PHOEBE Project ChangeLog: ---------------------------------------------------------------------------- 2004-07-23: - After a successful CVS restauration, a new release is in order: PHOEBE 0.25 testing release is out! 2004-06-29: - A fitting quickbar typo (Origin of time -> Argument of peri- astron) fixed. Thanks to Dirk Terrell for pointing this out! 2004-05-28: - PHOEBE successfully compiled under Fedora Core 2. Thanks to Dirk Terrell for reporting this! 2004-05-26: - Independent variable selection wasn't saved to the keyword file, it is now fixed. Also, HJD is now adopted instead of phase as the default choice. 2004-05-21: - Re-arranged the HJD->phase formula in phoebe_transformations.c to avoid precision problems. 2004-05-08: - The tooltip text of the Base PHOEBE directory in the Configu- ration window was misleading; fixed. Many thanks to Carla Maceroni for pointing this out! - PHOEBE compiles and installs under Gentoo Linux without is- sues. Many thanks to Mike Sallman for pointing this out! 2004-05-03: - Top-level README file updated. - PHOEBE-0.24 released. - PHOEBE 0.24 reported to install and run without issues under Linux SuSE 9.0. Many thanks to Jan Skalicky for pointing this out! 2004-04-26: - Stale WD98 label from the Configuration window removed. - Top-level README file updated for the upcoming 0.24 version. Other top-level files slightly updated. - Top-level AUTHORS updated with Dirk's affiliation. 2004-04-25: - Plot limits for 3D stellar plots increased to +/- 1.3a rather than +/- 1a, since a noticeable fraction of all plots went out of plot bounds. The plot is a bit smaller now, but it shouldn't be too inconvenient. 0.3x version will handle this properly. The limits are changed for both SM and GnuPlot. - configure.in updated with the plotting package test: if both SuperMongo and GnuPlot are missing, configure script stops and requests from the user to install either of the packages. Again, 0.3x version will remove this constraint so that the user will be able to compile only the scripter, but for 0.2x versions this isn't possible. 2004-04-23: - Added support for Portable Network Graphics (png) when using GNUPlot package. - Support for different devices added to Plot-to-File window. The devices are now initialized through a create_plot_to_file_ drop_down_menu () function in phoebe_gui_accessories.c. This function is called at initialization time and every time the configuration is permanently or temporarily changed. - PHOEBE About window expanded to include all authors along with their e-mails and specific work done in PHOEBE. - Added support for additional filters when using the Van Hamme limb darkening tables (Stroemgrem u and y, Johnson K L M N, and Cousins Rc and Ic) - Next PHOEBE version release 0.24 planned for 1st May. 2004-04-22: - GNUPlot checkup now used in plotting functions. - Added workaround lines in phoebe_allocations.c, functions read_in_synthetic_lc_data () and read_in_synthetic_rv_data () to assure correct synthetic data parsing output. - GNUPlot 3d plot now enlarged to use all canvas space. - The implementation of spot fitting didn't support updating the calculated values to the main PHOEBE window. The fix is rele- ased to the phoebe-dev CVS. Thanks to Dirk Terrell for poin- ting this out! - Third light values didn't get updated from the fitting window due to a silly OB1 bug. Fixed. - A new switch introduced to the Luminosities tab that couples or decouples log g1 and log g2 values of the model and of the atmospheres. By default this switch couples both values. It has been added to the keyword files with MODELLGG_ON keyword. - The string of the titlebar of PHOEBE main window changed during startup. 'BinariEs' -> 'Binaries'. Now it stays 'BinariEs'. - CVS write access granted to Dirk Terrell. Welcome aboard, Dirk! :) 2004-04-21: - Top-level configure.in hard dependency on SuperMongo finally dropped. PHOEBE should now compile without it! :) - src/Makefile.am updated to allow compiling without SM libra- ries. Two conditionals govern the library inclusions. - All SuperMongo-related functions #ifdef'ed to SM presence for compiling. - Radio button default in the configuration window changed to gnuplot. - GNUPlot checkup added to configure.in. If it is found, PHOEBE_GNUPLOT_SUPPORT is defined in config.h. All GnuPlot functions should thus check for this switch with #ifdef and #endif clauses. - CVS write access granted to Michael Bauer. Welcome aboard, Mike! :) - When updating true double-precision parameters such as HJD0, PERIOD, DPDT and DPERDT from the fitting window to the main window, the exponential form wasn't accurate enough and the fit, although numerically correct, was useless. This is now fixed. Thanks to Dirk Terrell for pointing this out! 2004-04-20: - Main CVS PHOEBE branch finally (and fully) supports GNUPlot! Many thanks to Michael Bauer for the complete implementation. The gnuplot support is in testing, but should reach stable in a couple of days! This means you can expect the new stable version soon! :) - PHOEBE GUI hooks for plotting packages implemented and tested. Plugging in new plotting packages is now as easy as writting a single package-dedicated function with specified list of arguments. So, if you want to see your favorite package sup- ported in PHOEBE, please consider contributing! :) - Error and exception handling wrappers downported from 0.30 version. They are in phoebe_error_handling.c. 2004-04-16: - Added debugging support to stable PHOEBE version. To compile PHOEBE with debugging enabled, use --enable-debug switch in ./configure script. - IBAND value was undefined if synthetic LC was plotted without any filter selected. It now falls back to the default value, 550nm Johnson V filter. 2004-04-15: - Different locales had different decimal digit symbol conven- tion. In example, German locale uses ',' instead of '.', which caused PHOEBE to crash. Now the locale is changed to "C" in- ternally and restored to the previous state upon exit. Please report any problems, since it's difficult to test all locales. This approach makes certain that the data/keyword file compa- tibility persists over all platforms. 2004-04-14: - Default RV vertices number was set to 10. It has now been in- creased to 100. - The UVLeo.phoebe file wasn't conforming to distributed experi- mental data files. Many thanks to Michael Bauer for pointing this out! - Files atmcof.dat and atmcofplanck.dat were installed as execu- tables rather than data files. This had no consequence for usage, but it is more elegant this way. - If the string was the last inputted value in a particular keyword file, its last character was still truncated. Fixed. Please check your old keyword files for this bug. - PHOEBE 0.23 released. - PHOEBE CVS updated to 0.23 version. This made 0.1x through 0.20 versions obsolete (using WD98). Please use 0.21 or higher for your work. 2004-01-19: - Mode logic made simplex border widgets sensitive, which shouldn't be. Fixed. - When changing mode logic, toggle buttons weren't checked if they were turned on. If they were, now they are turned off. Program logic already handled that correctly, but it was con- fusing for users. - A small bug that left some labels unsensitive while changing mode logic has been fixed. - A readout string that is used to read out widget data was allocated, which mustn't be according to GTK+ standards. This created spurious entries in keyword files (missing parts of strings, incomplete words, wrongly interpreted whitespaces...) 2003-12-17: - Hardcoded /tmp directory name for chi2 plotting changed to the proper temporary directory set in Settings->Configuration. - Symbolic linking to atmcof.dat and atmcofplanck.dat from the working (temporary) directory incorporated. Thus copying the two files isn't necessary anymore. However, this potentially reduces portability due to symlink () function usage. In the next PHOEBE version this will be handled differently. - Old e-mail changed with the new one in toplevel AUTHORS. - PHOEBE 0.21 released. 2003-12-16: - Administrative work for the new release finished. 0.21 will soon be released. 2003-09-18: - Some WD-related limitations set on spot parameter switches and a bug in PHOEBE removed (wrong KSP* setting). - Secondary spot values carried " signs, which were removed. 2003-09-16: - PHOEBE adopted WD 2003 version as the main back-bone. All functions changed and updated accordingly. 2003-06-10: - There was a problem in changing parameter values of the type string - the spaces weren't parsed correctly. Fixed. - Two generic scripts - one for detached binaries and the other for overcontact binaries - added to the scripts/ directory. The whole directory put under the CVS. The install script will install this directory to /usr/local/share/phoebe/scripts. 2003-06-09: - Normalizing magnitude parameter (MNORM) in Data tab had way too many decimals. From 5 it is now reduced to 1. - A script function Estimate Normalizing Magnitude () added to the scripter. It doesn't take any arguments. It calculates the median of all LC input files and returns the average. - If a keyword file was opened from within a script, the speci- cations info list didn't get updated. Fixed. 2003-06-08: - Due to the opening of the default keyword file at startup, the main window title got erased. Fixed. - Finding tokens with strtok () function modified the initial pointer, thus changing the contents of the Scripter window. By copying the contents of the original parameter string to a temporary token string, this problem is circumvented. - A small bug in startup line parsing called OpenKeywordFile function on a script if the script name was the last parameter in the command line. This was annoying because if scripts were called from the command line, a notice window popped up, comp- laining that the keyword file is a legacy file. Fixed. 2003-06-07: - Keywords LC_INDEP, LC_DEP, RV_INDEP and RV_DEP accidentally left out from keyword files. Fixed. - A new scripter command added: Adjust Parameters. The argument may be composed of many PHOEBE keywords separated by a comma. This means that PHOEBE may be used for something useful from a script - while it's working, you may actually get a cup of coffee! :) 2003-06-05: - Further work on PHOEBE tutorial has been done. First 18 pages are up-to-date for 0.20 version and basically ready to print. 2003-06-04: - Plot LC to File and Plot RV to File scripter commands lacked the filter name argument. Fixed. - Entry contents are now flushed each time the scripter command is changed. It just didn't look useful to preserve it. - Top-level SPECS file updated to contain instructions on adding PHOEBE keywords and scripter commands to PHOEBE. - A new keyword, LC_FILTER, added to PHOEBE. - Argument parsing is now done with strtok () function, which simplified this routine enormously. Why didn't I think of this before? - PHOEBE keywords RV_PLOT_SYNTHETIC, RV_PLOT_EXPERIMENTAL and RV_FILTER added to PHOEBE. - defaults/default.phoebe updated to the new format. PHOEBE now calls this file during startup, so changing this file changes startup values permanently. - UV Leonis photometric images are in fact acquired with Bessell filter set, not Johnson. Headers of both files have been fixed accordingly. Small headers added to RV data. - A heavy fixup has been done on the PHOEBE tutorial. - Bessell filter set incorporated in filter drop-down list. - A typo in Component-Related (capital R) tab label fixed. 2003-06-03: - Scripter backbone has been completely rewritten. The commands are now declared only on one place and from there they are added to the GUI. To extend the command list, one simply has to add a proper definition of the command. This makes room for user-generated scripter commands (also loops and run-time controls). 2003-05-29: - CLA adjust button state wasn't saved to the keyword file. It is now fixed, but only for the new keyword file format. - New keyword files are now fully parsed. Everything is now the same as before, only much simpler and more efficient. - Notice window added to notify the user when legacy keyword files are opened. - Top-level README file changed to contain this information. - PHOEBE has been promoted from "testing" to "beta"! The next version will be 0.20! - Autoutils' "missing" script added to autoutils directory and to the CVS. - Author's e-mail updated in the About window. 2003-05-28: - Cleaning-up of scripter routines has been performed. - A MAJOR CHANGE: keyword file format has been changed. This was done to conform to ISO/BASH standards. The structure of this file is now: KEYWORD=VALUE for numbers and KEYWORD="STRING" (quoted) for strings. This enables proper readouts from scripts as well. The routine for backward compability will remain for a year, but it will issue a warning of format change with an option of automatic conversion. - Log g and Metallicity widget ranges added to keyword files. - Keyword files' handling heavily rewritten. Everything has been cleaned up, routines should run even quicker. - Implementation of keyword saving isn't yet complete, so the CVS won't be updated. 2003-05-26: - Two new keywords added: LC_PLOT_SYNTHETIC and LC_PLOT_EXPERI- MENTAL. These two are booleans that govern the state of LC plotting window toggle buttons. They may be changed like any other widget value, where the argument has to be TRUE or FALSE. - Boolean type recognition in script wrappers implemented. - A script warning wrapper implemented. If a script calls a macro that isn't recognized, PHOEBE will complain about it, but will skip it and try to complete the script. 2003-05-24: - Change Parameter Value scripting function fully implemented and working flawlessly. 2003-05-23: - Change Parameter Value (keyword, value) added to scripting. The prototypes are fully implemented, but the actual script function still remains to be written, since the next step is to automate the change through PHOEBE keywords. 2003-05-21: - Working with global keywords furtherly simplified the work with data sheets. Flushing all parameter values to the sheet is now done fully automatically, so the value gets appended simply by adding a new keyword to phoebe_global definition. Printing out relevant parameters is still governed by hand, for obvious reasons. - Spectra number (keyword SPECNO) added to Open/Save keyword files. 2003-05-20: - Adding keywords to data sheet from scripts fully implemented. The user should use keywords listed in phoebe_global.h. There should be a table of these keywords with short descriptions in PHOEBE reference manual. 2003-05-15: - Gravity potential limitation values were too tight (their up- per limit being 20). It has now been increased to 1000. - Support for adding individual parameters to data sheet from within a script implemented. Parameter recognition still has to be properly implemented. - A lot of widget-naming cleaning up has been done. The inclina- tion keyword XINCL permanently changed to INCL. - A new global enum created that holds all PHOEBE keywords. The- se keywords have unique corresponding widgets which are atta- ched in a single function in accessories. Thus if a widget na- me changes, the only change imposed is in that function. 2003-05-14: - More work on Data Sheet window has been done, but it's still not appropriate for everyday usage yet. - A prototyped support for adding entries to the Data Sheet from the scripter implemented. Routines first have to be cleaned before I connect them to their callbacks. - Saving keyword files now uses generic file selector widget, so the original widget is deleted. A signal clean-up function is introduced that makes sure all signals are removed when a user cancels any action in file selector window(s). - Data Sheet printing to files implemented. It is called through a wrapper function (everything_ok_with_the_filename), which isn't yet written, so use *real* filenames, otherwise PHOEBE will crash. - Data Sheet scripting implemented. Currently the user may flush the contents of relevant data sheet to the given file. One may not yet make custom data sheets. This is next to be done. 2003-05-12: - A script for saving files added and is fully functional. - A tester remnant which outputted spots' parameters to terminal has been deleted. - Data Sheet prototype function add_parameter_to_data_sheet_list has been implemented. It is called much like the keyword hand- lers. - First couple of entries already put to the Data Sheet window. 2003-05-08: - A new generic widget PHOEBE_file_selector created. This window has only "delete-event" callbacks attached to the kill button and to the cancel button. The rest has to be assigned within the calling routine along with the actual callback that per- forms the specific action. See "on_scripter_open_script_*" as an example. - Open and Save script buttons in Scripter are now fully func- tional. For fool-proofing a new routine has been added to phoebe_gui_accessories.c called everything_ok_with_the_ _filename (char *filename). It is empty for now, but as it should perform generic tests, all openings/savings should be called only if this wrapper returns 1. This thus deserves my full attention for quite some time, so I'll be delving into this when I'm a bit less tired. ;) 2003-05-07: - Script name given in command line is now parsed and checked for existence; if it doesn't exist, PHOEBE will exit with the fatal signal. - Quitting PHOEBE from scripter now cleans after itself (all PHOEBE related /tmp files). 2003-04-26: - I'M MARRIED! :) 2003-04-22: - Keywords LC_VERTICES and RV_VERTICES added to the keyword fi- le. It works both for opening and saving keyword files. - Two new actions added to scripter for calculating gravity po- tentials. The calculated values are then inputed to their main PHOEBE window values. 2003-04-20: - Opening/Saving scripts algorithms written. GUI part is not fi- nished yet, but all read-in fool-proofing has been added. The user must be careful to write the proper action names, though. - A new option, "-s" or "--script", followed by the script name, added to PHOEBE command line. If one wants to execute a script named test.script, one would issue phoebe -s test.script (the script has to be in that directory, otherwise one must supply a path to it as well). There is no fool-proof support yet, I have to test it furtherly first. 2003-04-19: - Data sheet window created; it is empty for now, but all call- backs have been connected. - PHOEBE scripter completely rewritten. GUI-related stuff is completely implemented and fool-proofed, action-editing but- tons fully working. Four scripts already implemented: opening keyword files, LC plotting, RV plotting and quitting. It should be stressed that the contents fool-proofing is up to the user at present, if something is wrong, scripter will ca- use a segfault. - Four new keywords added to keyword files readout: LC_INDEP, LC_DEP, RV_INDEP and RV_DEP. These are introduced to easily manipulate what the user wants from scripts - all it takes is to create a keyword file containing *only* the parameters the user wants changed and opening it from a script. - A new documentation file, keywords.gnumeric, created. It con- tains all keywords that PHOEBE recognizes with concise desc- ription and remarks. It may be readily exported to postscript. 2003-04-17: - Filling issue resolved; PHOEBE issued a warning every time the solution "spilt" over L1 or L2, although it should issue this warning only when L2 is exceeded. Thus a new function for cal- culating critical L1 and L2 potentials is introduced and a new callback is attached to all widgets that are relevant to these critical values (RM, F1/2, PHSV/PCSV and E). The LC/RV plot- ting windows are extended to contain these values together with the information on exceeded state. The warning on L2 overfill is still kept in PHOEBE. 2003-04-16: - WD98 Notices temporarily disabled to sort out the L1/2 filling issue. 2003-04-09: - Fitting implementation for spots is now finished. However, up- dating still isn't done, since I first have to test the imple- mentation. 2003-04-07: - Further work has been done on spots. The GUI for adjusting spot parameters is finished and fitting callbacks provided. It still lacks the step values, but I'll get around to do it in a while. Fitting isn't yet fully implemented, this comes next, but not today. ;) 2003-04-03: - A new programming philosophy concerning memory allocation has been adopted: allocations and de-allocations of memory from now on *must* be performed within one function. If the func- tion would become bloated, *two* new functions may be intro- duced: allocate_memory_for_* and free_memory_allocated_for_*. Both of these newly created functions *must* be also called from within the same parent function. This way memory leaks shouldn't occur anymore. All error handling should also be performed in these supplementary functions. - Spots part completely rewritten to reflect the above philoso- phy. Multiple spots are now available (as opposed to only one spot situation acting until now) on both stars. All spots' parameters are now saved to and opened from keyword files. - Semi-detached binary with secondary star filling the Roche lobe wasn't constrained properly: both potentials were turned off. This was reported by Frederic Royer and is now fixed. 2003-03-26: - A small bug fixed which prevented both RV curves to be plotted simultaneously. It was introduced 30 minutes ago and is now fixed, so noone should really suffer from it. 2003-03-24: - There was a problem if WD98 reached an internal error (such as the temperature range being exceeded in ATM routine) before anything else: PHOEBE didn't check for successful allocation before trying to write out parameters into their frame and thus segfaulted when doing so. Fixed. 2003-03-21: - Data info table sorted out completely. It is now working pro- perly, but will remain in the Utilities tab as long as I don't decide where (and how) to put it permanently. Any suggestions are more than welcome. 2003-03-20: - Synthetic noise adding yielded yet another undocumented com- ment in WD98 output files, which caused PHOEBE to segfault. All readouts have now been changed to ignore all unexpected notices which occur before/after data. If they occur in the data itself, then something's very wrong and should be repor- ted/inspected. - Synthetic noise frame contents have been rearranged; the noise switch limited the Gaussian noise to scale only with the squa- re root of the light level, now the user may choose the weigh- ting as desired. Global switches struct was changed to make this possible. - GUI layout has been changed permanently. No major changes are expected to occur in three tabs: "System-related", "Component- related" and "Luminosities". Other tabs should be furtherly optimized. - Plotting routines have been cleaned of individual variables for seed and sigma in their declarations. Everything now goes through typedef'd structs. - A new configuration switch added to enable the user to disa- ble the warning on synthetic scatter applied to LCs. - The default SEED value changed to 1.5E8 instead of 1.0E8 to avoid underflowing the expected format in WD98 if a user for- gets to create its own seed. - A summary of important data on input data files, such as the number of points, the sum of all points, the median etc. will be given in a table that will get updated automatically. The prototype for it currently resides in the Utilities tab, but it will be moved to a separate window as soon as it is fini- shed. - The default directory for assigning data files changed to the data directory instead of the base directory, which appears more logical. - The default sigma value for input data files changed from 0 to 0.1, because fitting failed if sigma was accidentaly left at 0. - First automatic file analysis routines have been implemented, but are far from stable. If you downloaded today's CVS, please update it as soon as possible again, because stability is really low right now. I hope to update the CVS once more to- night. 2003-03-14: - Further changes on GUI have been performed. It still isn't fi- nal, but very close to. - Double contact binaries have both Omegas constrained, but un- til now only Omega2 was constrained. Fixed. 2003-03-13: - A major change in PHOEBE tabs has been introduced; tabs Basic and Advanced didn't make much sense, really; they will be re- placed with new (similar) tabs, one of them will contain all physical parameters of a binary as a whole and the other will contain individual, star-specific parameters. I still haven't decided upon their names, but probably something like General and Individual. I have to consider other naming conventions, though. Anyway, a General tab has already been implemented and all callbacks now point to new widgets. Old ones have been erased, although tabs still pertain. The reason is that I want to make a permanent, thus optimal outlook and I'm experimen- ting what looks good and what doesn't. If you have any sugges- tions, please let me know. - Another configuration option is implemented: the user may choose whether PHOEBE should ask each time for overwriting confirmation on Save. - While I was at it, a confirmation on quitting is also added to the Configuration window. - PHOEBE's main window "delete-event" was connected to gtk_main quit instead to a wrapper. Fixed. - A segfaulting bug that appeared when invalid keyword filenames were chosen (e.g. directories, devices, ...) fixed. 2003-03-10: - Aliasing was poorly implemented and it didn't alias some data points properly (for all-positive phase intervals). It has now been completely rewritten (simplified) and fully tested. - A tiny bug in Clear button in LC plotting window tried to era- se a non-existent widget. This only caused a warning to be displayed in the terminal, but it is fixed now. - A temporary tab is introduced in the main window. It will replace the Basic tab eventually, when all input entries are accordingly update it. Please ignore it for now. - A serious problem was discovered and reported by Frederic Ro- yer: WD98 algorithm sometimes suppresses the calculation of synthetic curves because of internal problems, such as excee- ded temperature range and such. The notice windows are now added to inform the user such error has occured. There is no workaround for this problem, it has to be communicated to R. Wilson. 2003-03-06: - A new source file pair "phoebe_scripting.c.h" introduced; it will contain all scripting routines. These scripting routines must perform basic actions such as opening/closing/plotting/ fitting/... with only one call with arbitrary arguments. This way the user will be able to write his own scripts. 2003-03-05: - Command line parsing implemented. Standard arguments -h and -v have been initialized (with their longer (--) counterparts). Parsing is done in phoebe_gui_accessories, but only to the ex- tent of writting to a global PHOEBE_args variable. This is then parsed furtherly in phoebe_init() function, after all re- levant variables have already been initialized. This goes es- pecially for keyword files given at input, which are now fully supported. Extending command-line options is simple. This means that PHOEBE will soon be able to run from batch scripts. - Filter management rewritten. Combo boxes are no longer used, because they got really crowded. Everything is now accessed through a button click and a pop-up menu bar with menues, sub- menues etc. Geneva filter system is now fully active. 2003-03-03: - CVS repository moved to /var/lib/cvs at Deneb; README file has been updated, PHOEBE homepage as well. Permissions have been tightened, although access is still available via :pserver:. - acconfig.h that is needed for make dist script was missing from Makefile.am. Fixed. 2003-03-01: - Plotting package directory removed from Configuration options since everything is now called through libraries and not exe- cutables. - Configuration file indentation changed by 1, to be aligned with the longest tag-name. This doesn't have any impact on the performance, it just looks better. - Configuration file keywords' readout has been fool-proofed. It should now be resistant to nonsense content. - Tooltip description for the defaults directory changed. - file_exists() function restricted detection only to those fi- les which the user had the full permission to handle (644), so the files in /usr/local/share/phoebe weren't properly tes- ted. Fixed. - A bug in configuration path assignment assigned a FF path to DC path, which caused wrong DC path *only* during the first- ever start of PHOEBE. Fixed. - When PHOEBE is installed, the working directory was set to the base directory, which is /usr/local/share/phoebe. Any user without root priviliges cannot write to that directory and that is why image generation failed. It is now fixed and fool- proofed (a notice window pops up if priviliges are missing). - Fool-proofing on directory names, priviliges and content has been implemented for Configuration options. - PHOEBE-0.15 released. It is substantially lighter that its precessors, because of the make dist rule which puts only the most relevant files into the distribution. 2003-02-27: - Location of pixmaps changed from share/pixmaps to pixmaps and it was merged with icons. This was done to furtherly clear up the root directory. All source files have been updated accor- dingly. - Base directory for installation changed to /usr/local/share. Within there will be a phoebe directory with slightly changed hierarchial structure as in the main distribution. All defa- ult values for .phoebe have been changed accordingly. - Install script is now in place. Makefile.am-s have been add- ed to all relevant subdirectories that should be installed: bin, include and pixmaps. - share/ directory renamed to defaults/ and it contains only defaults.phoebe file. It is used for restoring defaults in the main PHOEBE window. It has its Makefile.am. - Directory flux_files added to the main tree along with Make- file.am. - doc directory also added to the install script. - make install script works completely! There are still some small pixmaps' issues that need be resolved, but it's just a minor problem! Enjoy! - CVS server is currently unavailable due to maintenance. All changes will be commited after the server gets up and running again. 2003-02-24: - FEEDBACK file was outdated and was deleted from the distribu- tion. - File share/phoebe_init.sm deleted from the CVS, since it had no purpose. Same done with icons/under_construction.xpm, be- cause another instance of this pixmap was already in shared pixmaps directory. - Top-level Makefile.am finally adopted for automatic distribu- tion creation by issuing make dist in the top-level directo- ry. Install script has yet to be implemented properly. - Data files for UV Leonis have been updated (new RV curves). 2003-02-23: - Underdimensioning problem was present in weighting choices' names (defined 20, but contained all the way to 32 charac- ters), which caused PHOEBE to segfault if compiled with older compilers (gcc 2.95 or lower). All static dimensions have now been set to 255 to avoid any future problems of this kind. - 3rd light edit button had an incomplete label, since it con- tained opacity settings as well. It has been changed accor- dingly. - RV Plot Clear button referred to old (non-existent) widgets, which caused GTK-WARNINGs. Fixed. 2003-02-20: - Geneva photometric system added to filter list. This was requ- ested by Frederic Royer. Full implementation will be done af- ter defreezing PHOEBE. 2003-02-19: - PHOEBE development moved to Debian Linux (Woody 3.0r1). RedHat policy on new versions proved to be unstable and file locati- ons tend to change constantly. - README file updated, some minor changes in include directives have been made. 2003-02-10: - Finally RV plotting is completed. Plotting residuals now works for all different dependent variables, sigma is calculated for both curves as well. Hopefully nothing is left to implement at the time. - PHOEBE root directory got very cluttered, so all autoutils-re- lated utilities have been moved to the new autoconfig direc- tory. - GdkPixbuf library was used to put graphics on screen in gif format. This library was missing on some older systems and was thus abandoned for the time being in PHOEBE, because I believe in simplicity above all. All configure and make hooks have also been removed. Consequently there are no more installation issues related to this library's directory. - Configure script now checks for SuperMongo libraries. If they aren't found, configure script stops with an error. 2003-02-07: - Save and Save As buttons finally work as they are supposed to. If the filename isn't given (i.e. parameters inputted to de- fault startup screen), Save will prompt the user for filename, otherwise it will overwrite the current keyword file. On the other hand, Save As will always prompt for a filename and ask the user to confirm overwritting in case the chosen filename already exists. Save option has a shortcut CTRL+S. It would be beneficial to get feedback from users on which button is more useful in the button box: Save or Save As or maybe both. - Main PHOEBE window title now contains star name. A callback is connected to the Star Name entry field and whenever it is changed, the window title changes too. This is very useful to keep track if multiple PHOEBEs are open. 2003-02-06: - Input file parsing rewritten. Comments '#' are now completely supported: they may occur wherever and only the remainder of the line is discarded, everything else stays. Empty lines are also legal (as of now), arbitrary formatting is still preser- ved (use spaces or tabs). - Function parse_input_data introduced to accessories. This fun- ction is to be called whenever input file parsing is necessa- ry. - OB1 error in reporting data points number (1 too much) fixed. 2003-02-05: - PHOEBE goes into freeze state. This means that no new features are going to be implemented until the next release (0.15). In the next couple of weeks only code cleaning and bug fixing will be performed. - Code cleaning: obsolete PHOEBE_data definition was finally de- leted and new_PHOEBE_data was renamed to PHOEBE_data. This was a global change, so all source files are affected. 2003-02-04: - Configuration window has been improved. The main interface now contains tabs. Currently there are two tabs: Files and Options tab. Files tab holds whatever was there in the first place and Options tab contains (will contain) options. Right now there is only one option implemented: to plot or not to plot star images automatically when phase is changed. For slower compu- ters this seemed necessary to have, since it would take ages to compute the image every time. For faster computers this switch should of course be on. - Plot to File completely implemented and tested for star image plotting. Data for individual devices get rescaled because of aspect ratio problems in SuperMongo. 2003-02-03: - Star images are now fully supported. Everything works! Spots are displayed, all phases, there is a callback attached to the phase spin button that when it gets changed, the image is au- tomatically updated. Plot to File still isn't implemented and I am considering removing the Calculate button and exchanging it with Animate button. But this is long-term. 2003-01-24: - log g and metallicity for both stars added to main typedef. Their values are read in through the read_in_main_parameters function. 2003-01-23: - A median calculation routine has been implemented. Two new fi- les have been introduced: phoebe_calculations.c and .h. These hold all mathematical calculations performed in/by PHOEBE. - Simplex limits-to be are shaded out not to confuse new users. - Center-of-Mass velocity (VGA) moved to the orbit tab. - log g and [M/H] added to the Basic tab. - Teff configuration added to Utilities tab. Probably it will eventually become its own window. - Small bugs in configuration naming routines have been fixed. PHOEBE_CONFIG global variable contains a config filename, so phoebe.config, although generic, may be easily changed from the phoebe_gui_accessories.c. This may be eventually used for branching. - DCI files have been enumerated. They now follow the same pre- scription as the LCI files. This laid foundations for undoing in PHOEBE. - Dummy readout variables exchanged for %* switch. - "Flux Files Directory" added to the configuration window. It will remain disabled until the official flux files' support. All callbacks are connected, though. Saving/opening works. - Log g and metallicity values in keyword files were up until now in LD interpolation windows. Now they are in Basic tab. Values within interpolation windows are attached to the ones in Basic tab, e.g. they change on changed event. - About window version label changed to generic form in glade file. This has no impact on PHOEBE, just looks cleaner. 2003-01-21: - Import from DCI and Export to DCI entries have been added to the main menu, signals connected, but no code was yet written. - Luminosities tab finally set right. I'm pretty sure that after the today's change this tab will hardly be modified in future. It looks so nice and I'm sooooo happy with it! Aren't you? - Binary star image implemented in Luminosities tab. Another plot device, starxpm, defined in SuperMongo's graphcap.local. README file updated accordingly. Callbacks have been connec- ted, but actual routines still have to be written properly. As default, there is a nice UV Leonis plot at phase 0.4. - Main PHOEBE window was initialized before icon pathnames were set, so it failed to load pixmaps if run from any other di- rectory than src. Fixed. 2003-01-17: - Function read_in_data_file_info moved to phoebe_allocations.c from callbacks. - There was another units' problem in RV sigma, when RVs were inputted in km/s. Sigma was in km/s and internal data in 100 km/s. This must be improved when automatic rescaling will have been done. - HJD0 fitting didn't work properly because of spin button <--> entry conversion failure. Fixed. - When data files were removed, filter combo boxes didn't get updated. This caused PHOEBE to crash when the user selected the non-existent filter. Fixed. 2003-01-16: - System constraints have been fully implemented and tested. - Empty LD coefficient defaults changed from 1.0 to 0.5. It ma- kes more sense not to have a linear cosine proportion for an average star, 0.5 seems to be better. - A bug which prevented aliasing to work when both RV curves are simultaneously plotted has been fixed. 2003-01-14: - In experimental LC data absolute error (sigma) is now accepted as a weighting function. The transformation function takes this value and calculates 1/\sigma^2, then maps it on the weighting interval [0.1, 9.9]. These are internal weights. - RV luminosities info list deleted. Although there is support for RV HLAs and CLAs, I can't think of a single reason why it should be there. If you think of a reason, please let me know, I will support it again in an eye-blink. - Luminosities tab heavily rearranged. "Luminosities" name for HLA and CLA have been changed to "Level", since that is what they are, although of course still connected to luminosities. Opacity function has been introduced as a user-supplied para- meter which cannot be adjusted. It is wavelength-dependent and it accompanies third light in its info file. Some space is left spare in the tab for future additions, most probably 3D picture of the binary system. 2003-01-13: - PHOEBE 0.14 released. - Chi2 plotting routines have been thoroughly cleaned. Plot to File now works. This change will reflect in the next PHOEBE version and the CVS. 2003-01-12: - A small typo "differencial" changed to "differential" in the fitting tab. - There was a remnant of an LC function callback that crashed PHOEBE if magnitude was selected as a dependent variable. It has now been deleted and the problem is fixed. - A typo in LC plotting window: Experimental LC curve changed to Experimental light curve. - Extensive work on PHOEBE tutorial has been done. 2003-01-11: - Quit warning added; when the user clicks on the exit button or activates the Exit menu from the menu bar, PHOEBE will no longer blindly exit, but will prompt the user again. 2003-01-08: - LC Plotting window completely rearranged; all wavelength-de- pendent widgets have been deleted and a generic look was im- posed. RV plotting window already follows this scheme. - CTRL+q accelerator added to RV plotting window. - Chi2 functions tidied. - LC plotting functions and RV plotting functions have been mo- dified to be fool-proof without the need of shading certain widgets. If the settings are unfeasible, PHOEBE now displays a message (a notice window) about the problem. 2003-01-07: - Chi2 values have been replaced by weighted and unweighted sig- mas (standard deviations) in plotting windows. This is more meaningful to estimate offsets than chi2-s. This was suggested by Tomaz Zwitter. The routines for chi2 calculation are still preserved in both weighted and unweighted form. One accesses them by calling calculate_chi2 function with different "type" switch. - RV plotting window frames slightly changed. 2003-01-04: - Experimental data points (aka input files) weren't transformed correctly from HJDs to phases if DPDT was different from 0. Since this isn't totally trivial, I have added an appendix ex- plaining this to the tutorial. 2003-01-03: - HAPPY NEW YEAR! PHOEBE 0.14 will be out within the next week! - DPDT and DPERDT input formats have been changed to entries rather than spin buttons because of the desired double preci- sion. This will soon be obsolete (with GTK+2.0), but as an old branch this will be used as a workaround. - DCO readout values changed from F format to D format because of higher accuracy. Accordingly, DC output has been changed from %lf to %G (the conversion automatically chooses between floating-point and exponential form). - Automatic resizing of the correlation matrix window has been implemented, although not really elegantly. This could be im- proved with some dedicated GTK functions. 2002-12-26: - Step size in DC routine for orbital inclination was wrong be- cause of the OB1 error. Fixed. - "Update Selected" button in the DC window erased, since it doesn't make any physical sense to update only some parameter values, for this would certainly reduce the fit quality. The list is also changed to the "Single" mode. - Correlation matrix inspection fully supported. There is a button in the DC window that accesses the correlation matrix of the fit. It has been thoroughly tested and only one small issue persists: window dimensioning. It should be automati- cally dimensioned according to the CList size, but only if that size doesn't exceed some maximum size. If it does, it should use a scrolled window. At this point scrolled window is used for all CList dimensions. 2002-12-19: - DCI files were until now created in phase-space, because the minimization is significantly faster. Now the user may choose between phase and HJD for independent variable. It is done through a toggle button in the Fitting tab. 2002-12-18: - Constraints for mode 3 (overcontact binaries that are not in thermal contact) have been added. 2002-12-16: - RV plotting window completely changed. It seams that the HLA, CLA, VGA, ... values there were completely useless (they were originally provided for synthetic values (these were all wa- velength dependent quantities) and now they are fixed to their default values. If one wants to experiment with synthetic-only solutions, one may use the dummy input file. Please beware that the widgets aren't yet fool-proof. - Support for simultaneous plot of both lightcurves added. It has been tested preliminarly and indeed there are some mi- nor glitches with zooming and chi2 calculation, which have to be fixed as soon as possible, but they only occur at very unusual circumstances (e.g. 10000x zoom etc). 2002-12-12: - VGA description in the main PHOEBE window changed to prevent definition ambiguities. VGA should be given in km/s, regard- less of the units in input files (100km/s input is supported only because of earlier WD input files compatibility and ho- pefully everyone uses km/s these days!). - "Adopted units" in RV plotting renamed to km/s. It just isn't obvious to me that anyone would like something different. The output is limited to km/s accordingly. Choices have been re- arranged slightly in the dependent RV combo box. - "Systemic vel." label in RV plotting window changed to CoM RV. - switches.VUNIT variable erased; internal RV units will now be cemented to 100km/s (for WD98, invisible to the user) and shortly the units will be dynamically calculated (by log-ging we will measure the dimension and tune the units accordingly). - VGA DC calculation is now properly matching I/O units. 2002-12-11: - When a keyword file was saved, the file lists weren't updated and thus the user didn't see the newly created files, although they existed. Fixed. - TODO list updated. - Error handling improved on keyword files saving. If a user clicked on Save when there was no output file given, PHOEBE crashed. Now it issues a warning that a problem has occured and lets the user supply the filename. - Two new source files created: phoebe_configuration.c and .h. They contain all configuration-related functions, which have previously resided in callbacks. - OK button in the configuration window was functionless, it just hid the window. Now it applies the currently selected settings to the running PHOEBE session. These settings last until restart of PHOEBE or until the subsequent change in configuration, as is supposed to do. 2002-12-10: - Further work on 3D chi2 implementation has been done. Now routines plot to the main PHOEBE window instead of a tempora- ry file. The axes are still missing, but everything else works. - Symmetric derivatives switch was off by default. This is not the usual case, so it is now changed to be on by default. 2002-12-03: - Effective filter wavelengths, when changed in the Data tab, weren't changed in the limb darkening interpolation routines. This caused some limb darkening coefficients (most crucially Tycho/Hipparcos series) to be interpolated wrongly. Fixed. - Another problem persisted in interpolation routines: if there were multiple data files in the same filter bandpass defined (thus #num appended to filter names) LD routines didn't reco- gnise these filters as supported ones. Fixed. 2002-12-02: - There was a bug in the DC fitting window: if the user had only RV data or only LC data, the Update All procedure crashed PHOEBE because it tried to update the plotting windows. Fixed. - Major work has been done to incorporate spots into PHOEBE. GUI has been completely implemented, global declarations up- dated, read_in functions modified, a couple of new counters introduced. Yet to be done: keyword files open/save options, array dimensioning for spot parameters, DC fitting incorpora- tion and extensive testing. 2002-11-29: - After a long pause in development due to different article writting, I came back. A minor but significant change: the magnitude LCs are now calculated from the total flux and the given MNORM value in the Data tab. This rendered the MZERO switch obsolete. All routines cleaned to reflect this change. 2002-10-09: - There was a bug in PCSV calculation: q instead of 1/q was used and the calculated value was wrong. Fixed. - Phase ranges for LC and RV plots are now saved to keyword fi- les as suggested by Tomaz Zwitter. 2002-10-02: - VGA adjustability switch should not be available if there are no RVs present, because DC ignores it and this gets messy in the fitting window. Fixed. 2002-10-01: - UV Leonis LC data files changed: individual weights are added to the 3rd column based on nightly sigmas. The keyword file has been changed accordingly and the tutorial has been modi- fied to reflect this change. 2002-09-30: - The upper limit for the lightcurve spin button was set to 15, which doesn't really make sense, because there is practically no limitation for that value. It is changed to 100. If some- body needs a larger value, please let me know. 2002-09-26: - Reddening option added to PHOEBE Tutorial. 2002-09-24: - Some widgets were named in a VERY outdated fashion. Fixed. - Under Construction pixmap initialization moved from .glade file to phoebe_gui_accessories.c. This is done to prevent path-related problems that occured when PHOEBE is started for the first time. - Due to momentary lapse of reason I haven't updated the CVS documentation in ages. Now the newest phoebe_tutorial ver- sion is available for download via CVS. - PHOEBE HomePage updated to contain links to the tutorial and Van Hamme LD Tables. - Chi2 Plot to File button was sensitive although the Calculate button wasn't. Fixed. - Reddening may now be removed from within PHOEBE. The switch is placed to the LC Data Window and may be naturally used only when the dependent variable is magnitude. The input quantities R (the correlation coefficient) and E (B-V) (the color excess) are saved to the keyword file. These parameters have also been added to the switches typedef. Reddening is now hopefully com- pletely supported. 2002-09-20: - LC Plotting window outlook changed a bit. The upper part has been shrunk and a new check button has been added: Plot Re- siduals. - Plot Residuals functions have been implemented. There is a problem with aliased interpolation, but nothing can be done until BBL routine from WD98 is called directly from PHOEBE. This manifests in some weird outliers at the edge of the plot. Otherwise residuals work; the calculation depends on the plot settings (independent vs. dependent variable), it is *not* ge- neric. However, it is difficult to estimate the "usefulness" of the HJD residuals because of heavy extrapolation; this is why only phases are supported as independent variables at the time. - Sensitivity of the THE parameter was accidentally on initial- ly. Fixed. - A menu item created to hold all filters. These will eventually be connected to their flux files. This way it is much easier to select a particular filter than it was by browsing through a combo box. The menu is attached to the test button in the Utilities tab for now. 2002-09-19: - Fitting QuickBar connected to all main switches and values. It is now completely usable. - Mass ratio description changed -> secondary over primary part deleted because it was confusing to connect it to appreciated definition (secondary is *always* smaller than the primary). This was suggested by Panos Niarcos. - THE keyword for X-Ray binaries has been added to constraints handler. The entry is now disabled for other models. - More work on PHOEBE Homepage has been done. - Top-level files WISDOM, CAVEATS and QUESTIONS have been dele- ted, because their contents were obsolete. - A nice "under construction" xpm icon added to Orbit tab. :) This is the beginning of introducing small icons to PHOEBE buttons. This will hopefully be even more pleasent for the everyday user. - A lot of CVS repository directory cleaning has been done: the documentation was completely separated from phoebe source di- rectory, icons and pixmaps have been distinguished (icons are only those pixmaps which (will) appear on buttons and pixmaps in general are empty LC/RV/Chi2 plots and logos), all contri- butive files have been removed from the CVS. The data under the CVS is only for UV Leonis (along with basic parameters) and V505 Per (without any keyword file). - PHOEBE-0.13 testing version released. 2002-09-18: - Top-level README file SETTING UP section extended to include instructions for adding PHOEBE-related devices to SuperMongo's graphcap.local. Thanks to Frederic Royer for reminding me! 2002-09-17: - PHOEBE Homepage has been created. It is not completely fini- shed yet, but it is already functional and at present it re- sides on http://www.fiz.uni-lj.si/~prsa/phoebe. 2002-09-15: - Returned from the Monte Rosa GAIA conference. I have met Mr Wilson, Mr Van Hamme and Mr Milone, basically the alpha and omega of the eclipsing binaries world!!! - Fixed an OB1 error because of which mass ratio couldn't be fitted. This bug was fixed with help by Walter Van Hamme. 2002-09-05: - Added "tinygif" plotting device; it is useful for creating small animated gifs. Its size is 256x128 pixels. 2002-09-04: - PHOEBE Scripter window created. A button added to a Fitting tab that initiates the scripter. Scripter routines remain to be written. 2002-09-03: - Function prototypes that set widgets' sensitivity based on a solution constraints implemented. Constraining works for de- tached binaries and overcontact binaries. - IPB switch now also influences the sensitivity of CLA-depen- dent widgets. This tells the user when to expect CLAs to be computed from temperature (widgets are off) and when to be adjusted by user (widgets are on). - RV readout handling fixed. It should now be immune to arbit- rary comments within an .lco file. - Comments in .lco files rerouted to screen as notice windows. There are currently 4 supported notices (critical lobe and critical potential) and only one ever encountered! :) 2002-09-02: - WD98 .lco file obviously has only *more or less* standard sha- pe! Thus in some cases WD inserts a comment in a file (e.g. that the secondary star overfills the critical Roche lobe), which alters all positions of parameters in a file. Therefore PHOEBE crashed when such an unexpected comment was encountered during the readout. The whole concept of reading in values had to be changed. It has been done for LCs and remains to be done for RVs. 2002-09-01: - Cousins R and I filters added to the filter list. 2002-08-30: - A remnant of older versions which limited LCs to 5 fixed; now the top limit is set to 15 (Wilson 1998). - Drastic changes made to the filter list. Descriptive lines introduced for each filter set (Johnson, Stroemgren, Hippar- cos, RV and GAIA). Some wavelengths were catastrophically wrong! Fixed. It might be worth considering expandable list of filters, such as a pop-up menu with submenues instead of a combo box. - Comments still interfered with RV plots. Fixed. - Zillion new stars added to the data directory; they won't be put under the CVS, though, because although great, these data files bring nothing new, they just add to the bandpass amount and will be available for optional download from the PHOEBE homepage. 2002-08-29: - Data aliasing completed for phases (both LCs and RVs). The aliasing switch in both plotting windows initially turned on (there is no obvious reason why the user wouldn't want to use aliasing). - "Update Selection" button in the DC window disabled, because this option won't be implemented for quite some time (at least until much more pressing issues remain). - A Fitting QuickBar window created. Users have expressed the wish that adjustment switches and steps be gathered in an auxilliary window for faster access (so one doesn't have to browse through main PHOEBE tabs). The "Show QuickBar" button has been added to DC window. Callbacks haven't yet been con- nected to the DC window, so the QuickBar isn't functional yet. - Commenting supported in input data files. If a line in a data file contains '#', that line will be *completely* disregar- ded, wherever '#' is placed within that line. This could be improved in future so that PHOEBE disregards only the remain- der of the line, but that would mean too much work right now. 2002-08-28: - Chi2 3D plot wasn't working with orbital eccentricity. Fixed. - Two decimal numbers in Chi2 3D plot limits wasn't enough. It has been increased to 3. It would be useful to add automatic ranging in future, but it is a long-term plan. - PERR0 description changed. PERR0 value follows a strange convention where PERR0 = 0 and PERR0 = pi means that the elipse is oriented along the semi-minor axis. - A new callback added to all Adjust switches which updates the contents of the fitting window if these switches are toggled. - A bug fixed where original temperature values changed the units from K to 10000K. Now they are always in K. - Allocation in DCO readout wasn't correct for curve dependent parameters. This crashed PHOEBE when more than one such pa- rameters was to be adjusted. Fixed. - HJD -> phase transformation was off by 1 if HDJ0 was larger than HJD. Fixed. - Aliasing function implemented. GUI still has to be updated. 2002-08-27: - A new typedef "PHOEBE_dco_record" created. It holds all DCO quantities. It is associated with the "read_in_dco_values" function. - PERR0 units are radians. The range of PERR0 was [0, 100] and was obviously completely wrong. It is now set to [0, 2\pi]. - Chi2 info list created in DC window. It contains a sum of squares for each data curve including only individual weig- hts (supplied by the user). The list has been completely integrated into normal operation. - GAIA filter wavelengths (Munari 1998) added to filter list. - LC/RV plotting parameters' updating is enhanced. There is no need to re-click on the Data switch to reread their values. - Systemic velocity step reduced from 1 (initially meant only for km/s) to 0.01 (to conform with 100km/s too). - PHOEBE documentation directory phoebe/doc put under a sepa- rate CVS. This way the user doesn't have to download the documentation if he doesn't want to. - Van Hamme limb darkening tables (along with an auxilliary PHOEBE file) added to the CVS directory catalogs/ld. This way the user may, if he so desires, download the tables independently from PHOEBE. - Instructions for obtaining PHOEBE from the CVS added to the top-level README file. 2002-08-26: - After 20 days of brain-dead activities and outmost leisure I have returned from a wonderful Montenegrian Sun. I can't say I'm too happy about that but I did relax and now I hope that PHOEBE will grow even smoother! - A lot of useful comments recieved from TOmaz and Paola! Thanks, guys! - VGA entry added to PHOEBE_wl_dependent_parameters structure. - A bug fixed in RV window: the VGA value wasn't updated when experimental data was selected. It now works. - Proximity effect switch (ICOR) split to two switches, each for one RV curve (ICOR1 and ICOR2). GUI updated. 2002-08-06: - Velocity units haven't been used in RV plotting. Fixed. - VUNIT switch moved from the Switches tab to RV plotting win- dow. - LC Plotting window slightly rearranged. The flux factor and magnitude base widgets are overlapping and are shown based on the choice of the dependent variable. Grid and box buttons moved more to the left. Phase normalization switch moved from Switches tab to here. - Switches tab finally deleted. - Well, dear reader, I am happily signing off and going to the Montenegrian seaside! Bye bye! 2002-08-05: - Inclination fitting diverges if the starting value is 90. Thus this value has been changed to 80. - A bug introduced when NPPL switch was deleted that caused gar- bage to be entered to a DCI file. Fixed. - Chi2 calculation for RVs was accidentally left disabled. Now it is back on. - If only experimental data is selected in plotting windows, synthetic curve (LC or RV) was calculated redundantly. Fixed. - Zooming and moving the RV plot caused garbage to be written as data points number. Fixed. - A new typedef enum PHOEBE_plot_device added to phoebe_global. It is passed to all plotting routines as an indicator of the desired output. - Plot to File button in LC plotting window enabled. Plotting devices currently supported are x11 (the screen), encapsulated postscript (eps), X Windows Pixmap (xpm) and columned plain text (ascii). - Prototypes for RV Plot to File functions created, but not connected yet. - A few small glitches smoothened out and unified through all Plot to File related routines. - Graphics Interchange Format (.gif) and Portable Pixmap (.ppm) added to device list. - RV Plot to File fully supported. Prototype functions made for chi2 plot. The Plot to File button has been added to the fit- ting window and callbacks generated. Chi2 plotting routine first has to be rewritten to use SM library. Then and only then Plot to File may be enabled. 2002-08-04: - All DEL-XY delimeters changed to Step, because it seemed WD98 notation didn't help anyone, it just brought more confusion. - Interest combo box finally deleted. Button box was extended to include "Plot LC" button, "Plot RV" button and "Fitting" but- ton. The "Run" button was deleted along with all attached callback functions. - A parameter table in plotting windows was updated even if syn- thetic lightcurve wasn't plotted. Fixed. - A typing error Modelling -> Modeling fixed in both plotting windows. - IPB switch had an old callback dangling which crashed PHOEBE when the switch was turned on. Fixed. - The weighting "Edit Entry" button was enabled by default, which caused a segfault when clicked when no records were pre- sent. Fixed. - The tutorial extended significantly. 2002-08-03: - Column number spin button in the Data tab was obsolete and was deleted. Tutorial updated to reflect this change. - NPPL number spin button in the Data tab was obsolete and was deleted. Tutorial updated to reflect this change. - Tutorial heavily modified; all sorts of updates and improve- ments. 2002-08-02: - A bug caused a fatal PHOEBE termination if assigned data file didn't have any filters assigned. This was a consequence of non-consistent filter names checking for plotting routines. This bug is now fixed. - Another bug that changed multiple filter names has been fixed: if a data entry has been edited, but its filter remained un- changed, then an extra #2 was appended to its name. Fixed. - Fool-proofing added to data file format; if the format doesn't conform to the format defined in Data tab, experimental data output is suppressed. - Fool-proofing added to missing filenames; again, if a file- name is missing (invalid), PHOEBE warns the user about it and suppresses any further data output. - Corrected parameter values from fitting are now updated in plotting windows as well. This eliminates the work-around of turning off and then on the data switch in those windows. - A bug that resetted original values when clicked on Update button in the DC Corrections window, although there were no corrections calculated, fixed. Now nothing happens. - phoebe/data directory updated. 2002-08-01: - PHOEBE 0.12 released. This is the first testing version. - First 18 pages of PHOEBE 0.12 tutorial have been written. 2002-07-31: - Binning checked, some minor type inconsistencies fixed. - Code cleaning even furtherly purified PHOEBE. - Synthetic grid buttons have been temporarily disabled, be- cause of their unfinished state. They will be made available in the next version (or the CVS shortly). 2002-07-30: - After a long long time, RVs are working again without any ma- jor issues. Plotting of *both* RVs at the same time is cur- rently disabled, because it is very dangerous to examine both curves at the same time -> users don't bother to check if the right data belongs to the right synthetic RV curve. This decision will be given more thought soon. - A bug in filter naming (adding a #number to multiple curves in same filters), which added extra #number if already num- bered filter's record (other than the filter itself) was changed, is now fixed. - Level-dependent weighting (NOISE) added to wl_dependent type- def and is read in properly. - Fitting functions GREATLY simplified and working wonderfully. - Mass ratio (RM) spin-button steps were too large. Fixed. - An ambiguity in units of RVs fixed; fitting now *always* uses 100km/s as units, so PHOEBE takes care of the transformation from km/s if necessary. That was the reason why the fitting of semi-major axis failed; now it works without issues. How- ever, mass ratio is still returning nan after fitting. - A FIRST EVER preliminary result reached with PHOEBE: UV Leo. The configuration file may be found in the data directory. 2002-07-29: - A bug in RV dependent variable keyword entries fixed. - A LOT of work rearranging routines has been done. 2002-07-26: - phoebe_global.h created. It will hold all global variables and typedefs from now on. - phoebe_allocations.c and phoebe_allocations.h created. All memory allocations, readouts, struct readins etc. go in here. - All global declarations moved from main.h to phoebe_global.h and main.h erased, Makefile.am updated. - phoebe_gui_typedefs.c and phoebe_gui_typedefs.h erased. - Memory allocation for PHOEBE_data_record types moved to phoebe_allocations.c. Callbacks code cleaned with this. - Default dependent variable in LC plot window changed from Magnitude to Total Flux. - A humongous fixing has begun; all functions now go through a major rework and hopefully, everything after this fix will be ready for the first ever major public release. 2002-07-25: - FOUND THE BUG! Again, readout_str vs. working_str violation. - open_keyword_file () function simplified with another readout function. Implementation is finished, keywords still have to be adopted. - Secondary star temperature (TAVC) had wrong units in DC fit- ting window. Fixed. - DC Update All Corrections button functional. It works for both plain widgets and CLists. - LD interpolation window got hidden when values were updated. This wasn't practical. The window now closes only on the click to the Close button. - Started working on SuperMongo library calls instead of ex- ternal calls to SuperMongo itself. Library bindings work, but we just lost on general portability. Autoutils should check if these libs are available and if not, compile it without SM support. If nothing else, ./configure must fail if these libraries are missing. 2002-07-24: - Seeking a bug that crashes PHOEBE while opening keyword file on some machines (not including mine), but to no avail. Some small irrelevant details changed in affected routines. 2002-07-23: - A nonsense layout of X2A adjustability widget in LD tab fixed. - Marquardt lambda (XLAMDA) widget added. A switches typedef has been extended to include it, also read_in_switches () function reads it in as well. DCI creation routine has been updated accordingly. - Fitting parameters layout in Fitting tab changed. - All "Adjustable" switches renamed to "Adjust" for obvious reasons. Widget names have also been changed from _adjustable to _adjust. - Model combo box contents changed to be more user-friendly. - A warning window, although created, didn't show up on the screen in some circumstances. Fixed. 2002-07-22: - Chi2 calculation button was not fool-proof against undefined filter, which caused PHOEBE to crash. Fixed. - main.c cleared; it now contains only init functions. - An overflow bug in version number string fixed. - MODE combo box moved from Switches to Data tab. Also, Name entry for star name has been created. It has no significant purpose other than easier paperwork. - N1 and N2 separated and moved from Switches tab to Fitting tab. Same has been done for N1L and N2L. - ISYM switch (symetrical derivatives) moved to Fitting tab. - Fool-proofing added to environment readout; PHOEBE scans the environment to find the HOME entry. If it fails, it prompts the user to define it and restart PHOEBE. $HOME is definitely a must-have on Linux, so this doesn't seem too harsh. 2002-07-19: - Some minor glitches of already existent CLists fixed. - Limb Darkening Coefficients for RVs CListed. - Level-dependent weighting CListed. This is the last LC/RV num- ber dependent quantity, so hopefully we're through with chan- ging the interface. Now routines will be optimized for as many LCs and RVs as the user wants (actually, the current LC limitation posed by WD98 is 25) and soon a new release will be made. - Finally, the doubling, tripling, ... of LC and/or RV curve filters is handled correctly. To each filter name there is a #number appended to separate one curve from another. It is done in only one routine, which changes PHOEBE_*_data accor- dingly. This means that all other functions that depend on filters *must* read only PHOEBE_*_data[i].filter names. - LD interpolation routine fixed to reflect the above change. 2002-07-18: - 3rd light CListed. - LC plotting turned off because of extensive changes in main PHOEBE window. - SPECS file enhanced with What Goes Where section and Widgets Naming Convention section. - phoebe_init () function created to lighten the load of main(). - Further code cleaning has been done. - Limb Darkening Coefficients CListed. Interpolation routines fixed accordingly. 2002-07-17: - RV lists incorporated onto their designated places in Data and Luminosities tabs. RV plotting is still turned off at the mo- ment and will remain so until integration is complete. - Keyword entries precision changed from single to double. - Serious code cleaning done. - Defaults main button activated. 2002-07-16: - A small bug fixed: PHOEBE segfaulted when double-clicked on the empty CList record. Checking is now incorporated every- where. 2002-07-15: - Theory of fields PASSED!!! :) 2002-07-09: - Zero magnitude spin button added to Data tab. This is not to be confused with MZERO value of WD98. This is a value of mag- nitudes which is used to transform magnitude-given input fi- les to fluxes. One should set this value (from now on reffer- red to as MNORM) to have the flux in quarter phase after transformation around unity and then adjust luminosities to fit the exact values of normalization. - Synthetic grid changed: previously it was arranged filter after filter, which isn't really practical for reading out the values. This is why it has been changed to sort the grid by filters last, meaning different filters for same parameter values are together. 2002-07-08: - LC luminosities have been CListed and placed from Utilities to Luminosities tab. Same has to be done for RVs. 2002-07-05: - After much effort, CList for LC information has finally re- placed all obsolete widgets. Same has to be done for RVs. - Luminosities CList containers for both LCs and RVs created. Both CLists now reside in Utilities tab and will be moved to Luminosities as soon as possible. 2002-07-04: - A few fixes on observational weighting have been done. - Added new UV Leonis photometric data. 2002-07-03: - An upper limit for LC number set to 25, according to (Wilson 2001, WD98, pg. 17). - Created a toplevel WISDOM file, which contains physical in- sights to PHOEBE related modeling. - Normalization routine estimate_zero_magnitude added to trans- formations. It is based on unweighted average, which is not the best possible choice (median would be better). However, input fluxes to DC are in arbitrary units (they *must* be fluxes and not magnitudes, though), so normalization isn't really important. It is the question of style, really; it is nice to at least tend to normalize magnitude-supplied data to unity at quarter phase. This remains in the TODO. - All normalization procedures from magnitude-based input files now go over estimate_zero_magnitude function. MZERO influen- ces *only* the LC magnitude output. FACTORed LC output is unnecessary and will probably dissapear soon. - Luminosities and limb darkening coefficients all have as many values as there are lightcurves. Therefore DC info list must contain as many entries. Done. 2002-07-02: - phoebe_warning function added to improve error handling. - More work on DC calculation window integration has been done. - Further file-list integration has been done. Old widgets are slowly being deleted. - Luminosity spin button limitations fixed; the upper limit was 4\pi, now it's 100. - If there are no RV curves present, WD98 cannot calculate SMA, so the adjustment button has to be disabled. Fixed. - Temperature adjusting has to have changed units. Fixed. - Memory initialization bug that caused DC computation to crash when run without and RV curves fixed. - String allocation bug in LD typedef fixed. - Fixed a bug in DCI creation function that used the wrong va- lue for YBOL1. 2002-07-01: - DC calculation window created. It contains a list of all pa- rameters selected for adjustment with their original values. After the calculation all other cells are filled with read data from DCO file. 2002-06-30: - Added double-clicking editability of synthetic grid entries. Unfortunately, this demanded that the list widget remains active all the time, because disabling it during the editing also disables "event" signal propagation. The sensitivity is therefore kept on all the time. - Added double-clicking editability of lightcurve data. The same restrictions apply as above. - 2D correlations Calculate button sensitivity fixed. If clic- ked on that button when no lightcurves were defined, the program crashed. It is now fixed. - A DC minimization window created and adjusting parameter list populated. Calculation functions are already finished, all we yet have to do is to add individual experimental curves' sig- mas and write the calculated values into the list. 2002-06-28: - PHOEBE version number and date is now declared in main.c and main.h. Up until now the label in GLADE file had to be chan- ged and the whole package recompiled. - A small header created for grid files. - Grid-related source files phoebe_grid.c and phoebe_grid.h introduced and all relevant functions and callbacks moved there. 2002-06-25: - Keyword file saving/opening now works for both old and new LC declaration widgets. Old parts have been marked for dele- tion. - RV columned list has been added to Utilities window. Every- thing is the same as for LC columned list, it just needs to to be finished. The question is, of course, what is the mea- ning of more than 2 RV curves. - Synthetic grid calculator didn't refresh the list on conse- cutive button-clicks. Fixed. - Editing synthetic LC grid parameters is now possible; there is a special window dedicated to that. - Finally, grid calculation is about finished. Everything else (including defining different grid parameters) will be done in one of next releases. 2002-06-24: - Dialog windows have been destroyed if one would close them by pressing the (x) button on the window decoration. This beha- vior was wrong; the window should only be hidden. Fixed. - The effort to bring more LCs to PHOEBE has paid off during the past few days; with a lot of testing and learning, PHOEBE is finally capable of handling infinite light~ and radial velo- city curves. Currently there is a columned list in Utilities tab and it will remain so until everything works without any issues. Keyword file saving has been modified already to re- flect this change. - Unfortunately, explained development has pushed the date of 0.12 release for another few weeks. I hope it was worth wai- ting for. 2002-06-17: - Grid calculation added to Utilities tab. Window is created and templates have been set. Basic calculation is already availa- ble. 2002-06-14: - All variable names have been cleaned up for 0.12 release. 2002-06-09: - PHOEBE UNDER CVS :pserver:andrej@deneb.ago.uni-lj.si:[path]! - Chi2 plotting function cleaned of directory dependency; now all calculations are performed in PHOEBE_TEMP_DIR. 2002-06-06: - Extensive code cleaning performed. - DEVELOPERS text file renamed to SPECS, since it contains spe- cifications and suggestions for potential developers. - The worst bug ever discovered in PHOEBE: a humongous memory leak was produced by allocating, printing to and freeing string pointers generated by GTK. This is prohibited by GTK and that was the cause of earlier crashes (prior to multi- ple strings work-around). It is now mainly fixed (in all ma­ jor .c sources) and it will be thoroughly checked before the next release. 2002-06-05: - Wavelength-dependent typedef is extended to read out RV data next to already implemented LC readout. This drastically sim- plifies RV plotting routines. 2002-06-03: - An aspectxpm entry added to SuperMongo's graphcap.local. It will eventually make aspectgif obsolete, which is done because of GIF copyright problems. - A small bug that prevented 3D Chi2 initial (empty) plot to have a proper dimension has been fixed. - All image formats changed to XPM instead of GIF. - A new spin button has been put to PHOEBE_plot_lc window: standard deviation. It is easier to "measure" how good the experimental estimate for sigma is. - A complete style unification has been started. At this point no new features will be added until the code is cleaned com- pletely. - Clean-up functionality added; prior to exiting, PHOEBE now deletes all the files it created in PHOEBE_TEMP_DIR directo- ry -> all phoebe_* files. 2002-06-02: - A small bug in file_exists function fixed; it returned errno handler, which is volatile and thus not appropriate for re- turns. The switch is now boolean and we use it like boolean: if (file_exists ("bla")) or if (! file_exists ("bla")). All functions have been updated accordingly (hopefully ;) ). - Finally, believe it or not, A MAJOR FIX: the consecutive fi- lenames for all config/run files have been numbered and no- thing is lost anymore! The fine-tuning will be done as soon as gif->xpm conversion is finished. 2002-05-30: - Error handling for malloc, realloc and fatal events implemen- ted. Memory access is now called through those wrappers; this is not a sound policy, but for as long as memory testing goes, this seems to be best. - Chi2 2D correlation works! Well, kind of; parameters' limits still have to be taken from their spin buttons, a bunch of non-important, stylistic buttons have to be finished, but already a user can calculate and draw SMA vs. XINCL! 2002-05-29: - An ugly bug fixed: when calculating chi2, the values further- most to the right were treated as they were on the left due to inaccurate inequality. It is now fixed. 2002-05-27: - Independent variable switch deleted. It seems that that combo has no purpose, since independent variable is set from the LC Plotting window. The entry has also been deleted from keyword files. - Spin buttons FACTOR and MZERO were moved to LC Plotting win- dow. It just makes more sense for them to be there. - Third light and opacity support finished in LC Plotting. - RV Plotting Window rearranged to reflect the changes in LC Plotting window. It would be worth considering what are the benefits of separate plotting window; is it really necessary? Wouldn't it be more intuitive to have them integrated in the main PHOEBE window? - VGA spin button added to RV Plotting window. It seems like a useful variable in the plotting window itself. - Most of RV plotting routines have been enhanced and modified so to follow the example from LC plotting. 2002-05-24: - Synthetic noise support has been finished; the button has been moved to Luminosities tab (for correspondence with weighting) and keyword files hold the information on the switch state and the seed value. A random number generator for seed has been made. - Standard deviation for synthetic noise background label and spin button have been added; now hopefully synthetic noise makes sense. - Detailed reflection effect transferred to luminosities tab as well; keywords and read_in functions have been updated to reflect that change. - 3rd light support finished: initial values changed from 10.0 to 0.0; wl_dependent read_in function extended to contain 3rd light as well. - Plot LC window rearranged to contain 3rd light and Opacity function spin buttons. They are supported formally, but fil- ter-dependent callbacks still have to be adjusted. 2002-05-23: - A small inconsistency fixed in callback functions: the PHSV/ PCSV calculation routines used the separate function for shading regulation, which already exists generically. - RM and F1 values from main PHOEBE window are now transferred to PHSV/PCSV calculation window prior to calculation. - Status bar now informs the user that PHSV/PCSV calculation has been initiated. - A small bug fixed: if LC number was changed, the change didn't reflect in fitting and plotting combo boxes. Now it does. 2002-05-22: - Added filename parameter in create_dci_file routine. This eliminated the "no write privilidges" warning. - Deleted obsolete function variables for LCI file creation: IFSMV1, IFSMV2, PHNORM, THE, VUNIT, FACTOR, MZERO. They have already been included through typedefs. A new typedef has been introduced for wavelength dependent parameters. - Extensive code cleaning has been made around LC plotting. - Another nasty bug has been eliminated; due to rounding error in WD98 the number of synthetic points doesn't always corres- pond to chosen number; thus renumbering was necessary and now it has been done so. - A Chi2 plotting function has been further implemented; it now has a nice progress bar to ease your waiting. Also, the implementation triggered a heavy code change; it is now even more mature and I hope it now has more or less standard form. 2002-05-21: - Default setting for semi-major axis changed from 1 to 10. - A nasty bug fixed that caused segfault in LC plot: if binning was active and independent variable was set to time and expe- rimental data was turned on, the plot segfaulted due to bin- ning attempt. 2002-05-20: - Started working on DC section; a two-parameter correlation GUI has been created in the Fitting tab, a routine for dcout read- out has been prototyped. 2002-05-17: - Higher Quantum Mechanics PASSED!!! :) 2002-05-07: - LD interpolation button was active even if the LD switch in configuration window was off. It is now fixed. - MMS support finished. It is to be considered if it is impor- tant to save MMS settings to keyword file. Currently it seems unnecessary. 2002-05-06: - Preliminary support for the method of multiple subsets has been introduced to Fitting tab. Clicks and shades appear to work correctly; the DCI creation still doesn't reflect the implementation. 2002-05-05: - About PHOEBE window created and connected to Help menu entry. 2002-04-28: - Digits after a decimal place have been changed for temperature from 5 to 0, since it makes no sense to have temperatures to 5 decimals accurate. - Monochromatic/bolometric buttons have been joint to only one interpolation button; the choice is now reflected in the se- parate interpolation window through a combo box. - All interpolation callbacks have been finished. Interpolation now works for bolometric and monochromatic values. - Log g and metallicity values added to keyword files. 2002­04-27: - LD tables readout has been changed; return value is now a typedefed PHOEBE_data, where .weight is used to carry error handler and .indep and .dep return values for x and y. The routine for bolometric readout is no longer necessary, for the new routine has been introduced that reads out data based on the desired filter, supplied as an argument. - Fixed a rounding bug in metallicity string conversion. - A file ld_availability.data has been created and put to LD shared directory; it contains triplets (T, log g, M) that are available in LD tables and is used for interpolation. - Interpolation routine has been completed and tested. There are potential flaws in out-of-range values, but for now just avoid them; it will be fixed with global error-handling. - Temperature spin-button limits have been changed accordingly: from 2000-20000 to 3500-50000 interval of allowed values. - A new GUI window for interpolation has been created, but only the GUI layout. Callbacks still have to be connected. 2002-04-21: - Error handling for LD table readout is provided through an integer return value: 0 for OK and -1 for failure. 2002-04-05: - Started working on limb darkening coefficients interpolation. Configuration window has been modified to contain LD switch and directory entry. Two new source files have been introdu- ced: phoebe_limb_darkening.c and phoebe_limb_darkening.h. A prototype for the interpolating function has been created. 2002-03-22: - Transformation routines have been enhanced - they now use typedefs instead of static variables. There was also a problem with FACTORing, which is now fixed. - All climb rates for spin buttons have been unified and set to zero. Values change much more smoothly now. - Normalization routine has been implemented for correct nor- malization. Up until now the "normalization" only multiplied the input data by FACTOR, now it actually normalizes data at PHNORM and then multiplies it by FACTOR. Possible problems exist if there are no data points close to PHNORM. 2002-03-20: - Spin-button step for XINCL changed from 0.01 to 0.1. - IPB and IFAT* switches have been moved to Luminosities tab. IPB influences whether CLAs are to be taken into account, so it makes more sense to have it there. Shading for CLAs has been fixed to reflect this switch. IFAT2 switch never existed and now it does. 2002-03-17: - Prototypes for fitting functions have been made. - A bug in DCI file creation has been fixed: due to cfortran.h "feature" that images C's [0,n] to FORTRAN's [1,n+1] auto- matically produced an OB1 error in PHOEBE code. Now it has been fixed and tested. - Due to ANSI C compatibility all GTK+ related typedefs (gint, gchar, ...) have been replaced by original C declarations. As PHOEBE progresses, it would be much better to check the float/double/.../ dependent accuracy from a Makefile.am and not by GTK+ typedefs. - A bug in RV plot due to inconsistency between data VUNIT and synthetic VUNIT has been fixed. It now works for arbitrary synthetic VUNIT values. 2002-03-14: - LC Binning finally fully supported. Any sense in binning RVs? - A serious code-cleaning has been made concerning static and dynamic space allocation. Hopefully everything will work much better now that RAM is dinamically accessed. 2002-03-11: - A bug fixed that wrote PHNORM value as integer instead of double into a keyword file. 2002-03-10: - Switches tab changed; it is eventually expected that this tab will dissapear and its entries distributed to other tabs. 2002-03-08: - Fixed-size arrays changed to dynamically allocated memory blocks. The source hasn't been tested for leaks yet; the main reason for this is to allow PHOEBE to receive as large input data block as limited by RAM itself. Practically this means that all limitations are set by the WD98, not by PHOEBE. 2002-03-05: - Weighting supported in Luminosities tab. Now you may choose between three different weighting options. 2002-03-04: - Adjustable toggle buttons for LD coefficients have been fixed: it is not bolometric values that can be adjusted, but monochromatic ones. 2002-02­28: - PHOEBE-0.11 pre-release version released. - GUI has been changed even further: "Exotic" tab was renamed to "Orbit", since it contains orbital parameters. Some space is reserved for Roche lobe graph and slice diagrams. A new tab has been created, "Luminosities", where all HLA and CLA variables are put. Third light (EL3) support has been added. The order of tabs has been changed to more meaningful order. 2002-02-26: - RV plotting was accidentally turned off in PHOEBE-0.10. It is hanged to more meaningful order. 2002-02-26: - RV plotting was accidentally turned off in PHOEBE-0.10. It is now fixed and it uses FORTRAN routines for LCI creation. 2002-02-25: - GUI has been changed to contain DEL parameters for fitting. 2002-02-22: - PHOEBE-0.10 pre-release version released. This version is in- tended only and explicitly for LCI compatibility testing pur- poses. ---------------------------------------------------------------------------- Prior version of PHOEBE-0.10 was called WDGUI - Wilson-Devinney Graphical User Interface. Since that version was only a model for PHOEBE, this is a ChangeL