Changes between Initial Version and Version 1 of RenderingPluginStatus


Ignore:
Timestamp:
Feb 21, 2010, 1:32:24 PM (14 years ago)
Author:
sdlime
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • RenderingPluginStatus

    v1 v1  
     1'''MapServer Rendering Plugin Status'''
     2
     3This page documents the rendering plugin implementation status for MapServer version 6.0.
     4
     5What Works:
     6
     7  * create image function (into an imageObj)
     8    * notes...
     9  * basic vector API is in place (e.g. draw line, fill polygon, etc...)
     10    * Cairo Graphics is implemented in trunk
     11    * AGG is minimally implemented in a sandbox
     12    * KML SoC work uses this API
     13    * ...
     14  * basic text API is in place
     15    * performance concerns regarding font handling
     16    * per glyph or per string?
     17  * symbols support works for the most part
     18    * exception in PIXMAP symbols
     19
     20Remaining Tasks:
     21
     22  * implement lazy PIXMAP loading
     23  * write our own image loaders (into an imageObj), this is part of the rendering API
     24    * using libpng, libjpeg and something for GIF
     25    * drivers could choose to implement their own (e.g. GD)
     26  * write our own image writers (from an imageObj), again, this is part of the rendering API
     27    * using libpng, libjpeg and something for GIF
     28    * drivers could choose to implement their own (e.g. GD)
     29    * Thomas has implemented PNG and JPEG writers already, minus color pre-preocessing and some other output optionss
     30  * raster API implementation
     31    * how to deal with potentially different image buffers (e.g. GD vs. everything else)?
     32  * implement color pre-processing
     33    * quantization
     34    * is there still a need for a partial palette?
     35  * existing renderer re-writes
     36    * GD
     37    * vector drivers (PDF/SVG/Flash), may, in some cases, be replaced by Cairo
     38  * symbol API gaps
     39    * use an imageObj instead of a gdImagePtr
     40  * vector API gaps
     41    * polygon fills using a tile
     42    * hatching support
     43      * possibility exists to create a hatch generator (e.g. all the lines) and we'd pass those to the renderer
     44      * GD driver has code to compute the lines, but they aren't clipped
     45  * remove GD dependencies from certain functional areas, specifically scalebar, legend and reference map generation
     46    * embedding may be difficult, however, if the symbolObj changes to use imageObj's for PIXMAP symbols then the current process can probably be retained
     47
     48   
     49