| Age | Commit message (Collapse) | Author |
|---|
|
Fixes consist in:
- no use of comma operator for multi statement purpose
- avoid confusing comma in for loops first and third clauses
- avoid implicit long to int casts by storing to long, or explicit bound checking before cast
|
|
|
|
|
|
|
|
original ellipsoid parameters (before any projection mess with them)
|
|
As suggested in [0], steps in a pipeline are now checked for compliance.
If the right side units in step n differ from the left side units in
step n+1 the pipeline can't be constructed and an error is raised.
[0] https://lists.osgeo.org/pipermail/grass-dev/2018-March/088123.html
|
|
null pointer dereference
|
|
|
|
* Avoid using angle brackets for PROJ headers.
Switching from #include <proj.h> to #include "proj.h",
and correspondingly for projects.h and proj_api.h,
reduces the risk of accidentally picking up a system
installed older version of the header while building
a new version of PROJ.
* Also handle geodesic.h
|
|
Due to the slightly involved way a pipeline is set up, only a small
subset of the definition parameters are directly read by the pj_init
code. The remaining parameters will not get their "used" flag set,
and for that reason will not be included in the projection definition
element of a PJ_PROJ_INFO, returned by proj_pj_info.
For now, we force the "used" flag of all elements of a pipeline to be
set. The code is tested by introducing cct functionality for printing
the projection definition used.
|
|
With the introduction of the "inverted" flag on PJ objects
you can no longer rely on checking that the inv, inv3d and inv4d
functions are available on said PJ object. The function is used
internally a few places and otherwise exposed in proj_api.h to
ensure that users of the old programming interface can safely
check if an operation has an inverse.
|
|
"+proj=pipeline +inv +step +urm5 +n=0.5 +inv" now works as expected,
returning the forward operation of urm5. In principle adding more +inv's
should also work, resulting in the forward operation when an even number
of +inv's are present, and the inverse when an odd number of +inv's
are present.
"+proj=pipeline +step +urm5 +n=0.5 +inv" fails at initialization since
no forward operation can be performed. This is a new requirement, but
aligns perfectly with the rest of the library since no operation without
a forward method exists.
|
|
Some projections do not have an inverse mapping. If such a projection is used
as a (forward) step in a pipeline we won't be able to perform an inverse
operation using the pipeline. By setting the inv, inv3d and inv4d pointers to
zero we signal to the caller that an inverse mapping is not available.
|
|
arbitrary dimensionality in test code
* Call trans func of same dimensionality as input in gie
* Refactor prep/fin code for pj_fwd/pj_inv 2D,3D,4D
* Remove prime meridian handling from pj_transform (now handled in pj_fwd_prepare/pj_inv_finalize)
* Introduce prep/fin skips, mostly in support of axisswap and pipeline drivers
* Refactor fwd/inv subsystem
* pj_transform: Let pj_fwd/inv handle scaling
* Let pj_fwd/inv3d fall back to 2D eventually
|
|
Having non-const variables of static lifetime or even global scope is
usually a bad idea. These variables are inherently constants, and this
should be enforced.
This required marking some functions as not modifying input parameters
and marking some pointers as pointers to const.
One advantage is that the compiler usually puts const static variables
in a read-only code segment, so they can't be modified physically.
This can be verified with `nm` (on POSIX systems).
To avoid changes to the public API, functions returning non-const
pointers to data tables were left intact, but the returned data may not
be modified. Internally we prefer using the proj_list_* functions over
the pj_get_*_ref functions, because the former return const pointers.
|
|
* Free format now in cmd lines, in gie, and in init files
* Corrected handling of defaults
* Add demo of integrated definition and validation
* Repair stack-smashing memmove in get_init
* repair paralist corruption, clean up debug output
* Install test files for nmake builds
* Add many improvements following suggestions by @schwehr
* Be consistent in requiring lower case everywhere in gie.c
Also, this Fixes #703 and Fixes #697
|
|
Decisions regarding the size of certain types are always made at
compile time, hence this is where the assertion should take place.
If the assertions fail, there is now a compiler error.
|
|
Improve error messaging for cct and gie, and do some clean ups in the ellipsoid handling - partially to squash bugs, partially to improve naming consistency which, in turn, improves the readability of the ellipsoid handling code.
Renamed functions:
pj_inherit_ellipsoid_defs has been renamed pj_inherit_ellipsoid_def, while pj_calc_ellps_params has been renamed pj_calc_ellipsoid_params.
The code in get_opt (part of pj_init.c), which handles whether or not an ellipsoid definition should be dragged in from proj_def.dat, has been rewritten. I suspect this was buggy beforehand, and at least the new code is easier to follow (although it may be slightly slower, which is not really a problem as it sits in the setup code, and hence is executed only once).
|
|
* Replace pj_ell_set with reimplementation supporting ellipsoid inheritance
* remove unreachable code from pj_ell_set.c
* Swap steps, so ellps args are read first, in accordance with historical behaviour
* Add ellipsoid tests to CI targets
* Reduce some optimistic tolerances
OS/X appears to have a slightly off float handling, resulting in differences at the nanometer level. Switching to 10 nm.
|
|
* Removed remaining traces of the builtin selftest system. Moved all functionality to test/gie
* Updated Appveyor and Travis build scripts
* Another appveyor script update
|
|
of special cases to handle in pj_init.c
|
|
PJ_OBS eliminated, API adjusted to reflect that we now have only one 4D data type.
2 new API functions added to determine output types of a PJ.
|
|
* Remove a number of debugging calls and memory management from the inner loop of the pipeline drivers
* An attempt at handling inverted steps in a more straightforward way
|
|
|
|
Allow nested pipelines when wrapped in +init's.
The previous behaviour was to quit pipeline initialization when
encountering a nested pipeline definition. With this commit that
behaviour is changed so that it is possible to nest pipelines as
long as they are defined elsewhere in a init-file.
This is useful in init-files where steps in complicated transformations
can be grouped in "sub-pipelines". These "sub-pipelines" can then be
used as individual steps in a larger and more complicated pipeline.
Nested pipelines are governed by the following rules:
1. You can't have more than one literal +proj=pipeline in a proj-string
2. Pipelines can be nested if they are wrapped up in a +init
3. More than one +init is disallowed in non-pipeline proj-strings
4. +inits are expanded as late as possible, that is they will only be
expanded in single operations (that can be a part of a pipeline)
|
|
* Linguistics: Clarify this and that, here and there
* Revert nullification of PJ_cart->fwd, inv
|
|
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=3620
Credit to OSS-Fuzz
|
|
Fuzz issue reports
|
|
In most cases memory deallocation is completely removed from the
code since it can be handled by the default destructor. In a few
special cases a local destructor overrides the default destructor
and makes sure that locally allocated memored is cleaned up correctly.
Move all deallocation from pj_free to pj_default_destructor
Rename pj_latlong.c to fit with the conventional format PJ_latlong.c - freeup was missed here due to wrong naming
Clean up pj_init to avoid double deallocation; Also resolve #576 by adding z_0 and t_0 options in pj_init, while cleaning
Add a prototype for dealloc_params
Added missing errno.h include in pj_ctx.c
Temporarily removing ob_tran from testvarious, to be sure that is where the trouble is
Make PJ_ob_tran.c use proper initialization for the chained projection
proj=ob_tran: make it clear, that we disallow ellipsoidal projections, and, for improved backwards compatibility, turns off default settings, which could inject unwanted ellipsoid definitions
... then also remove the ellipsoid definition from the testvarious test case - which is probably buggy anyway
Work around cs2cs spherical init bug in testvarious; Forbid defs for ob_tran in pj_init
|
|
|
|
* obs_api revision and improvements: new namespace etc.
* Minor corrections: use unused functions, add missing prototype, added fwd/invcoord to PJ
* Minor correction: MSVC barfs on va_arg type specification. Trying once more with added parens
* Reverting paren correction, which appears to be a non-solution
* Significant improvements to the OBS_API, plus a number of corrections, mostly in test code, to reflect API changes
* Added two missing prototypes
* Adding the proj_transform function and some related checks
* Improvements to proj_transform etc. based on suggestions from Even Rouault
* Reducing the libc include footprint of proj.h - again based on suggestions from Even Rouault
* A few minor corrections
* Eliminate a MSVC warning about non-initialized usage. Not an actual issue, as another check has locked the path, but at least this should calm down MSVC
* Improved support for the errno reset/restore paradigm
* Introduced the internal header proj_internal.h; Moved most non-API stuff from pj_obs_api.c to pj_internal.c
* Adding proj_internal.h to HEADERS_LIBPROJ to solve build problems
* Correcting a few pj...proj blunders in PJ_pipeline.c
* Correcting a few additional blunders in PJ_pipeline.c
* Changed angle-brackets to quotation marks in includes of proj_internal.h
* Some minor build system repairs
* Some PJ_CONTEXT usage simplifications following suggestions by Kristian Evers @kbevers
* Added version numbering to proj.h - Fixes #529
* remove proj_errno_restore macro, provide function implementation
* Add proj_get_definition. Fixes #538
* Added library specific deallocator proj_buffer_free, eliminating a potential cross heap issues on Windows. Thx to Even Rouault for spotting this
* Got rid of a meaningless cast in proj_buffer_free
* Added some missing functions to proj.def (again spotted by @rouault); removed some not-yet-implemented material from proj.h
* Renamed proj_get_definition/proj_buffer_free to proj_definition_create/proj_definition_destroy, for symmetry and clarity
* Renaming the definition handlers to proj_definition_retrieve / proj_free
* Renaming proj_free to proj_release
|
|
With the introduction of transformation pipelines it is now a valid
use case to have more than one +init argument in a proj string. This
commit makes sure that all +init's are expanded.
Added test case to PJ_pipeline.c for use of multiple +init's in a
proj-string.
Fixes #479.
|
|
|
|
once in the pipeline which break tests on OSX. Moving the +ellps parameters to each +step of the pipelines seems to fix the OSX tests.
|
|
Introducing the Horner polynomial evaluator also introduces the need for
very long +init:tag arguments (a n'th order 2D polynomium has
(n+1)(n+2)/2 coefficients, and n is typically in the range 5-10, i.e. up
to around 60 coefficients for each polynomium, and there are 4 polynomia
in a complete back/forward transformation set).
Hence, in this commit, along with the first part of the Horner code, the
code for reading +init files has been modified in a (for all practical
purposes) backwards compatible way, by making it possible to introduce
line continuations by escaping line breaks, i.e. preceding them with a
backslash.
An escaped line break works (as it would in TeX), by skipping all
following whitespace, including interspersed #-comments. This simple
extension makes it possible to create very long initialization elements
without losing track of the structure (cf. s45b.pol and pj_init_test.c
in the examples-directory for a demo).
The s45b.pol file was created by hand-editing the output of the software
doing the original constrained adjustment for the polynomial
coefficients. The simple adding of the “skip following whitespace and
comments” feature has made it possible to retain almost all metadata
from the source material.
This is considered very important, since 1) For the lack of a prior
common file format for geodetic polynomial coefficients, there is a good
chance that this will become THE standard, at least for the time being,
and 2) Without the metadata represented, it will be very hard for a
human to debug code involving a slightly misrepresented polynomium.
Due to the current architecture of the pj_init.c code (mostly around the
fill_buffer() function), it is next to impossible to implement the line
continuation functionality in full generality. Hence, it has been
necessary to limit this format extension to files smaller than 64 kB.
* Correction of spherical HEALpix test case
The first HEALpix test case in nad/testvarious is clearly intended to
invoke the spherical form of HEALpix.
It does, however, specify the spheroid using the +a=1 size parameter,
without specifying any shape parameter.
But since +no_defs is not specified either, a shape parameter is picked
up from the nad/proj_def.dat file (where ellps=WGS84 is given in the
<general> section).
It appears that this has not happened before I updated the pj_init code to support projection
pipelines (see below). I do, however, believe that the present behaviour is the correct one,
and rather than retrohacking the pj_init code, to (incorrectly, I
believe) reproduce the prior behaviour, I have corrected the test case
invocation in nad/testvarious to specify the spheroid using the +R=1
size parameter (which was already used in the following test case).
* Repair scaling of projections stomping on value of semimajor axis
* Workaround MSVC HUGE_VAL misimplementation.
The "return const err object" idiom (i.e. const <type> err =
{HUGE_VAL,...}; ... if (bad) return err) is problematic to implement
due to MSVC's misimplementation of HUGE_VAL as a non-const.
Hence, we need to run-time initialize these. In the pj_inv functions,
this was mistakenly done to the wrong object.
For pj_fwdobs/invobs and the remaining part of the obs-based API, this
is now worked around by providing functions returning a run time
HUGE_VAL initialized PJ_OBS or PJ_COO resp.
Obnoxious, but given MSVC's market penetration there is really not much
else we can do.
|
|
* re-enter pipeline
The pipeline interface is now internally based on the pj_obs_api, which
simplifies the implementation significantly.
This is the first mock up - it compiles fine, but is currently untested
* pipeline code cleaned up
The pipeline code is now based on the PJ_OBS api (although you can still
invoke a pipeline through pj_fwd/pj_inv and their 3D brethren).
This has made it possible to eliminate scores of funky casts and
convoluted workarounds. The code is now way more straightforward and
mostly conforming with common C idioms..
Also, the proj.h / obs_api interface to the logging system has been
streamlined through the introduction of the pj_log_error, pj_log_debug,
and pj_log_trace functions.
* Geodesics + minor changes
First proj.h style interface to Charles Karney's geodesics code:
pj_lp_dist.
Also, renamed pj_apply -> pj_trans
* Extended Ellipsoidal Parameters
Second eccentricity, second and third flattening etc.
* Rename pj_debug_set -> pj_log_level
... and add self test code for PJ_pipeline
* Clean up missing pj_apply->pj_trans
* Clean up missing pj_obs_dist_2d rename
* pj_strerrno bug fixed. Some doc/comments added
(In response to a review by @kbevers)
|
