[Mapserver-dev] Antialiasing State of Affairs

[Steve Lime]

1a) Report them as bugs! The cartoline stuff should be pretty good and
the developer is
still around. I can try to fix things if he can't/won't.

1b) GD only supports single pixel lines (last time I looked). I'd need
to seem samples of "odd"
to comment further. Cartolines are the way to get wide, antialiased
lines.

2 a/b) Not really. I'd prefer a more unified model. The code was
developed at seperate times with
2 or 3 authors. Cartolines and the ttf lines were special cases of a
normal line. Ideally they will
become unified. I spoke briefly with the cartoline developer at MUM2
about the possibility.

3) Tough to say, but it's not trivial. There are a couple of major
tasks:

  - cartoline cleanup
  - generic symbol rotation for vector symbol descriptions and images.
  - logic to know when cartolines are a more appropriate approach to a
brush.

My guess would be several weeks of development. Futhermore we may want
to look at another
rendering engine, OpenGL has been suggested, as opposed to GD.
Something to think about.

4) Yes, see Peter's follow up.

Steve

>>> Titus von der Malsburg <malsburg at cl.uni-heidelberg.de> 9/9/2004
7:38:08 AM >>>

Hi,

yesterday I spend a hard time to figure out how antialiased rendering
with gd works in Mapserver (including: a lot of reading in mapdraw.c,
mapgd.c and friends).  Finally I got it working.  Now I have some
questions:

1.a) Since I get a lot of segfaults, I wonder how developed the
code may be.  Are there only a few minor bugs or is this cartoline
stuff
more like a feasibility study?

1.b) Drawing antialiased lines works only for lines with a thickness
of
one pixel.  Thicker lines look odd.  Is gd responsible for
that or is it due to deficiencies of the cartoline code?

2.a) Are there good reasons why one would want to have three
spereate techical concepts for lines (normal, dashed lines with
symbols,
cartoline)?

2.b) Would it be a good idea to unify these?

3.) How much work would it be to
 * to make cartoline rendering stable,
 * to realize the unification of the line concepts?

4.) Is there anyone working on this currently?  Is there any interest
in
these things at all?


Thanks,
	Titus
_______________________________________________
Mapserver-dev mailing list
Mapserver-dev at lists.gis.umn.edu 
http://lists.gis.umn.edu/mailman/listinfo/mapserver-dev