[Mapserver-dev] Fix for broken pixmap transparency in 4.2 and 4.3 -> 4.2.4 release?

[Sean Gillies]

On Oct 4, 2004, at 7:48 AM, Frank Warmerdam wrote:

> Sean Gillies wrote:
>> Hi all,
>> I've come up with a fix for the long standing bug regarding
>> the rendering of pixmaps with transparency onto RGB map images.
>>   http://mapserver.gis.umn.edu/bugs/show_bug.cgi?id=926
>> Commit is pending review by authors of mapdraw.c and mapgd.c.
>
> Sean,
>
> I looked int he bug report but only see a general description of the
> solution, not the patches themselves.  Am I missing something?  I 
> didn't
> find the description quite clear.
>
>> This would make 6 significant bug fixes since 4.2.3, and I'd
>> like to request a 4.2.4 release before we get completely
>> occupied with the work of releasing 4.4.0.  Yes?
> >
>> While looking into the map rendering code, it appears to me that
>> a nice feature for a future 5.0 release would be 7 bit alpha
>> support for vector layers.  Say, if a style's color was defined
>> like:
>>     COLOR 255 0 0 64
>> then the vectors would be rendered with ~50% transparency.
>
> I believe that we should treat any alpha transparency values as 8bit
> and just downgrade to 7 bit internally when applying to GD.  It would 
> be
> confusing to the user to user, and to developers of other rendering
> modules to use 7bit alphas just because that is a limitation of GD.
>
> Note that your change is basically to use RGBA for colorObj 
> everywhere, right?
> I do think that is very desirable though it may be a fair amount of 
> work to
> apply across the board.  No harm in rolling it out in sections though.
>
> So, is the next version after 4.4 really 5.0 instead of 4.6?
>

Frank, please disregard my wishful thinking about full RGBA in layer
styles.  That paragraph really had nothing to do with my fix for 490,
and I should have forseen that it would lead to confusion.

Sean