a fix for NVIZ site's dependency: elimination of sites lib dependency

[neteler]

Was: ​http://wald.intevation.org/tracker/?func=detail&atid=205&aid=371&group_id=21

Date: 2007-04-24 18:32, Bob Covill wrote in communication off-tracker:

The issue of the X icon displaying poorly is unrelated to the other issues. The X icon is drawn with the function gsd_x which in turn uses gsd_line_onsurf to draw the X. In other word it tries to drape the X on a surface. It appears that the gsd_line_onsurf function does NOT work.

This could be fixed (for the X icon) by simply drawing the X flat like the other icons.

Other functions which use gsd_line_onsurf include Query Raster. A line should be drawn between query points draped over the surface. This should probably be listed as a bug.

---

In a followup of ​http://www.nabble.com/Re%3A-GRASS-6.2.1RC1-published-p7837804.html Bob has sent me a new version of Gp3.c file, which gets rid of the sites dependency in NVIZ. Hovever, Bob had one doubt:

One of the parts I am not sure about is on line 138 and 139. I am not sure what this is intended to represent in the geopoint. Any insights would be appreciated.

In spite of this, it works: fetched the 6.3 cvs, overwrited the Gp3.c, built on Ubuntu Dapper using GCC 4.0.3 tcl/tk 8.4.12 on a 32bit Pentium M and all is OK!

No line artifacts, strange icons colors or random icon types. Adding/removing 2d/3d points at runtime and at NVIZ startup via 'points' option, displaying 3d points flat - all works OK.

Remaining issue is that the 'X' icon is still always rendered kindoff hovering over the surface in 2d mode, not on the surface (though other icons are fine in 2d mode). It is rendered on the correct elevation for 3d points though.

Since there's been no response from Bob on this subject for few weeks know, I decided it would be better to post this patch to the tracker so we don't loose an almost working fix. Hopefully somebody would like to look into lines 138, 139 and the 'X' icon issue.

Maciek

[neteler]

update to use vector lib

[neteler]

I have applied the patch as changeset:31074

The

Markus

[marisn]

X icon issue is fixed in r34305

[helena]

I am wondering whether this patch was tested for points with attributes functionality before it was applied? Maybe it fixed x draping but broke the scaling for attributes? I don't see anything obvious in the patch but it may be in how the vect library handles the attributes compared to sites

Helena

[hamish]

see r38604

[marisn]

Replying to helena:

I am wondering whether this patch was tested for points with attributes functionality before it was applied? Maybe it fixed x draping but broke the scaling for attributes? I don't see anything obvious in the patch but it may be in how the vect library handles the attributes compared to sites

Helena

Indeed this patch broke attribute dependent point rendering. I'm currently fixing parts of it, still I have problems understanding some parts of code and thus I would like to read some doc's how it should work:

• seems like only 1 layer was supported
• what's an idea behind multiple attribute rows?

[helena]

Replying to marisn:

Replying to helena:

I am wondering whether this patch was tested for points with attributes functionality before it was applied? Maybe it fixed x draping but broke the scaling for attributes? I don't see anything obvious in the patch but it may be in how the vect library handles the attributes compared to sites

Helena

Indeed this patch broke attribute dependent point rendering. I'm currently fixing parts of it, still I have problems understanding some parts of code and thus I would like to read some doc's how it should work:

• seems like only 1 layer was supported

yes, that is right - old sites map would have only one layer - I am not sure we need support for several layers for this functionality

• what's an idea behind multiple attribute rows?

I am not sure what you are refering to, there may be a paper about this functionality from the GRASS conference 2002 in Trento where this was introduced (there was a workshop as well).

[marisn]

Related to #702

Some reading: ​http://lists.osgeo.org/pipermail/grass-dev/2009-October/046563.html