* Rebuild for python 3.6 support.
* New upstream version
* Build-Depend on default-libmysqlclient-dev. Closes: #845833
* Updated debhelper compatibility level to 10
* New upstream version
- fixed rounding error on i386. Closes: #834784
* New upstream version
- depend on newer libwreport
* Added missing help2man dependency
* New upstream version
- fixed build failures on 32bit machines
* New upstream version
- Redone documentation from LaTeX to Markdown
- Introduced new SQL schema "V7"
- Refactored Fortran bindings to not need CNF
* Dropped LaTeX dependencies
* Dropped CNF dependency
* Increased libdballef soname
* Updated Standards-Version, no changes required
[ Scott Kitterman ]
* Change python:Depends to python3:Depends for python3-dballe so that
correct binary depends are generated
* Remove obsolete python packaging features from debian/control
[ Enrico Zini ]
* New upstream version
* Applied Scott's patch. Closes: #810062
* New upstream release
- big refactoring, extracting the beginning of a stable API/ABI
- ported to wreport 3
- removed internal provami, now a separate package:
https://github.com/ARPA-SIMC/provami
* New upstream release
- implemented encoding and decoding of radiation data
- implemented handling of temperature change information in synop messages
- deal with SYNOP messages which do not transmit the pressure of the
standard level
- implemented native PostgreSQL connection
- implemented native MySQL connection
- removed support for V5 database formats.
- if DBA_LOGDIR is set, it creates a log file in that directory with a
JSON trace of all significant usage of the library
* New upstream release
- rewritten database connection interface to support multiple connector
backends: now supported are ODBC and libsqlite3
* New upstream release
- several bug and documentation fixes
- new fortran API: unsetb
- if DBALLE_TRACE_FORTRAN is set to a file name, a trace of fortran API
calls will be written to it to assist in filing bug reports
- new fortran APIs: idba_messages_open_input, idba_messages_open_output,
idba_messages_read_next, idba_messages_write_next
- work around getting rowcount bug in new MariaDB
- do not build/package provami, since I am unable to maintain the
old wxPytyhon codebase and a QT rewrite is in progress. Closes: #759043.
Thanks Olly Betts <olly@survex.com> for patches in the meantime.
- improvements to dballe::Record's API
- run with autoreconf. Closes: #744563.
- updated standards-version, no changes required.
* New upstream release
- added a third database type: "mem:", which uses in-memory data
structures instead of a SQL database
- removed unused support for attr=add in Fortran API
- removed obsolete support for rep_cod
- removed need of having mobile=1 when inserting a mobile station. The
presence of 'ident' is enough to know that the station is mobile
- implemented var= and varlist= support for quantesono/query_stations
- added support for Lua 5.2
* New upstream release
- clarified and remplemented the meaning of data_filter
- auto-insert new rep_memo values when importing data, if they are
missing from the database
- work around SQL having undefined behaviour with NULLs on an UNIQUE
index. You need to run the following query to fix existing v6
databases: UPDATE data SET id_lev_tr=-1 WHERE id_lev_tr IS NULL
* New upstream release
- changed SQLite queries to avoid losing existing values when overwriting
a data value
* New upstream release
- refactored cursor+querybuilder for v6 databases
- do not leave SQL_TIMESTAMP_STRUCT.fraction uninitialised, which caused
duplicate imports in some cases
- fortran message api: deal gracefully with uninterpretable messaes
- fixed data overwrite on V6 databases
- provami: documented the possibility of passing initial key=val filters
on command line
* New upstream release
- postgresql fixes
- accept hypens as MISSING_INT
* New upstream release
- python: allow levels and time ranges to be set to None
- provami: filtering by one station doesn't raise a SQL exception anymore
* New upstream release
- new, faster database layout
- old layout still supported for existing databases
- changes to C++ and Python API to make it possible to have multiple
database engines
- Python API redesigned from scratch with clean, hand-written, documented
bindings
- support for WMO SHIP messages
* Updated sonames
* Removed dependency on swig
* New upstream release
- added --rejected=filename option to dbamsg and dbadb
- dbamsg convert -o/--output
* New upstream release
- tests and fixed for WMO pilot import/export
* New upstream release
- support for reading and writing WMO PILOT messages
* New upstream release
- low-level ODBC test suite
- removed ODBC workarounds, which used features not supported by
"industry leader" Oracle
* New upstream release
- dbadb import --fast disables journals and flushes in sqlite (same as
setting DBA_INSECURE_SQLITE)
- provami works with rpy and rpy2
* New upstream release
- 1:255:12 and 1:255:14 now recognised as ships
- fixed AIX-compatibility intexp10 in pollution importers/exporters
- help swig pattern match Varcode lists correctly, so that Python
bindings now work also for attribute management
* Added Vcs-Svn header
* Updated standards-version, no changes needed
* New upstream release
- various python bindings cleanups
- wrapped record.parse_date_extremes
* New upstream release
- fixed dbadb --report: the option had no effect
- fixed B01194 appearing twice in exported generic messages
* New upstream release
- added a missing m4 dependency
- fortran api: ana_id is not automatically reset after a prendilo. It is
still automatically reset after setting latitude or longitude
- added dbadb --wipe-first
- when DBA_INSECURE_SQLITE is set to anything, sqlite is initialised with
directives making it very fast, but losing any data integrity in case
of errors
* New upstream release
- dbatbl: fixed format formatting when len < scale
* New upstream release
- dbadb import now accepts --precise and -t csv
* New upstream release
- clear old vars on Cursor::to_record
- fixed compiler errors when building without Lua
- prevent compiler flags to end up in dependencies in tools/Makefile.am
- dbamsg convert: documented 'csv' as a valid input type
* New upstream release
- workaround to be able to build on swig 1.3 with newer gcc
- added dbamsg convert --recompute-categories
* Non-maintainer upload with maintainer's agreement:
- Depend on python-wxgtk2.8 instead of python-wxgtk2.6. Closes: #645626
* New upstream release
- in rare cases, SYNOP exports would select the wrong context for
temperature exports, incorrectly exporting minimum or maximum
temperatures instead
- tweaked sounding group detection to deal with an obscure TEMP recently
added to the test suite
- make check works on sqlite
- flight import/export: all data moved from station information level to
flight level, to prevent overlaps for flights on different days at the
same coordinates
- Added variables B01006, B01023, B20042, B20043, B20044, B20045, B02064,
B07010, B08009, B11031, B11037, B11039, B11075, B11076 and B11077 for
GTS AMDAR reports
- Make sure that height levels for flights are expressed in millimeters
as documented
- Initial support for importing/exporting experimental WMO ACARS
templates
- When scanning flight identification numbers, give preference to B01008
over B01006
* Avoid installing .la files and fixed doxygen dependencies, thanks to Luk
Claes. Closes: #631898
* Migrated to dh_python2. Closes: #616784
* Depend on swig 2.0
* Fixed timestamp syntax in old changelog entries
* dballe-common: make arch:all
* fixed lintian spellchecking errors in debian/control
* New upstream release
- Preserve message type, subtype and localsubtype during template
conversions
- Added 'wmo' autodetection template
- Deal with TEMPs with missing pressure in levels
- Deal with missing sensor height in total precipitation when doing
precise imports
- Record keyword documentation is back
- Fixed to fapi documentation
* New upstream release
- revert height->heightstatin change in variable aliases
* New upstream version
- ported to new wreport
* Added dbamsg dump --dds
* Added dbamsg dump --structured
* New upstream version
- provami: fixed disappearing 00:00:00 time in date fields
* New upstream version
- Fixed an issue with present but undefined values in the Fortran API
- Local B table change: B22042->B22043
- Local B table change: removed B08001 (replaced with B08042)
B08001 found in old generic BUFR messages will transparently be
converted to B08042 on import.
- Local B table change: removed B10009 GEOPOTENTIAL HEIGHT
(redundant, since we already have B10008)
- Lots of testing and fixes in WMO SYNOP templates import/export
- TEMP templates: all pressure imports now use B10004
- Local B table change: B07001->B07030
- Autodetect ECMWF/WMO template to use if not explicitly requested
- Imported past weather time range now is 3h or 6h depending on the
reference time hour
- Precise import of ECMWF templates use standard SYNOP levels
- Do not import undef attrs into Msg
* New upstream version
- Improved TEMP sounding group identification
- Show template numbers in bulletin exporter list
* New upsteam version
- Implemented dbamsg dump --csv
- Implemented dbamsg dump --csv --interpreted
- New local B variable: 013002 MIXING RATIO
- TEMP import tries to identify sounding groups and scan them in a
smarter way
- Also use new geopotential B code when looking for data exporting old
pilots
* New upstream version
- No transactions are left open (at least in the test suite)
- Reenabled generation if Python API docs
- Handle all 2:x:x as TEMPs
- Command line tools now properly turn exceptions into errors
- Ported rconvert to rpy2
- Fixed timeranges when importing synops
- In fortran message API, treat levels and time ranges that are not set
as "missing" instead of "leave as they were before"
- In fortran message API, fixed bug whereby it would fail to advance from
the last subset to the next message
- Fixed formatting of time ranges
- Fixed dbamsg dump --text
- Instant time range is now 254,0,0
- Messages with ECMWF templates are created with old version of tables
- Implemented message filters for all commands where it makes sense
* New upstream version
- Handle negative priorities
- voglioquesto in fortran msgapi moves to the next subset, or the next
message when there are no more subsets
* New upstream version
- Rewritten in C++
- The dballe++ module has been removed (now everything is in C++)
- BUFR/CREX functions have been split into a separate wreport library
- Bumped libdballe soname as ABI is now obviously different
- Fortran ABI is unchanged, so we still have libdballef4
* Smarter lookup of tables, with fallback on standard WMO tables
* Access variable codes via Lua
* Tweak to SIM table entries for pollutants
* Many improvements to BUFR decoder: compressed strings, compressed delayed
repetitions, C05 modifiers
* Stricter range tests on variable assignments
* Handle BUFR update sequence numbers
* Tweaked unit conversion factors
* New upstream version
+ Lua bindings to be used from C
* Dropped sqliteodbc build-dep and disabled make check as a consequence.
Closes: #581008
* New upstream version
+ Complain if trying to encode a generic message containing a variable
with an attribute which is not a B33YYY variable
+ Added dbatbl_makeb to edit B tables
* New upstream version
+ Allow to use priority filters on query=best queries
+ Implemented dbadb --dbfile to directly use sqlite file databases
+ Added dba_db_create_{from_file,from_url,test}
+ Allow to provide a url instead of a dsn as the dsn parameter in presentati
+ Implemented dbamsg convert --template to force output template
+ Implemented dbamsg convert --report to force output message type
* Build-depends on python-dev instead of python-all-dev. Closes: #571228
* Applied patch by Alessio Treglia to work with python 2.6. Closes: #535839
* Stricter shlibs
* Added fortran bindings doxygen documentation
* Updated Standards-version (no changes required)
* Fixed a misbehaviour in the Fortran API to create BUFR messages
* Allow to encode and decode an optional section in BUFR messages.
This change may break ABI, but I am not bumping SONAME because as far as I
know I can still handle rebuilds for each library user.
* Implemented "dbamsg head"
* Added msg/repinfo.h with dba_repinfo_default_filename
* Allow to filter by variable names in dbamsg filter
* Lowercase rep_memo before use
* Added support for CO, SO2 and PM2.5
* Redefine Fortran<->C constants to be more portable
* Test database using SQLite if no test DSN is specified
* Favour use of rep_memo instead of rep_cod
* New upstream version.
+ Don't give error in case of empty input
+ Fixed byteswapping of AOF on big endian machines. Closes: #526021.
* Updated Standards-Version. No changes needed.
* New upstream version
- Implemented dbamsg bisect to find messages with problems among large
message sets
- ship direction and ship speed flags were swapped when reading from AOF
- Ignore CREX data when loading tables for BUFR decoding/encoding
- Handle negative heights in AOF aireps, somehow
* libdballe-doc is now Architecture: all. Closes: #517607
* Updated Standard-Version (no change required)
* Build-Depend on ghostscript instead of gs
* Added missing ${misc:Depends}
* Updated debhelper compatibility level
* Move libdballe-doc to Section: doc. Closes: #515920
* New upstream version
- Fixed export of data with same context except for rep_cod
- Importing with --full-pseudoana will try to integrate existing data
without overwriting them
* New upstream version
- Try searching table files using BUFR4 table names even for BUFR3
- Added missing headers for GCC 4.4. Closes: #504903
- Do not allow CREX data to interfere with units and scale when decoding
and encoding BUFR
- Fixed template alias for synop messages
* New upstream version
- fixed handling of queries using query=best on MySQL
* New upstream version
- dbadb now accepts aliases with --template
- allow to select a template when creating BUFR/CREX messages from
Fortran
* New upstream version
- Import the right numpy module
- Fixed checks for real range in idba_enqr
* New upstream version
- Allow to read and write BUFR files from Fortran
* New upstream version
- Fixed import of 12-hours cumulated rain
- Test against NaN on setr and setd
- Improved manpages
* New upstream version
- Added missing include. Closes: #481397, #481738
* New upstream version
* New upstream version
+ Builds with gcc 4.3. Closes: #455391.
* Heavily reduced the amount of binary packages
* Fixed syntax of dependency fields. Closes: #453199.
* Updated standards-version
* New upstream version
+ Allow to only parse the header of a BUFR and CREX message
* New upstream version
+ Builds on ARM
* Various fixes in the Fortran API code and documentation
* New upstream version
+ Properly import and export synop high level stations
+ Improved database independence: now it can connect via ODBC to Postgres,
SQLite and Oracle
* Applied GVV 4.3 patch from tbm. Closes: #417155.
* Use ${binary:Version} instead of ${Source-Version}. Closes: #414337.
* New upstream version
+ Moved provami into the package
+ Added dballe.volnd and dballe.rconvert Python modules
+ Improved C++ and Python bindings
* Conflict with old libdballe-dev. Closes: bug#411975.
* New upstream version
+ Fixed database export problems.
+ More control on database import: please see "dbadb help import" for a
list of the new options
* New upstream version
+ Split the package in various libraries to allow use of just a part (for
example, the BUFR or CREX encoder and decoder)
+ Much improved C API documentation
+ C++ interface
+ Python interface
+ Support for SQLite
+ BUFR importer can now parse compressed BUFR messages
+ Improved level and time range formatters
+ Redesigned the dba_file API
+ Allow to run all non-database tests, and run them when building the
deb
+ Database support works on 64bit systems. Closes: bug#405357.
* New upstream release
* Build-depend on tetex-bin, tetex-extra and gs-common. Closes: bug#393047.
* Initial upload. Closes: bug#389776.