| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
MSVC, in some versions has _isnan(), rather than isnan()... but why?
|
|
euler, murd1, murd2, murd3, pconic, tissot, vitk1 (7 projections in one
file - PJ_sconics.c)
|
|
tcc, tcea, tmerc, tpeqd, urm5, urmfps, wag1
|
|
calcofi, cass,cc,wag2, wag3, wag7, wink1, wink2
|
|
The projection list maintained by code in pj_list.c depended on stuff in
projects.h.
This stuff is non-obvious and since it is only used in pj_list.c, it has
been moved from projects.h to pj_list.c (The material moved is based on
some really clever X-macro techniques implemented by numerous
redefinitions of the PROJ_HEAD macro, followed by immediate re-inclusion
of te pj_list.h file).
Also, the self test runner pj_run_selftests has been remodelled, now
depending on a list of self test functions, generated in a similar way
to the projection list, and like the projection list generated by
X-macro techniques in pj_list.c.
For this to work, self test functions need to be available for all
projections. To fulfill this, stubs for all projections not having self
tests already, have been added to PJ_aea.c
Along with the recent build system patches by @kbevers, this first part
of the macro-refactoring project is considered close to finished.
|
|
|
|
projections natearth2 and comill to dll-build.
|
|
|
|
Added stubs for selftest functions for all projections supported.
For now, the stub collection has been addad at the bottom of PJ_aea.c
Each time a real selftest function is added to a projection source file,
the stub should be removed.
Stubbing the stuff makes sure a selftest function exists for all
projections, and hence allows us to build a list of all selftest
functions using the X-macro trick (with PROJ_HEAD as X-macro) used for
building pj_list[] in pj_init.c.
Actually, the plan is to add a selftest entry directly in the struct
PJ_LIST, and let pj_init build the list directly as a part of pj_list[]
|
|
|
|
sdfe-refactor-macros--and-repair-generic-constructor-bug
|
|
Continuing alphabetically, with a few detours due to the occasional case
of more than one projection in one source file.
|
|
Removed som debugging leftovers, and corrected the behaviour of new proj
option -C / -VC, for checking internal regression tests.
|
|
Need these to reduce the chance I'm screwing up something during this
rather intrusive code surgery
|
|
|
|
|
|
Correcting a blunder!
|
|
The brief version::
In an attempt to make proj.4 code slightly more secure and much easier
to read and maintain, I'm trying to eliminate a few unfortunate design
decisions from the early days of proj.4
The work will be *very* intrusive, especially in the PJ_xxx segment of
the code tree, but great care has been taken to design a process that
can be implemented stepwise and localized, one projection at a time,
then finalized with a relatively small and concentrated work package.
The (very) long version: See the comments in PJ_minimal.c
|
|
|
|
algorithm, and algorithm lineage.
|
|
|
|
Document a (the?) convention for PJ_VERSION for versions 4.10 and up.
|
|
|
|
Add support for MISR SOM projection
|
|
|
|
|
|
|
|
|
|
|
|
|
|
unnecessary assignment in Lambert Azimuthal Equal Area
|
|
adapt patch for #325 to fix static-only library building
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
* upgrade geodesic.[ch3]
* add test suite geodtest.c and invoke via cmake's add_test
* increment version to 4.9.3 and library version to 11.0.0
|
|
|
|
This partially reverts https://github.com/OSGeo/proj.4/commit/338ea581
(#271) to solve issue with naptrans2008.gtx grid mentioned in
http://lists.maptools.org/pipermail/proj/2016-February/007327.html
|
|
|
|
|
|
On systems with long paths (e.g. iOS, which uses UUIDs for bundle
directories), the existing maximum path length of 128 resulted in path
truncation, and caused the grid shift file loading to fail.
This change increases the buffer size for the path to 1024 chars,
enabling the nadgrids file to be loaded on such systems.
|
|
|
|
|
|
This is to address PR #341 in which Visual C++ stops compile due to encountering a #warning.
|
