2016-05-27  Mohammad Akhlaghi  <akhlaghi@gnu.org>

        Following the discussions in task #13779, the ChangeLog has not
        been updated since 2015-11-05 (previous log). Instead, the commit
        messages are very descriptive. Git's blame tool can be used to
        check the history of any part of the code very effectively.

        Gnulib has scripts to generate a ChangeLog from the commit history
        so building this file automatically is possible and will be
        implemented (task #14007). Once this ChangeLog is automatically
        generated, it will be removed from the version controlled
        source. Until then this file will remain unchanged. There is also a
        discussion for adopting the ChangeLog format in the commit messages
        in task #14008.

	* ChangeLog: no longer updated.

2015-11-05  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* doc/gnuastro.texi (Why C): Clarified explanations.

2015-11-04  Mosè Giordano  <mose@gnu.org>

	* .gitignore: Add new regexps to the list of ignored files.

	* doc/gnuastro.texi (Arguments): Mention new accepted extension.

	* lib/fitsarrayvv.c (nameisfits): Add "fits.fz" to the list of
	accepted extensions.  Fixes task#13767.
	(nameisfitssuffix): Ditto.

2015-10-18  Mosè Giordano  <mose@gnu.org>

	* bin/imgcrop/imgcrop.c (imgcrop): Do not use modefunction to set
	the function, run the functions manually instead.  Fixes
	bug#45380.

2015-10-16  Mosè Giordano  <mose@gnu.org>

	* ./*: Change in many files the name of the package from
	"gnuastro" to "Gnuastro".  Fixes bug#46212.

	* doc/forwebpage.sh: Use canonical path for bash shell.

	* doc/plotbin/conversions.sh: Ditto.

2015-10-02  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/noisechisel/noisechisel.c (makeoutput): Added a "raw" term
	in the comments of the noise statistics so emphasize that this is
	from the raw mesh grid and not the interpolated and smoothed grid.

	* bin/mkcatalog/mkcatalog.c (makeoutput): As explained below.

	* doc/gnuastro.texi (Depth and limiting magnitude): Changed the
	sigma which representes depth from the maximum to the median. The
	maximum has a strong scatter and even one mesh can completely push
	to unreasonable values. But the median is not affected by such
	minor mesh variations, so it is a more robust measure of the
	dispersion.

2015-10-01  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/imgcrop/ui.c (checkifset): Removed the hstartwcs and hendwcs
	options as mandatory arguments.

	* lib/fitsarrayvv.c (readfitswcs): Now applies the hstartwcs and
	hendwcs options.

	* include/fitsarrayvv.h: readfitswcs now takes the hstartwcs and
	hendwcs values as arguments. This was done so all programs could
	take in these options.

2015-09-23  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* doc/gnuastro.texi (Depth and limiting magnitude): Updated
	explanations in for the depth and limiting magnitude to be more
	realistic.

	* bin/mkcatalog/mkcatalog.c (makeoutput): Removed the limiting
	magnitude calculation that was previously defined since it's
	physical meaning is very unclear (it was mainly put as a
	test). Also changed the surface brightness limit to a per-pixel
	basis (independent of the projected area of the pixel) as
	explained in the manual.

2015-09-17  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/mkprof/mkprof.c (mkprof): Only print a log file if the
	`--nolog' option is not called.

	* bin/imgcrop/imgcrop.c (imgcrop): Only print a log file if the
	`--nolog' option is not called.

	* bin/mkprof/main.h (mkprofparams): Removed the p->dir0file1
	parameter from the main structure.

	* tests/mkprof/mosaic1.sh: Corrected the name of the first output
	for the new individual name configuration (where the output file
	name is suffixed to the row number).

	* bin/mkprof/ui.c (sanitycheck): Now everything based on the
	situation the basic information for the next steps is set here.

	* bin/mkprof/mkprof.c (saveindividual): Removed the name checking
	here. All the necessary information is now set in ui.c.

2015-09-13  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/imgwarp/imgwarp.c (imgwarponthread): Now calculates the
	fraction of area that is blank (NaN) and if that fraction is
	larger than a user defined value, then the output pixel will be
	blank. Before the blank fraction was ignored.

	* bin/imgstat/imgstat.c (reportsimplestats): Now uses the value to
	the `--mirrordist' option instead of the fixed value of 1.5.
	(reportsimplestats): Report the `NOT ACCURATE' warning in a
	separate line to make reading by AWK easier, also removed comma
	between outputs for the same reason.

	* bin/*/ui.c (readconfig): Now uses macros for common options.
	(printvalues): Similarly uses macros for common options.

	* lib/statistics.c (sigmaclip_converge): Removed commas between
	the outputs to make reading by AWK easier.
	(sigmaclip_certainnum): Similar to above.

	* include/fixedstringmacros.h (GNUASTROBIBTEX): New name from
	ASTRUTILSBIBTEX. Also the content was updated to the full ADS
	BibTeX output after the paper was published.

	* include/configfiles.h (CHECKSETCONFIG): Now checks for the
	version of the program and also has a more cleaner/readable path
	for reading/writing to the user configuration directory and
	reading from the system configuration directory.

2015-08-16  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/noisechisel/thresh.c (snthresh): The check on the number of
	objects in finding the S/N threshold is now done before sorting
	the S/N table.

	* bin/convolve/convolve.c (removepaddingcorrectroundoff): When the
	maximum radius is smaller than the input images, then the final
	kernel will be shrinked to have no only-zero column or row.

2015-08-05  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/convolve/convolve.c (frequencyconvolve): Divide the arrays
	when the kernel should be made, multiply them when convolution is
	to be done.
	(frequencyconvolve): When in deconvolution mode, it will go to
	another function to make the final product.

2015-07-31  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/convolve/convolve.c (makepaddedcomplex): Added a check to
	make the padded image size an even number in both dimensions by
	adding one pixel if it is odd. This was done because FFT
	operations work much faster on even sized arrays.

2015-07-22  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/noisechisel/thresh.c (snthresh): Had mistakenly used
	`p->detquant', for both detection and segmentation! This is now
	corrected.

	* bin/mkcatalog/mkcatalog.c (secondpass): Sets the river flux to
	Sky for when the grown clumps are used, see the explanation
	bellow.

	* bin/mkcatalog/columns.c (brightnessmag): Corrected the
	definition of brightness. It was incorrectly dividing the total
	brightness by the area when the user asked for brightness.
	(sncol): Can now work accurately when grown clumps are also
	included. In a grown clump image, full detections that aren't
	harmed by deblending are are fully labeled. So there is no river
	pixels there and the S/N equations had to incorporate this.

2015-07-18  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/noisechisel/segmentation.c (nextavailablelabel): Now
	accounts for the `--grownclumps' option.

	* bin/noisechisel/args.h (options): Corrected the type of output
	for `--gthresh'. I had forgot to include any for it!

2015-07-17  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* lib/fitsarrayvv.c (pixelareaarcsec2): Previously called
	pixelsteradians. Changed the units to arcsec^2 because it is more
	useful.

	* bin/noisechisel/detection.c (detlabelsn): Now using a new sum of
	pixel values for flux weighted center of profile. The values are
	stored as one column in the

2015-07-07  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/mkcatalog/mkcatalog.c (firstpass): Use only positive (after
	subtracting sky) fluxes as weights for the flux weighted
	center. Also calculate the geometric center (irrespective of
	flux), so in case there are no positive pixels, the geometric
	center is output instead (after all, the object should be
	represented by some position and the NaN values that were used
	before are not useful).
	(secondpass): Similar to the above. Also I had simply replaced the
	getclumpinfo method of NoiseChisel's clumps.c here. But I had
	forgot that over there, all the neighboring clumps were within one
	detection. But here, two clumps could be one river pixel appart
	but belong to separate objects. So the wngb array has to have two
	values for each clump, not one, see the comments for more. Also I
	had to check the object IDs from the neighbors, not the river
	pixel.
	(setcpscorr): Removed. The minimum and maximum Standard deviations
	are now stored by NoiseChisel into the header of the standard
	deviation image. MakeProfiles then gets the value from those
	header parameters, so there was no need for this function.
	(makeoutput): The magnitude and S/N information is printed even if
	the user has not asked for magnitudes and S/Ns.

	* bin/mkcatalog/columns.c (sncol): Renamed variables to fit the
	clumpsntable function in bin/noisechisel/clumps.c. Also, the river
	flux is subtracted here, not before.

	* lib/fitsarrayvv.c: The numblank variable did not actually store
	the `number' of blank pixels, it was only a 0 or 1 value to
	specify if there are any blank pixels or not. It is now changed to
	`anyblank' to be more clear and not confuse the readers. All the
	Gnuastro programs that used this variable, were also corrected.

2015-07-06  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/noisechisel/noisechisel.c (noisechisel): After dilation, 4
	connectivity was used to find the labels of the connected
	components. However, when we are looking around the river pixels
	in the segmentation process, we are looking in their 8-connected
	neighbors. So when two 4-connected labeled regions are separate,
	they can still be 8-connected and this will mess up the river flux
	association functions where the flux of a river is given to the
	clump that is touching it. It took me several frustrating hours to
	find this very simple bug!

	* bin/noisechisel/clumps.c (oversegment): NaN (masked) pixels are
	no longer fed into this function. So the conditions to check for
	them, especially in setting the top index are removed.
	(getclumpinfo): Now uses a different number for the sum of flux
	when trying to find the flux weighted center of the clump. This
	sum is made of only positive values. Negative values cannot act as
	weights and will mess up the center calculations. The new number
	for each column is stored in a third column to the xys
	array. Also, the xys array is no longer needed outside this
	function.

2015-07-05  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/mkcatalog/columns.c (brightnessfluxmag): Previously
	fluxmag. Now this function gives the brightness, flux and
	magnitude depending on its input parameters.

	* bin/mkcatalog/args.h (options): Corrected all usage of "flux"
	and "brightness".

	* doc/gnuastro.texi: Corrected all usage of "flux" and
	"brightness".

	* doc/forwebpage.sh (thismonth): Removed type=\"text/javascript\""
	from the string since it seemed to be interfering with LibreJS.

2015-07-03  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* lib/fitsarrayvv.c (readkeywords): Now reads multiple keywords.

2015-07-02  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/mkcatalog/mkcatalog.c (firstpass): Now accounts for blank
	labeled pixels.
	(setcpscorr): Now finds both the minimum and maximum of the STD,
	maximum for finding the 5sigma magnitude.

	* bin/mkcatalog/main.h (CATUNITMAG): Changed unit for magnitudes.

2015-06-26  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/noisechisel/thresh.c (qthreshonmesh): Only sort and find the
	quantile if there actually is any pixels!
	(applydetectionthresholdskysub): Now accounts for blank pixels.

	* bin/noisechisel/segmentation.c (segmentation): Now only works on
	non-blank pixels.

	* bin/noisechisel/label.c (BF_concmp): Now accounts for blank
	pixels in both unsigned char and long arrays.
	(labareas): Accounts for blank pixels.
	(removesmallarea_relabel): Accounts for blank pixels.
	(labindexs): Accounts for blank pixels.

	* bin/noisechisel/detection.c (detlabelsn): Now also removes
	detections in the sky area that cover a detection after filling
	holes.
	(removefalsedetections): Changed to applydetsn.
	(detsnthreshonmesh): Changed to detectpsedos. Practically this
	function just fills the holes in each large mesh and opens the
	binary result then puts the field back into the main p->dbyt
	array. It does not calculate anything any more. Such calculations
	have now become full image wide to find one S/N value over the
	full image.
	(detsnthreshongrid): Changed to detsnthresh. This function now
	does the image-wide calculations that were previously done by
	detsnthreshonmesh.
	(dbytolaboverlap): Now accounts for masked unsigned char and long
	arrays.

	* bin/noisechisel/clumps.c (clumpsntableonmesh): Previously
	clumpsnthreshonmesh. Moved all the countings and conditionals to
	another function to find the S/N threshold from one distribution
	over the image and not in the large meshs.
	(findclumpsn): Previously clumpsngrid. Now it collects all the S/N
	values from all the meshs into one array and uses that to find the
	S/N threshold using a fixed function for both detection and
	segmentation (snthresh in thresh.c).

	* bin/noisechisel/binary.c: The fixed value of 2 is changed
	BINARYNAN which is defined in binary.h. This makes managing the
	special values to use more easy, especially now that FITSBYTEBLANK
	is also being used from fitsarrayvv.h.
	(fh_makeinv): Replaced the old functions to fill the inverse array
	with new ones that don't use an index variable but pointers. It is
	much more cleaner, faster and easier to read now. It also now
	accounts for blank pixels and completely ignores them.
	(fillboundedholes): Similar to fh_makeinv.

	* bin/noisechisel/: The mesh based signal to noise calculations
	are now gone and one value is calculated for the full image. This
	allows for much better accuracy. Most of the steps in NoiseChisel
	were modified to take this into account. Also blank pixels are now
	included in the binary and labeled images too.

	* lib/fitsarrayvv.c (arraytofitsimg): Moved the BLANK keyword
	before EXTNAME.

2015-06-25  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/imgcrop/args.h (options): Changed inpolygon to
	outpolygon. Everywhere that inpolygon appeared, it is now
	outpolygon.

	* bin/imgcrop/wcsmode.c (wcscheckprepare): Check for both axises
	having the same pixel scale now incorporates floating point
	errors.

	* bin/imgcrop/crop.c (imgpolygonflpixel): Now gets the vertices
	and their number instead of the cropparams structure. This was
	done to make it be more general and use it in both Image mode and
	WCS mode.
	(onecrop): Allocation of the crp->ipolygon array was moved to the
	cropflpixel function.

2015-06-13  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* lib/fitsarrayvv.c (fileorextname): New name for the setmaskname
	function. This was done to be a generic check, for all kinds of
	input, not just the mask.

	* bin/ : All the functions to make a copy of an input and print
	string values were changed. These were mainly in ui.c and args.h
	for each program.  Until now, for each option, I would allocate,
	check and set the proper values, but now, I wrote a small function
	in checkset.c (allocatecopyset) which does the job. This makes
	reading the code significantly more easy and less bugy.

2015-06-12  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/subtractsky/args.h (options): Removed the
	--checkinterpolation option.

	* bin/subtractsky/subtractsky.c (subtractsky): Similar to below.

	* bin/noisechisel/thresh.c (findapplyqthreshold): Similar to
	below.

	* bin/noisechisel/sky.c (findavestdongrid): Similar to below.

	* bin/noisechisel/detection.c (findsnthreshongrid): Now using
	meshvaluefile from mesh.h instead of checking its self. Once the
	--meshbasedcheck was added, doing all the checks in each function
	was too long and repetative. So I just made this function to do
	the job with one call and not distract the readers.

2015-06-11  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/noisechisel/args.h (options): `minbfrac' and `minnumfalse'
	options moved under the `Input' category from the detections
	category. This move was because these two options are used by both
	the detection and segmentation steps and thus keeping them under
	detection would be confusing. Also `checkthresh' was changed to
	`checkthreshold'. Also, the option `numerosion' was changed to
	`erode' to be consistent with the opening and dilate options.

2015-06-10  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/mknoise/astmknoise.conf: Default background magnitude
	changed to -10 and zeropoint to 0.0.

2015-06-09  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/noisechisel/clumps.c: All functions now use the
	clumpsthreadparams structure as input instead of a list of
	arguments.
	(oversegment): ctp->topinds now keeps a good value when the clump
	has NaN pixels over it.

	* lib/mesh.c (imgindextomeshid): Corrected to work correctly when
	the pixel might be lying on the last mesh that might be larger.

2015-06-08  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/noisechisel/thresh.c (applydetectionthresholdskysub):
	Similar to below, uses p->imgss for sky subtracted image.

	* bin/noisechisel/sky.c (subtractskyimg): Does not touch the input
	image, puts the sky subtracted value in p->imgss.

	* bin/noisechisel/noisechisel.c (noisechisel): Removed the
	allocation of the mesh structures to ui.c

	* bin/noisechisel/detection.c (detlabelsn): Uses p->imgss for the
	sky subtracted image.

2015-06-07  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/mkprof/oneprofile.c (randompoints): The random number
	generator is now allocated (cloned) with the build thread and for
	every profile the seed is independently set once. Before each
	pixel would allocate and seed the random number generator
	individually!

	* bin/mkprof/astmkprof.conf: Changed default zeropoint magnitude
	to 0.00.

	* bin/mknoise/mknoise.c (mknoise): moved the random number
	generator to ui.c to make things faster and more easier to
	understand.

	* bin/mknoise/args.h (argp_option): Removed the `backgroundinmean'
	option, because I couldn't figure out what use it would have. If
	it is needed we can add it again later!

	* bin/imgstat/args.h (argp_option): Changed `binonzero' option to
	the more general `onebinvalue' option.

	* bin/convolve/convolve.c (removepaddingcorrectroundoff):
	Corrected bug: would return *d in both cases! It is corrected now.

2015-06-06  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* lib/mesh.c (chbasedidfromgid): Chaged the old setmeshid to
	chbasedidfromgid along with a new function (gidfromchbasedid). The
	explanations on top of the first clearly explain everything
	extensively. The old function (my understanding) was not yet
	mature enough but now it seems to be very nicely working. All the
	functions that used this were also changed.

	* bin/noisechisel/astnoisechisel.conf: Changed default
	lastmeshfrac to 0.51 from 0.6.

2015-06-04  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/subtractsky/subtractsky.c (avestdonthread): No more need for
	setnan. Since all the meshes are initialized to NaN anyway.

	* lib/mesh.c: Previously, the garrays would be allocated and freed
	on every set of operations that were done on the mesh grid. I made
	some modifications so there is no more need to allocate and free
	them every time. A new prameter was added (ngarrays) so that on
	each use of the garrays, the number of garrays that are needed are
	kept here. All checks to work on the second garray are now done
	through this parameter. In this fashion, when multiple operations
	are to done on each mesh, there is no need to repeatedly free and
	allocate the garrays. Nearly all the functions in mesh.c were
	corrected to account for this.

2015-06-03  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/noisechisel/astnoisechisel.conf: Changed default dthresh to
	-0.1.

	* lib/statistics.c (histogram): Corrected the function. Until now,
	there was a problem when an element was on the last bin: it would
	go onto the next bin and thus would not be counted. This issue is
	now corrected.
	(cumulativefp): Same as above.

2015-06-02  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* lib/mesh.c (operateonmesh): mp->indexs is now allocated in
	makemesh. Because all the programs that will need threads on meshs
	will need it, so it is best to set mp->indexs once and for all in
	makemesh.
	(meshinterponthread): Removed the naninds array, a new option was
	added to interpolate only on blank pixels when needed.
	(preparemeshinterparrays): Similar to the above.

2015-06-01  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* lib/statistics.c: Removed all isnan checks in the comparison
	functions for the minimum and maximum values. NaN values will
	automatically fail all comparisons.

	* lib/timing.c (reporttiming): Removed the `in' from reporting
	seconds.

	* include/timing.h (VERBMSGLENGTH_V): Changed to 45 from 40.

2015-05-27  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/convolve/args.h (options): Remove the `noedgecorrection'
	option. Since it was useless. Also changed the old short option
	for spatial (`s') to `p'. Because `s' is now needed for the mesh
	size option. Also changed the old short option for khdu (`H') to
	`U', to make it similar to all the other programs that do
	convolution.

	* include/mesh.h (meshparams): the garrays are now only
	pointers. The allocated arrays are either cgarrays or fgarrays and
	based on the user's will, garray will be set equal to one of these
	two options. All the functions in mesh.c were changed to
	accomodate this change.

2015-05-25  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* lib/mesh.c (operateonmesh): New name for `fillmesh'. Now the
	function that will be spinned off from each thread can be
	specified from the outside.

2015-05-24  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/imgcrop/crop.c (sectionparser): When only a `*' was given
	with no positive or negative following number, a segmentation
	fault would happen. This bug is now fixed.

	* lib/spatialconvolve.c (sconvonthread): No convolution will be
	attempted for a NaN pixel.

2015-05-23  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* lib/mode.c (modesymmetricity): Corrrected topi to account for
	the case when 2*mi is larger than size.

	* bin/subtractsky/args.h: --numnearest and --kernelwidth moved to
	the `Mesh grid' part of the help output. --checksmoothing and
	--checkinterpolation also moved to the `Mesh grid' part.

	* lib/mesh.c (checkgarray): Modified to show both the full, or
	contiguous over full image, garray and the default one which is
	contiguous only over each channel.

2015-05-20  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* lib/mode.c (modesymmetricity): Corrected topi to the end of the
	mirror distribution.
	(modesymmetricity): If no diff point is found, the end of the
	mirror distribution is used. Before it was the end of the data!
	(valuefromsym): af set to the quantile of the mirror distribution.

2015-05-18  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/convolve/convolve.c (complextoreal): Changed the two states
	to three to add only conerting the real part of an array.

2015-05-15  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/convolve/convolve.c (removepaddingcorrectroundoff): The
	limit to check for floating point errors is now set by a macro and
	changed to 1e-10 instead of 1e-17.

	* lib/statistics.c (sigmaclip_converge): Now, there is a limit on
	the number of times to try for convergence.

	* lib/mode.c (makemirrorplots): All the output paramters are now
	given as arguments so the output can be fully configured from the
	outside. Also, the method that it sets the plot ranges is now
	better configurable.

	* include/mode.h (MODESYMGOOD): Changed to 0.2 from 0.15 based on
	new simulation tests.

2015-05-11  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* lib/statistics.c (setbins): Changed method to find the bin that
	has the value zero within it. Before it would just choose the
	first bin with a positive value. But now, it will check the signs
	of the two interval sides of the bin and when ever the
	multiplication becomes negative, that bin is chosen.
	(histogram): Changed so the last bin is a closed bin, not half
	open like the rest of the histogram bins. This also includes
	floating point rounding error.
	(cumulativefp): Same as the histogram function.
	(sigmaclip_certainnum): Printing is now an option.
	(sigmaclip_converge): Printing is now an option.

	* lib/mode.c (makemirrorplots): New name for the old
	savemodeplots.

	* include/mode.h: Changed input arguments to
	modeindexinsorted. Plots can now be made through the separate
	makemirrorpolots function.

2015-05-05  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/subtractsky/ui.c (preparearrays): Moved the function that
	reads any input as float and incorporates the mask into it.

	* include/fixedstringmacros.h (MOREHELPINFO): Removed long
	explanation about help-gnuastro. Instead an info link to the part of
	the manual that introduces it was included.

2015-05-03  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* include/fixedstringmacros.h (MOREHELPINFO): Corrected to include
	help-gnuastro mailing list.

	* doc/style.css (a:hover): Changed to bright orange color.
	(a:active): Changed to dark orange.

2015-04-24  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/mkprof/oneprofile.c (makepixbypix): When in circumference
	mode, if the truncation radius is reached, it is filled with the
	profile value. This is done so very elongated profiles can at
	least show a one pixel circumference.

	* bin/mkprof/main.h (MINCIRCUMWIDTH): Changed to 0.5f.

	* lib/checkset.c (doublelvalue): Corrected bad value report when
	options are called.

2015-04-21  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/mkprof/ui.c (freeandreport): Avoided double-free by only
	freeing mergeedimgname when it is not equal to output.

	* bin/mkprof/oneprofile.c (makepixbypix): Will only do monte carlo
	integration when the profile is not a constant value.
	(setprofparams): Changed the function of point to `Fixed'.

	* bin/mkprof/mkprof.c (write): Changed to replace and read from an
	image and add profiles on that.

	* lib/statistics.c: All float and double functions can now work
	with NaN too.

	* bin/mkprof/ui.c (setparams): Catalog is always given, no need to
	check.

	* doc/gnuastro.texi: Removed all "Future updates ..." subsections
	in the manual. They are now managed in the Gnuastro project
	webpage on Savannah.

2015-04-13  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* lib/fitsarrayvv.c (updatekeys): When no value is given, it will
	use fits_update_key_null.
	(readwcs): Uses fits_free_memory instead of clib's free.

2015-04-06  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/mkprof/mkprof.c: Corrected condition when a thread would
	finish `build' profiles but the last set of its profiles wouldn't
	pass onto `write'.

	* bin/mknoise/: Changed background flux unit to magnitudes, not
	flux.

	* THANKS: Added institutions which also helped.

2015-04-03  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* lib/fitsarrayvv.c: Changed all occurences of `nul' or `NUL' to
	`blank' or `BLANK'. The FITS standard defines a BLANK keyword for
	integer types.
	(copyrightandend): Will now right extra headers.
	(arraytofitsimg): Can now accept header to write to output.

	* include/fitsarrayvv.h: Changed all `NUL' macros to `BLANK'.

2015-04-02  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* lib/fitsarrayvv.c (changetype): Round floating point input types
	for integer output types.
	(arraytofitsimg): Added numblank as a variable, to deal with blank
	pixels.

	* bin/imgwarp/imgwarp.c (imgwarppreparations): Coorected the point
	(1.0f, 1.0f) to be center of the first pixel, not zero.

	* lib/fitsarrayvv.c (copyrightandend): Removed the Copyright
	notice of Gnuastro from the output image header. Since it might be
	confused with a copyright on the data its self.

2015-04-01  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/imgwarp/args.h: Temporarily removed the wrap option. Will be
	added later.

2015-03-24  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/imgcrop/ui.c (preparearrays): Aborts if in WCS mode and the
	wcsprm structure could not be started.

	* bin/imgcrop/crop.c (firstcropmakearray): Only adds WCS headers
	if the wcsprm structure is defined.

	* lib/fitsarrayvv.c (readwcs): If any error occurs, a warning
	message will be printed and the returned pointer will be NULL.

2015-03-17  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* tests/convertt/: If libjpeg is not present, the tests will be
	skipped.

	* bin/convertt/: libjpeg is now optional.

	* configure.ac: Changed libjpeg to an optional requirement. So if
	it is not available, Gnuastro can still be built.

2015-03-16  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/convolve/ui.c (preparearrays): The flip parts was missing a
	-1.

	* doc/gnuastro.texi: Changed all `gnuastro' occurrences in the
	text to `Gnuastro'.

2015-03-14  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/convolve/convolve.c: Added frequency domain convolution.

2015-03-12  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/mkprof/ui.c (setparams): Added condition for no output given.

	* bin/mkprof/astmkprof.conf: Removed output.

	* bin/imgcrop/ui.c (sanitycheck): Moved check of reentrancy in
	cfitsio to the end so if numthreads is set to 1, there is no
	problem.

	* bin/convertt/ui.c (sanitycheck): Added check if output is not set.

	* bin/convertt/astconvertt.conf: Removed output.

	* lib/spatialconvolve.c: First version complete.

	* configure.ac: Removed sqrt and pthread tests (they are done by
	Gnulib). Added the AX_PTHREAD macro from Autoconf archives.

2015-03-11  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/convolve/: Adding new functions and steps.

	* ./: Copyright transferred to the FSF. Copyright notices in all
	files changed.

2015-03-10  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* tests/imgcrop/*.sh: Set --numthreads=1 for all the tests so if
	CFITSIO is not installed with reentrancy, there is no failures.

	* tests/Makefile.am: Made the fitstopdf result conditional on if
	GPL GhostScript is installed.

	* include/astrthreads.c: Removed the old function and structure
	for replacing the pthread_barrier type and functions that was copy
	and pasted from another webpage and replaced it with my own
	implementation.

	* configure.ac: bug report email set. pthread_barrier checked. GPL
	GhostScript checked.

2015-03-09  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* include/astrthreads.h: Added a temporary function so the
	pthread_barrier type and functions can be run on systems that
	don't support it. I have contacted the author and it will only
	remain if he clarifies the license. I have also asked the Gnulib
	people to try to include it so this issue can be completely fixed.

	* configure.ac: Added checks for nproc and gs along with warnings.

2015-03-02  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/convolve/: Work started on Convolve.

	* lib/fitsarrayvv.c: Changed array types from intXX_t to short,
	long and LONGLONG. Unfortuantely this is how CFITSIO works.

2015-02-22  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* doc/gnuastro.texi: Added index, changed "Science and its
	software".

2015-02-20  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* doc/gnuastro.texi: Removed list of programs at first, added a
	programs index. Added an examples chapter along with some one line
	examples of each program. Moved the Why C and design philosophy
	subsections to the Developing chapters.

	* src: All 'error' functions are now called with a string literal
	and not a string variable. Also all 'system' function calls now
	check the return value.

	* lib: Similar to above.

	* include: similar to above.

2015-02-18  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* lib/Makefile.am: Changed all static libraries to libtool static
	libraries.

	* doc/gnuastro.texi: Updated.

	* configure.ac: Added Libtool checks. Corrected cfitsio and wcslib
	dependencies and removed FFTW as a dependency.

2015-02-17  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/convertt/eps.c: Functions for binary image added.

	* lib/fitsarrayvv.c (nameisfits): Check length.

	* doc/gnuastro.texi (Internal libraries): Removed details of
	libraries.

	* README: Details removed.

2015-02-16  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* lib/statistics.c (dminmax): Works with NAN values too.
	(fminmax): Similar to dminmax.

	* doc/gnuastro.texi (ConvertType): Updated.

2015-02-13  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/convertt/args.h (argp_options): `convert` changed to
	`change`. Since the program name is also convert this could cause
	confusions.

	* doc/gnuastro.texi (Invoking astconvertt): Option explanations.

	* bin/convertt/jpeg.c (savejpeg): Corrected the pixel ordering.

	* bin/convertt/convertt.c (doubleto8bit): Find minimum and maximum
	for all colors to set the scale, not just one color.

	* lib/fitsarrayvv.c (arraytofitsimg): Will not remove comments.

	* include/fixedstringmacros.h (SHORTCOPYRIGHT): Corrected year.

	* lib/fitsarrayvv.c (nameisfits): Dot no longer included.
	(nameisfitssuffix): Added.
	(changetype): Corrected input and output pointers.

	* doc/gnuastro.texi (ConvertType): Added explanations.
	(Future updates): Chapter removed.

2015-02-11  Mohammad Akhlaghi  <mohammad@akbari>

	* bin/convertt/ui.c (preparearrays): Added.

	* bin/convertt/jpeg.c (preparejpeg): Added.

	* doc/gnuastro.texi (Files): Added chapter
	(ConvertType): Added section
	(FileInfo): Added section

2015-02-08  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/convertt: Started working on ConvertType.

	* lib/checkset.c (intsmallerequalto): Function Added.

	* include/fixedstringmacros.h (ASTRUTILSBIBTEX): Corrected article
	name

	* include/checkset.h: Added intsmallerequalto.

2015-01-25  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* ./*: In all files, the name `AstrUtils` was changed to
	`gnuastro` and the `astr` prefix was changed to `ast`.

2015-01-20  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* tests/imgcrop/makerandomcat.py: Removed, because now the points
	are very specific.

	* tests/imgcrop/*.sh: Corrected the inputs to the outputs of
	MakeProfiles tests.

	* tests/Makefile.am (TESTS): Added tests for MakeProfiles.

	* Makefile.am (EXTRA_DIST): Moved the distribution files for each
	folder to the Makefile.am in that folder.

2015-01-19  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/mkprof/ui.c: Steps for reading and checking WCS parameters
	added.

	* bin/mkprof/oneprofile.c (ispsf): Added for a fixed place to
	check if a profile is a PSF.
	(makeoneprofile): Finding mkp->xc and mkp->yc moved here.

	* bin/mkprof/mkprof.c (preparewcs): Added.
	(build): Removed finding mkp->xc and mkp->yc.

	* bin/mkprof/args.h (argp_option): Added WCS related options.

	* bin/imgcrop/imgcrop.c (imgcrop): Number of barriers, corrected
	for when the number of jobs was less than the number of threads.

	* lib/fitsarrayvv.c (addwcstoheader): Changed the WCS structure to
	the WCS header string as input.
	(arraytofitsimg): Convert the WCS structure and remove the
	comments CFITSIO in here.
	(atofcorrectwcs): Added.

	* doc/astrutils.texi: Updated.

2015-01-18  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/mkprof/mkprof.c (writelog): Added.
	(write): Completed.

	* bin/mkprof/args.h (arg_option): Added `--nomerged`.

	* lib/txtarrayvv.c (doformatting): Added option to use `f` in
	printf or `g`.

2015-01-16  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/mkprof/mkprof.c (builtqueue_addempty): Added.
	(saveindividual): Added.
	(build): Is working until the profile is built.

	* bin/mkprof/ellipse.h: Function taken to `box.h`.

	* bin/mkprof/ellipse.c: Function taken to `box.c`.

	* bin/mkprof/args.h (argp_option): added `--numrandom`, changed
	`mginimg` to `psfinimg`.

	* bin/imgcrop/main.h (imgcropparams): iwidth is now a 2 element
	array.

	* bin/imgcrop/crop.c: Removed `borderfromcenter` and
	`correctflpixels` to `lib/box.c`.

	* lib/stats.h: Changed to `statistics.h`, all non-used functions
	removed.

	* lib/arraymanip.c: Removed all non-used functions.

	* include/stats.h: Changed to `statistics.h` and all non-used
	functions removed.

	* include/fixedstringmacros.h (ASTRUTILSBIBTEX): Added.

	* include/commonargs.h (argp_option): Added `--cite` option and
	added the `main.h` and `cite.h` header for each program to use
	their names.

	* include/checkset.h: function name `nameiswritable` is changed to
	dir0file1 which is much more descriptive.

	* include/arraymanip.h: Removed all the non-used functions from
	the past. New functions will be added as they are needed.

	* doc/astrutils.texi: Updated.

2015-01-13  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/mkprof/ui.c (preparearrays): Added.
	(setparams): numthreads initialized to configured value.

	* bin/mkprof/mkprof.c (build): Added.
	(write): Added.
	(mkprof): Parallel jobs configured.

	* bin/mkprof/ellipse.c: Added

	* bin/mkprof/args.h (argp_option): Changed psfprofsinimg to mginimg.

	* bin/imgcrop/imgcrop.c (imgcrop): Used attrbarrierinit.

	* lib/astrthreads.c (attrbarrierinit): Added.

	* doc/astrutils.texi: Updated.

2015-01-12  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/imgcrop/imgcrop.c: Using pthread_barrier_wait for
	synchronizing threads instead of the mutexes and conditional
	variables before.

	* lib/txtarrayvv.c: When the value couldn't be read as a number,
	it would be saved as the lowest possible double value. Now it will
	be saved as a NaN.

	* lib/fitsarrayvv.c (nultovalue): Convert blank values in images
	to a given values.
	(changetype): Change the type of an image.

	* lib/checkset.c (nameisawritablefile): Corrected to accurately
	check if the given name is usable as output.

	* lib/arraymanip.c (converttofloat): Removed conversion
	functions. These were mainly for converting a FITS image type to
	the desired type. Now there is one function in fitsarrayvv.c which
	will do the conversion.

	* include/checkset.h (CHECKCOLINCAT): Added to check column number
	and all values within a column.

	* doc/astrutils.texi: Updated the text, changing names and
	explanations.

	* configure.ac: Changed MockGals, mockgals, name to MakeProfiles,
	mkprof.

	* Makefile.am (EXTRA_DIST): Moved the library header files to the
	Makefile.am in lib/.

2015-01-06  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* bin/mockgals/args.h (parse_opt): Removed the check below.

	* bin/imgcrop/args.h (parse_opt): Removed the check below.

	* include/commonargs.h (cparse_opt): Added check for not calling
	`--setdirconf` and `--setusrconf` together.

2015-01-05  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* doc/formath.texi: Added. A set of macros to write equations
	nicely in TeX, HTML (using MathJax).

	* bin/imgcrop/astrimgcrop.def: renamed to
	bin/imgcrop/astrimgcrop.conf.

	* lib/defaults.c: renamed to include/configfiles.c.

	* lib/checkset.c (stringhasspace): Function added.

	* include/fixedstringmacros.h (SHORTCOPYRIGHT): 2015 added.
	(MOREHELPINFO): Corrected info explanation in help output.

	* include/defaults.h: renamed to include/configfiles.h

	* include/commonparams.h (commonparams): Similar to the change in
	commonargs.h.

	* include/commonargs.h (argp_option): `--dirdefaults` and
	`--userdefaults` changed to `--setdirconf` and `--setusrconf`.

	* doc/astrutils.texi: Included formath.texi to display math
	equations, updates to the text.

	* doc/Makefile.am (headers): Used a shell '*' instead of calling
	each program separately. This was enabled with the new file
	structure.

	* configure.ac: Changed all default names to configure for a
	configuration file, not a default file (only name).

	* Makefile.am: MockGals conditional compilation added.

2014-12-30  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	* tests/basicchecks.sh: New file (common tests for all tests).

	* tests/Makefile.am (AM_TESTS_ENVIRONMENT): Corrected test
	environment.

	* bin/imgcrop/ui.c (setparams): Moved all similar actions to the
	CHECKSETDEFAULTS macro in defaults.h

	* bin/imgcrop/Makefile.am (AM_CPPFLAGS): New (previously in
	AM_CFLAGS).

	* lib/defaults.c (addhomedir): Now returns char *.
	(writelocaldefaultstop): Informs the user of how to make the
	needed directory with `make -p`.

	* include/defaults.h (SAVE_LOCAL_DEFAULTS): changed addhomedir.
	(CHECKSETDEFAULTS): A new macro to read defaults.
	(END_OF_NOTSET_REPORT): Added some extra information.

	* doc/astrutils.texi: A lot of updates in the text! Listing them
	all would be too much!

	* configure.ac: `--with-numthreads` and `--enable-progname` were
	created along with all their necessary checks.

	* Makefile.am: Set conditional subdirectories. Added
	`basicchecks.sh` to the files to be distributed.

2014-12-28  Mohammad Akhlaghi  <akhlaghi@gnu.org>

	GNU Astronomical Utilities created with the ImageCrop as the first
	package. Other packages will be added in the next few days.

;; Local Variables:
;; coding: utf-8
;; End:

 Copyright (C) 2015-2020, Free Software Foundation, Inc.

 This file is part of GNU Astronomy Utitlies (Gnuastro).

 Gnuastro is free software: you can redistribute it and/or modify it
 under the terms of the GNU General Public License as published by
 the Free Software Foundation, either version 3 of the License, or
 (at your option) any later version.

 Gnuastro is distributed in the hope that it will be useful, but
 WITHOUT ANY WARRANTY; without even the implied warranty of
 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 General Public License for more details.

 You should have received a copy of the GNU General Public License
 along with Gnuastro. If not, see <http://www.gnu.org/licenses/>.
