aboutsummaryrefslogtreecommitdiff
path: root/src
AgeCommit message (Collapse)Author
2019-03-17Helmert: avoid potential division by zeroEven Rouault
Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=11893 Credit to OSS Fuzz
2019-03-17sterea: fix lat_0 = -90Even Rouault
A division by zero caused NaN values to be returned. Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=12062 Credit to OSS Fuzz
2019-03-17Hammer: avoid division by zeroEven Rouault
Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=12799 Credit to OSS Fuzz
2019-03-17ellps_spherification(): detect invalid semi-major axis valueEven Rouault
Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=12909 Credit to OSS Fuzz
2019-03-17aea: validate |lat_1| and |lat_2| <= 90Even Rouault
Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=13028 Credit to OSS Fuzz
2019-03-16Use 1st eccentricity instead of 2nd eccentricityKristian Evers
By mistake the second eccentricity was used in a few places in the Molodensky transform. According to the literature the first eccentricity should always be used in the eccentricity. This only affects the output slightly. Reported test coordinates differed by less than a millimeter. Fixes #1321
2019-03-16Run scripts/reformat_cpp.shEven Rouault
2019-03-16Fix doc generation with Breathe 4.12.0Even Rouault
Breathe 4.12.0 (as pulled by MacOSX builds such as https://travis-ci.com/OSGeo/proj.4/jobs/185395222) does not seem to like default initialization in documented C++ structs (regression/bug) /Users/travis/build/OSGeo/proj.4/docs/source/development/reference/cpp/io.rst:6:Parsing of expression failed. Using fallback parser. Error was: Error in postfix expression, expected primary expression or type. If primary expression: Invalid definition: Expected identifier in nested name. [error at 67] std::string osgeo::proj::io::AuthorityFactory::CRSInfo::authName = {} -------------------------------------------------------------------^ If type: Invalid definition: Expected identifier in nested name. [error at 67] std::string osgeo::proj::io::AuthorityFactory::CRSInfo::authName = {} -------------------------------------------------------------------^
2019-03-15Molodensky: avoid division by zeroEven Rouault
Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=13069 Credit to OSS Fuzz
2019-03-15Reject eccentricity values larger than oneEven Rouault
Valid eccentricity should be between 0 (included) or 1 (excluded) Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=13665 Credit to OSS Fuzz
2019-03-06Make gie return non-zero exit code when file can't be openedKristian Evers
2019-03-01Doc: rename to ISO-19111:2019Even Rouault
And publish link to corresponding promoted and public OGC doc: http://docs.opengeospatial.org/as/18-005r4/18-005r4.html
2019-02-27Fix default context fallback in proj_log_func.Elliott Sales de Andrade
2019-02-27Fix null dereference warning.Elliott Sales de Andrade
Counter-intuitively, this means removing checks for nullptr. The compiler sees these checks and thinks the one remaining dereference is missing a check. However, since pj_get_default_ctx cannot return nullptr, these tests are redundant.
2019-02-26Fix build issues on SolarisEven Rouault
2019-02-26Rename internal constant to avoid conflict with macro in Solaris system headersEven Rouault
2019-02-26io.cpp: cast to 'unsigned char' before isspace() to avoid undefined ↵Even Rouault
behaviour, for example on netBSD 8
2019-02-26mutex.cpp: remove _XOPEN_SOURCE definition that causes compilation problem ↵Even Rouault
on netBSD 8 and is useless on Linux
2019-02-24Robinson: fix wrong values for forward path for latitudes >= 87.5 (fixes ↵Even Rouault
#1172), and fix inaccurate inverse method
2019-02-24Merge pull request #1285 from rouault/doc_pattersonKristian Evers
Doc: add minimal documentation for patterson, compact miller, misrsom, natearth2, sch and times projections
2019-02-24ParameterValue::_exportToWKT(): fix null pointer dereference with ↵Even Rouault
-D_GLIBCXX_ASSERTIONS on WKT with PARAMETERFILE (fixes #1290)
2019-02-24jniproj.cpp: fix compiler warningEven Rouault
2019-02-23Doc: add minimal documentation for sch projectionEven Rouault
2019-02-21Transformation: reintroduce the term of 'Null geographic offset' for ↵Even Rouault
transformations between geographic CRS of same datum (typically 3D to 2D)
2019-02-21Geog2D+Height -> Geog3D of same datum: avoid inserting a useless ballpark ↵Even Rouault
horizontal transformation
2019-02-21projinfo: add information about missing gridsEven Rouault
2019-02-20typo fixes: s/Explictly/Explicitly/ and s/instanciat/instantiat/Even Rouault
2019-02-20CoordinateOperation: add a hasBallparkTransformation() method that can be ↵Even Rouault
used to know if it includes a very approximative transformation term
2019-02-20CompoundCRS to Geog3DCRS: in synthetised transformation, document in the ↵Even Rouault
name we are lacking an ellipsoid height to vertCRS height correction
2019-02-20PROJStringFormatting: change order of emission of push/pop w.r.t axis ↵Even Rouault
swap/unitconvert to avoid useless simplification rules
2019-02-20projinfo: advertize the use of '--spatial-test intersects' when it can bring ↵Even Rouault
more results
2019-02-20Apply multiplier in proj_vgrid_value()Even Rouault
2019-02-20PROJStringFormatter: remove useless push/pop in another caseEven Rouault
2019-02-20Operation sorting: tweakEven Rouault
2019-02-20Vertical CRS transformation: synthetize a vertical unit change ↵Even Rouault
transformation when needed, and also sort Null geographic offset transformation in last
2019-02-20compoundCRS to compoundCRS: avoid emitting dummy 'Null geographic offset ↵Even Rouault
from X to X' in transformation name
2019-02-20vgridshift: propagate multiplier to avoid false-positive detection of nodata ↵Even Rouault
values in the grids with US VERTCON grids that are in millimeters
2019-02-20Add VERTCON grid name alternatives in database, and handle filename ↵Even Rouault
substitution for VERTCON method
2019-02-18Fix a few issues with i386 buildsEven Rouault
Relax tolerances in a few unit test, and in laea code. Seen with gcc 5.3 and also 7.1 Related to the use of the 387 floating-point math, since they disappear with gcc 7.1 if using non-default -mfpmath=sse -msse
2019-02-17Modify the default strategy of researching intermediate CRS to do it only if ↵Even Rouault
there is no direct transformation
2019-02-16Make sure proj_get_source_crs() and proj_get_target_crs() work on 'pseudo' ↵Even Rouault
PJ object returned by proj_create_crs_to_crs() when there are several alternatives
2019-02-16Make sure that the PJ object returned by proj_create_crs_to_crs() is a ↵Even Rouault
proper CoordinateOperation so that we can call proj_get_source_crs() on it for example
2019-02-14Bump ABI version numberKristian Evers
2019-02-14Merge pull request #1264 from kbevers/remove-t_obsKristian Evers
Clean up time handling in helmert and deformation
2019-02-14Coordinate operations where a grid is present, but not in grid_alternatives ↵Even Rouault
table, should be considered as known to be sorted first
2019-02-14deformation: Replace +t_obs with +dtKristian Evers
The +t_obs parameter was confusing for users since it effectively overwrote the observation time in input coordinates. To make it more clear what is the operation is doing, users are now required to directly specify the time span for which they wish to apply a given deformation. The parameter +dt has been added for that purpose. The new parameter is mutually exclusive with +t_epoch. +dt is used when deformation for a set amount of time is needed and +t_epoch is used (in conjunction with the observation time of the input coordinate) when deformation from a specific epoch to the observation time is needed.
2019-02-14Merge remote-tracking branch 'osgeo/master' into remove-t_obsKristian Evers
2019-02-14Reverse direction of deformation operations/transformations/deformationKristian Evers
Changed the direction of dt-calculation to follow the same convention as helmert. Changed from dt = t_c - t_obs to dt = t_obs - t_c, which effectively reverses the direction of the operation. Existing projstrings using deformation can simply reverse the direction of the operation to achieve the same results as before this commit.
2019-02-13EPSG Helmert and Molodensky-Badekas methods in the Geographic 2D domain: use ↵Even Rouault
the push/pop v_3 operator to preserve the Z component
2019-02-13PROJ string parsing: no longer attempt to build a CoordinateOperation object ↵Even Rouault
from +proj=helmert/molodensky since there are ambiguities
generated by cgit v1.2.3 (git 2.25.1) at 2026-05-02 15:04:24 +0000