0.9.2: new features, changes and fixes since FITSH version 0.9.1
* fix: many parts related to "unused-but-set variables" are fixed.
* new: a custom random generator library `librandom` has been
added to the source tree: the standard libc functions (srand48(),
drand48(), etc.) have been replaced by the calls provided by
this `librandom` library.
* fix: some minor fixes in the Makefile.in templates.
* fix: Makefiles can safely be run in paralell now. This is due to a tiny
hack (i.e. the make calls itself on the same directory) but works fine.
* fix: some minor fixes in the ./deb/fitsh.control template.
* new: the ./configure.mingw32 and ./configure.mingw64 script has been added
and the configure.in emplate has been fixed in accordance with the
requirements of the MINGW compilers. Hence, is it possible now to
create MS Windows executables on systems where the MINGW compilers
have been properly installed. Note that some FITSH features
which are highly UNIX specific, won't work in these MS Windows
environment or executables.
* fix: ./Makefile.in: `make deb` target has been fixed
(there isn't any /bin/bash specific commands)
* fix: lazy evaluation bug in fitsmask.c:213 and :216 has been fixed.
* obs: `grtfilter` and `grselect` has been removed from the source tree.
The functionality intended to be providedby `grtfilter` can be
implemented by `lfit`. The task `grselect` can be replaced by
using `awk` or some similar basic text processing tool.
* new: support for high dynamics (16-bit) PNM output is added to `fiinfo`.
* new: support for --help-wiki output, which results a simple but elegant
summary page about the available command line options using the
Mediawiki markup language. The output can then be copied to any
Mediawiki-based system (see also http://fitsh.net/wiki/).
* new: support for simple polygon shaped apertures. See e.g. Fig 2.
of http://arxiv.org/abs/1609.02760 for an application of such
* obs: the support for various coordinate basis systems has been expunged
from the full source. for backward compatibiliy, only the
interpretation (and silent ignorance) of the "basissift" keyword
is kept which is found in transformation files. The only coordinate
basis is now the native one, in which the coordinate of the
lower-left corner of the lower-left pixel is (0,0). This implies
a shift w.r.t to the system used by IRAF where the center of the
lower-left pixel is (1,1) since in IRAF system, the lower-left corner
of the lower-left pixel is (0.5,0.5).
* fix: parts of the code related to varios --help outputs have been fixed
* fix: some common declarations and header files have been fixed
* fix: minor fixes after testing at debian/stretch/gcc-6.1.1
* new: testing and minor fixes related to the support for CLANG and TinyCC
* fix: some obsolete checks have been removed from ./configure
* fix: the subdirectory ./src/parallel has been removed from the source tree
* fix: internal debian package creation files have been moved to ./misc/deb
* fix: some fixes in the `make install` and `make uninstall` targets
* new: the directory ./man has been created with the man pages of the
various tasks found in this package. This directory is shipped with
the upstream package as well.
0.9.1: new features, changes and fixes since FITSH version 0.9.0
* fix: lfit/linear.so: -dylib argument is fixed on Mac/OSX targets.
* new: fiphot: from now, a formally zero gain (--gain 0) imply an "infinite"
gain, i.e. no photon noise from the detector. It can be conveniently
used for photometry on images of instruments where photon noise is not
an issue (e.g. Hersche/PACS images) or when photon noise is completely
negligible (e.g. photometry of very faint targets).
* new: fiarith: new per-image functions sign(.) and theta(.) are added.
sign(.) is the standard sign function while theta(.) is
the Heaviside step function (that is zero for negative numbers
and unity for non-negative numbers).
* fix: fiphot: some minor bug fixes related to the polynomial gain
* new: fitrans: simple image shift operations can be implied now
using the -e or --shift command line options.
* fix: lfit: the LIBC/math calls of finite(.) have been replaced to
* chg: The `fizip` and `fiunzip` tasks are removed from the whole package.
We suggest to use the more sophisticated, faster and standard
FITS compression tools `fpack` and `funpack`, that works nicely
in parallel with the task of the FITSH package (not transparently, but
standard input/output pipes can be built easily). As of now, a native
support for the tile compression method used by `fpack` and `funpack`
is not planned.
* chg: Some standard project files (./AUTHORS, ./CHANGELOG, ./COPYING,
./INSTALL and ./README) have been added to the source tree.
* new: The target `make deb` creates a Debian package on the current
architecture (this feature requies the `dpkg-deb` utility and also
depends on the `help2man` program in order to create the manual pages
that also should go into the Debian package).
* new: Related to the new `make deb` target, the subdirectory ./deb (currently
with two files) has been created in the source tree.
* new: The first public release of the FITSH package. See ./README and the
inline reference manuals (--long-help) for further details.
See also ./INSTALL for installation tips.