Revision history for Perl extension Geo::ReadGRIB.
0.01 Fri Apr 21 05:45:54 2006
- original version; created by h2xs 1.23 with options
-X -b 5.6.0 -n Geo::ReadGRIB
0.40 Sat May 27
- Added a tempfile() method based on File::Tmp to name
temp filed so had to change Perl version to "use 5.6.1"
- Each method that uses wgrib.exe now uses tempfile() to
name temp files and then removes them when done.
- Changed tests to use full path to test GRIB file
- POD doc tweaks
0.50 June 12
- New method extractLaLo() extracts forecast data for a given
type and time for a range of locations. The locations will be
all (lat, long) points in the GRIB file inside the rectangular
area defined by (lat1, long1) and (lat2, long2) where lat1
>= lat2 and long1 <= long2. - that is, lat1 is north or lat2
and long1 is west of long2 (or the same as...)
0.51 July 10 2006
- Added END block and signal handlers to clean up temp files on
close or interruptions.
- Added new method getParam() to return values of selected
parameters useful to users.
0.98 March 6 2009
- Added Geo::ReadGRIB::Place and Geo::ReadGRIB::PlaceIterator
classes.
- extractLaLo() now returns a PlaceIterator object;
- Deprecated getCatalog() and getCatalogVerbose(). Getting the
offset index for each data type and time is now done during
object creation.
0.99 Oct 2009
- A major bug fix release to support Canadian high res gribs.
- Add a check for GDS byte 6 = 0. Croak if not (yet) supported
lat/long grid
- Discover scan mode from byte 28. Many changes to support south
to north scanning. Will also print advisory messages if it
detects as yet unsupported modes.
- Some refactoring and new tests.
- extract() now returns a PlaceIterator object like extractLaLo
does.
1.0 Oct 2009
- Some documentation changes and minor adjustments. Verification
of new GRIB type support justified finally moving to v1.0
1.1 Oct 2009
- Bug fix release for rt.cpan.org ticket #50820 where new test
fails for some values only on 64bit uselongdouble Perl
1.2 Nov 2009
- Improved extractLaLo() by reading all rows from (la1, lo1) to
(la2, lo2) into scalar in one read. This will get more data
than needed in many cases, but the temporary increase in memory
footprint will typically be less than a megabyte. Speed increases
grater that %35 were measured on large extracts.