[SIS-329] Transformation of envelope from UTM to WGS84 sometime wrongly expanded to the ±180° longitude range

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.7
• Fix Version/s: 0.8
• Component/s: Referencing
• Assignee: Martin Desruisseaux
• Reporter: Martin Desruisseaux
• Created: 28/Jun/16 14:27
• Updated: 02/May/17 12:30
• Resolved: 29/Jun/16 17:55

Description

Consider the following envelope in UTM 30N (EPSG code 32630). Note that the longitude range crosses the limit of the CRS domain of validity, which is -6° to 0°:

BOX(199980 4490220, 309780 4600020)

Transformation to WGS84 (EPSG code 4326) should give something close to:

BOX(-6.5941 40.5085, -5.2462 41.5292)

but we get:

BOX(-180 40.5085, +180 41.5292)

The longitude range (-6.6° to -5.2°) has been expanded to ±180°. This is because the Envelopes.transform(…) method verifies if the given envelope contains a "singularity" (typically a pole), and the ±180° longitude are wrongly identified as contained singularities in this particular case.