implementation, that should be looked at
- patch is likely a bit too big and should be sliced before inclusion

A+

[James McKenzie]

Eric Pouech wrote:
> James McKenzie a écrit :
> > All:
> >
> > I would like to see the problems with EM_Fontrange be fixed.  The 
> > attached patch is for comment/review and I would like SPECIFIC 
> > guidance on how to implement this patch as it relates to a Wine 1.0 
> > bug (bug/issue 6254) and fixes the problem specified there.  The 
> > conformance tests are already in tests/editor.c and all they involve 
> > is removing some todo_wine blocks related to this issue.  The entire 
> > patch will do this.  Sorry that the patch is in UNIX diff format, I 
> > could not get git to do what I wanted.  The actual patch will be git 
> > format and use the daily git.
>
> I think that a couple of things should be improved in the patch:
> - there are 3 hDC stored in the context, while only 2 are of use
>
> (rendering and measure)

Actually, there are four, but it appears the one that is not needed is 
fr.hdcTarget.  Does it appear to be not used in normal usage of this 
feature?  If so, then the portion of code that was written for this use 
of this handle would not be needed. That would remove a few lines of code.
> - there is some interdependency with the EM_SETTARGETDEVICE message
>
> implementation, that should be looked at

I will look at this when I can.  My time has become very limited as of 
late.  That may be changing in a few months, however.
> - patch is likely a bit too big and should be sliced before inclusion

This is the most difficult part.  Where should I make the 'slice' at?  
With only the part that implements the called code first and then the 
calling code?  That would make the most sense to me.

James McKenzie