Re: [Emacs-diffs] Changes to emacs/src/xfaces.c,v

Re: [Emacs-diffs] Changes to emacs/src/xfaces.c,v

[Juanma Barranquero]

On Fri, Aug 1, 2008 at 13:48, Adrian Robert <Adrian.B.Robert@gmail.com> wrote:

> +           cmap = Fcons (Fcons (build_string (name),
> +                                make_number ((red << 16) | (green << 8) | blue)),
> +                         cmap);

This is a change over the original function, and in fact it doesn't
work on Windows because of byte order issues.
Either the attached patch, or adding a new RGB_TO_EMACS_INT macro, is needed.

   Juanma


Index: src/xfaces.c
===================================================================
RCS file: /cvsroot/emacs/emacs/src/xfaces.c,v
retrieving revision 1.409
diff -u -2 -r1.409 xfaces.c
--- src/xfaces.c	1 Aug 2008 14:01:08 -0000	1.409
+++ src/xfaces.c	1 Aug 2008 14:43:37 -0000
@@ -6601,5 +6601,9 @@
 	      name[num] = 0;
 	    cmap = Fcons (Fcons (build_string (name),
-                                make_number ((red << 16) | (green <<
8) | blue)),
+#ifdef WINDOWSNT
+				 make_number (RGB (red, green, blue))),
+#else
+			         make_number ((red << 16) | (green << 8) | blue)),
+#endif
 			  cmap);
 	  }

Re: [Emacs-diffs] Changes to emacs/src/xfaces.c,v

[Adrian Robert]

On Aug 1, 2008, at 10:54 AM, Juanma Barranquero wrote:

> On Fri, Aug 1, 2008 at 13:48, Adrian Robert  
> <Adrian.B.Robert@gmail.com> wrote:
>
>> +           cmap = Fcons (Fcons (build_string (name),
>> +                                make_number ((red << 16) | (green  
>> << 8) | blue)),
>> +                         cmap);
>
> This is a change over the original function, and in fact it doesn't
> work on Windows because of byte order issues.
> Either the attached patch, or adding a new RGB_TO_EMACS_INT macro,  
> is needed.

Thanks, in interest of heading off problems, implemented quickest fix  
for now, which is committing patch as-is.  (Sorry about any  
disruption, I could not find any #def for RGB but assumed it was like  
what I put there.)  If folks would prefer a macro we should find all  
the places it might be used and figure out the right file to put it in..

Re: [Emacs-diffs] Changes to emacs/src/xfaces.c,v

[Lennart Borgman (gmail)]

Adrian Robert wrote:
> 
> On Aug 1, 2008, at 10:54 AM, Juanma Barranquero wrote:
> 
>> On Fri, Aug 1, 2008 at 13:48, Adrian Robert 
>> <Adrian.B.Robert@gmail.com> wrote:
>>
>>> +           cmap = Fcons (Fcons (build_string (name),
>>> +                                make_number ((red << 16) | (green << 
>>> 8) | blue)),
>>> +                         cmap);
>>
>> This is a change over the original function, and in fact it doesn't
>> work on Windows because of byte order issues.
>> Either the attached patch, or adding a new RGB_TO_EMACS_INT macro, is 
>> needed.
> 
> Thanks, in interest of heading off problems, implemented quickest fix 
> for now, which is committing patch as-is.  (Sorry about any disruption, 
> I could not find any #def for RGB but assumed it was like what I put 
> there.)  If folks would prefer a macro we should find all the places it 
> might be used and figure out the right file to put it in..


Adrian, could you please explain what will happen when this code is run 
on w32?

Re: [Emacs-diffs] Changes to emacs/src/xfaces.c,v

[Adrian Robert]

On Aug 1, 2008, at 12:49 PM, Lennart Borgman (gmail) wrote:

> Adrian Robert wrote:
>> On Aug 1, 2008, at 10:54 AM, Juanma Barranquero wrote:
>>> On Fri, Aug 1, 2008 at 13:48, Adrian Robert <Adrian.B.Robert@gmail.com 
>>> > wrote:
>>>
>>>> +           cmap = Fcons (Fcons (build_string (name),
>>>> +                                make_number ((red << 16) |  
>>>> (green << 8) | blue)),
>>>> +                         cmap);
>>>
>>> This is a change over the original function, and in fact it doesn't
>>> work on Windows because of byte order issues.
>>> Either the attached patch, or adding a new RGB_TO_EMACS_INT macro,  
>>> is needed.
>> Thanks, in interest of heading off problems, implemented quickest  
>> fix for now, which is committing patch as-is.  (Sorry about any  
>> disruption, I could not find any #def for RGB but assumed it was  
>> like what I put there.)  If folks would prefer a macro we should  
>> find all the places it might be used and figure out the right file  
>> to put it in..
>
>
> Adrian, could you please explain what will happen when this code is  
> run on w32?

I wasn't trying to change any code, just accidentally introduced a non- 
workalike for a platform-defined macro.  Juanma's patch restores the  
previous version of the line I changed under Windows.  Again,  
apologies for the temporary bug.

Re: [Emacs-diffs] Changes to emacs/src/xfaces.c,v

[Lennart Borgman (gmail)]

Adrian Robert wrote:
> 
> On Aug 1, 2008, at 12:49 PM, Lennart Borgman (gmail) wrote:
> 
>> Adrian Robert wrote:
>>> On Aug 1, 2008, at 10:54 AM, Juanma Barranquero wrote:
>>>> On Fri, Aug 1, 2008 at 13:48, Adrian Robert 
>>>> <Adrian.B.Robert@gmail.com> wrote:
>>>>
>>>>> +           cmap = Fcons (Fcons (build_string (name),
>>>>> +                                make_number ((red << 16) | (green 
>>>>> << 8) | blue)),
>>>>> +                         cmap);
>>>>
>>>> This is a change over the original function, and in fact it doesn't
>>>> work on Windows because of byte order issues.
>>>> Either the attached patch, or adding a new RGB_TO_EMACS_INT macro, 
>>>> is needed.
>>> Thanks, in interest of heading off problems, implemented quickest fix 
>>> for now, which is committing patch as-is.  (Sorry about any 
>>> disruption, I could not find any #def for RGB but assumed it was like 
>>> what I put there.)  If folks would prefer a macro we should find all 
>>> the places it might be used and figure out the right file to put it in..
>>
>>
>> Adrian, could you please explain what will happen when this code is 
>> run on w32?
> 
> I wasn't trying to change any code, just accidentally introduced a 
> non-workalike for a platform-defined macro.  Juanma's patch restores the 
> previous version of the line I changed under Windows.  Again, apologies 
> for the temporary bug.

Thanks for the explanation.