MapGuide Open Source:  Home |  Download |  Internals

Ticket #4 (closed defect: fixed)

Opened 2 years ago

Last modified 9 months ago

tile egde effects with raster layers

Reported by: pagameba Assigned to: traianstanev
Priority: high Milestone: 1.2
Component: Tile Service Version: 1.1.0
Severity: critical Keywords:
Cc: External ID: 839964

Description

When using raster data sources in tiled maps, the edges of the tiles sometimes have a single pixel border on some or all edges of some tiles. This appears to be a rounding error in the code that requests the tile from the FDO provider since the same error occurs with both the ADSK raster provider and the GDAL raster provider.

Tom said on the list:

Hi Paul,

This is a problem with the tiling code when it creates a tile with a raster (using the Autodesk raster provider or the GDAL provider). I debugged into this a while back and it looks like either: 1) the extracted image (from the source feature source) used to create the tile is sometimes offset incorrectly (like you say, this is a rounding error) and/or 2) the image is not being stretched to fill the whole area of the tile (GDRenderer.cpp line 553, has comment //TODO: do we need +1?).

I tried #2 but the lines still came up a bit faded. Someone with more time is going to have to look into this.

Tom

Change History

02/20/07 17:03:40 changed by Tom Fukushima <tom.fukushima@autodesk.com>

  • external_id set to 839964.

Added External ID.

02/27/07 13:10:52 changed by anonymous

  • status changed from new to assigned.
  • owner set to anonymous.

03/28/07 15:39:20 changed by anonymous

  • status changed from assigned to new.
  • owner changed from anonymous to traianstanev.

03/28/07 16:18:22 changed by anonymous

  • status changed from new to closed.
  • resolution set to fixed.

03/28/07 17:20:36 changed by pagameba

Traian, I read your commit comments where you mentioned the last row might not be filled in ... does this also apply to the vertical edge effects? You also said 'Maybe fix ticket #4 ...' ... should we close this ticket if we are not sure?

03/28/07 17:23:45 changed by anonymous

It looks fixed. The transform code and image draw code had changed since this bug was reported, so I didn't see any vertical lines. The reason I'm not sure is because the code change affects too many things for me to fully test. It just needs more testing. If the bug is still there, the ticket can be reopened.

03/28/07 17:26:46 changed by pagameba

Thanks Traian, I'll try to come up with some way of testing this.

03/28/07 17:52:18 changed by tomfukushima

Paul, This fix will be in the MGOS 1.2 beta scheduled for next Monday.

04/27/07 10:57:39 changed by anonymous

now leverages Spring's wow gold? interface for generating a unique bean name for the anynomous top-level beans provided by a namespace handler [fixes #498]; updated spring.jar and springsrc.zip of the upcoming release 2.0.3 to the current head version

12/18/07 20:54:59 changed by jbirch

  • status changed from closed to reopened.
  • resolution deleted.

12/18/07 20:56:55 changed by jbirch

  • status changed from reopened to closed.
  • resolution set to fixed.

Darn, thought I could delete spam comments as admin. Sorry.