Opened 16 years ago

Closed 16 years ago

#2600 closed defect (fixed)

entity_encoded_follow_text test broken in 5.0 and trunk

Reported by: pramsey Owned by: sdlime
Priority: normal Milestone: 5.2 release
Component: MapServer C Library Version: unspecified
Severity: normal Keywords:
Cc: tbonfort

Description

The result image (as generated by 5.0) appears to be wrong, in that the entities are not encoded in the second example. The result from trunk appears to be wrong, in that two of the cases in the second example don't work at all.

Attachments (2)

trunk_entity_encoded_follow_text.png (20.8 KB ) - added by pramsey 16 years ago.
Trunk result
50_entity_encoded_follow_text.png (23.7 KB ) - added by pramsey 16 years ago.
5.0 branch result

Download all attachments as: .zip

Change History (6)

by pramsey, 16 years ago

Trunk result

by pramsey, 16 years ago

5.0 branch result

comment:1 by pramsey, 16 years ago

Owner: changed from sdlime to tbonfort

comment:2 by tbonfort, 16 years ago

Cc: tbonfort added
Owner: changed from tbonfort to sdlime

Steve, from my testing, this regression was introduced in r7231

the "follow" labels reappear if the label->buffer is set to a smaller value. it seems like the label's polygon is incorrectly calculated, as the labels are discarded at http://trac.osgeo.org/mapserver/browser/trunk/mapserver/maplabel.c#L253

you'll find the test data in msautotest misc/agg_entity_encoded_follow_text.map

comment:3 by dmorissette, 16 years ago

Milestone: 5.2 release

comment:4 by sdlime, 16 years ago

Resolution: fixed
Status: newclosed

The problem was that when the code was reshuffled the label->force check was omitted and in the test case that caused path following labels not to be drawn. The buffer checks are correct if you remove the FORCE TRUE line. Fixed in r7749.

Steve

Note: See TracTickets for help on using tickets.