changes

解吸SEED格式的源代码

Changes to rdseed 

version 3.44 :
	
	added support for mini seed output
	added "r" option to output response information
	changed "Printers/*.c" to eliminate null pointer crashes
	merged ddl support
	fixed small bugs in parse_nchar.c & parse_double.c

version 3.45

	fixed bug in process_data.c, where under certain circumstances
	a null pointer resulted in program crash. 

	fixed minor bug, if they specified options:ct, then the times data
	output complained. 

	Also added a printout header to the time series output.

	To do....
		update parse routines to skip if zero bytes need mallocing
		add support for multiple alt_response_files
		add date and time to RESP file output.


	modified and fixed some bugs in the free() routines. Put in 
	checks to make sure we weren't deallocating a NULL pointer, etc.

	added code to clear allocated memory to zeros.

	Added code to use multiple alt_response_file

	Added new way of calculating the A0 for AH output.

	If the output is SAC, added response file output.

	Fixed bug wherein program complained if outputing a mini-seed
	and it encountered a blockette 1000 already.

	Fixed a bug wherein the program didn't put the correct
	time out to mini-seed files.

	Fixed a steim compression bug. Program wasn't reading the
	start of data correctly.

		
	Fixed a off-by-one bug in parse_61 - number_of_numerators	

version 3.46

	fixed mini_flag not being reinitialized after every loop thru
	rdseed. Minor bug.

	fixed small problem in process_data where the program returns if
	the logical record's time was outside any selected time. Just let
	it do its thing.

	**** added seed subset generation.  ****

	expanded the decode string too hold complete name (from 5)

	added frequency multiplication to calib factor in output_css.c	

version 4.0

 	Finished seed subset generation. Generates the same seed version
	as the original, if greater than 2.0

	Added more support for network code. Touched a lot of modules. 
	Added new command line for network code. This will allow for 
	duplicate station names.

	Added support for blocks 43..48 in output_ah.

	Changed process_data so it will look for and use the blockette 1000
	information, if there.

	Added individual station/channel/network selection when the user 
	selects the 's' option.
 

	Untangled some minor spagetti which was causing some missed waveforms.

	Changed AH output so it doesn't do inapproriate calculations if 
	channels are AT, SHO, LOG

	Added network code to output file name of RESP files.

	updated the man page	

	Made changes to the way rdseed reads in the header information. 
	Previously, rdseed would read in all stations/channel regardless of
	a specifications made by the user. It would then eliminate station/
	channels at output time. Now it only reads in the requested station/
	channels. Should save memory, especially on dataless seed volumes 
	being linked to miniseed stuff.

version 4.1

        changed all references to patchlevel.h in include statements to a new file
	(called it version.h).  In addition to the PATCHLEVEL, also defines the
	VERSION for use in print statements

		files effected by this procedure:

		Main/output_resp.c
		Main/rdseed.c
		Printers/pagetop.c
		Decoders/process_time_span.c

        changed the output format of the response file information in the following
	files:

		Main/output_resp.c
		Printers/print_53.c
		Printers/print_54.c
		Printers/print_55.c
		Printers/print_56.c
		Printers/print_57.c
		Printers/print_58.c
		Printers/print_60.c
		Printers/print_61.c
		Printers/print_43.c
		Printers/print_44.c
		Printers/print_45.c
		Printers/print_46.c
		Printers/print_47.c
		Printers/print_48.c
		Printers/print_41.c

	this includes changes that add the comment string 
	(through the creation of the
	Include/resp_defs.h include file) and the blockette/field number 
	addresses for parsing.
        Also added the Transfer Function Type to the output in 
	Printers/print_53.c (so that more
	than just the string generated by rdseed is given as the 
	Transfer Function Type)
	and aligned the dictionary reference integers in [53] and [54] so 
	that they
	look better.

        finally, the output for the print_60, find_41 and find_43 through 
	find_48 files were changed
        considerably.  Breaking comments were added to each type of filter, 
	the blockette info from blockette 60 was expanded, and blockette 60
	had a breaking comment 	added.  A new command line option was also
	added to rdseed (-R) to print this new format.  For the time being,
	both RESP file formats will be supported.  Eventually, only this new 
	format will survive.

	the files that were changed to support this "dual" RESP format are:

                Main/output_resp.c
                Main/rdseed.c
                Printers/print_53.c
                Printers/print_54.c
                Printers/print_55.c
                Printers/print_56.c
                Printers/print_57.c
                Printers/print_58.c
                Printers/print_60.c
                Printers/print_61.c
                Printers/print_43.c
                Printers/print_44.c
                Printers/print_45.c
                Printers/print_46.c
                Printers/print_47.c
                Printers/print_48.c
                Printers/print_41.c

	the changes should be easy to remove, since they involved defining a 
	new flag (outresp_old_flag) to output the old format (defined in rdseed.h
	and rdseed.c), adding an option to the command line interpretation section
	of rdseed.c, and duplicating most of the code in the rest of the files
	to add an equivalent function where the dual functions can be identified
	easily by the presence of the string 'old' in the function name that is
	called to support the old format (e.g. old_print_type43 output_old_resp,
	and print_old_resp).  The R option was also added to the interactive part
	of the rdseed program (and made the default if the user asks to extract 
	responses there).

	tjm; 21 July, 1995

	--------------------------------------

	Added support for "weed" 

	Added the 'S' extract station info option

	Added the 'e' extract event info option

	Found a bug in the output_seed stuff, where the program "forgot" to
	output the block 74s.

	Added a routine to output the LOG records, if there. 

	Added summary file processing to rdseed. Added new input line
	to the interactive mode. Input summary file >>>>

	updated man page
	

version 4.11

	added ability to ignore network codes if there. Farm products don't 
	all have network codes.
	fixed bug in free_43.c and free_44.c 

version 4.11a

	If new style resp file output, no Network Code.  Needed a line added.
	fixed RESP file output so it indents properly - an issue for evalresp.
	added the ability to skip channels and network from interactive cmdline.

version 4.12
	An error reading the tapes was fixed. - rdseed.c
	A0 * gains was added to SAC POLES & ZEROS file output for SAC data
	Some internal date format changes for use with WEED - summary.c
	A minor bug with blockette 1000 and steim 2 - output_mini.c
	Minor bug fix to parse_50.c - make sure network code is a null
		string instead of a null byte.
	Added some code to clean up the logical record memory area, 
		problem crops up if logical record size < 4096 
		read_logical_record.c
	Added change so it processes "I" of family 50 - DDL - parse_key.c &
		parse_30.c 

	Also fixed a bug where it separated the timespan if they crossed a
	day boundary.

	rdseed wasn't positioning itself properly when reading 
	multiple tapes volumes.

	AH output was changed, bug fix wherein it didn't get the
	right instrument abbrviation, also added a unit type to the
	comment field, also added hypocenter info in available.

	output_mini.c had a lot of changes needed to fix orientation
	code missing in the geoscope volumes. Also put check in to 
	make sure the number of blockettes is accurate. 

	Fixed a bug in time_span_out.c where the sample count was
	overflowing the sizeof(int)!

	Also, if they specified multiple time slices, it wasn't 
	processing them all.

	output_sac.c - added conversion of CONSTANT value to what we
	do with AH output data, convert from velocity or accel to 
	displacement. 

	Also changed the SAC poles and zeros file to reflect the 
	3 letter channel name in the case of geoscope.

	minor problem with decode_32bit.c, needed unsigned changed to
	a signed int.

    -------------------------------------------------
    fixed a problem with the way that the 'station' information was
    output.  The 'new' RESP file format needed to be consistent with
    the station information so that the blockette and field number
    prefixes could be removed with a simple 'cut' command.  The files
    effected by this change were

        rdseed/Printers/find_30.c
        rdseed/Printers/find_33.c
        rdseed/Printers/pagetop.c
        rdseed/Printers/print_10.c
        rdseed/Printers/print_11.c
        rdseed/Printers/print_12.c
        rdseed/Printers/print_30.c
        rdseed/Printers/print_31.c
        rdseed/Printers/print_32.c
        rdseed/Printers/print_33.c
        rdseed/Printers/print_34.c
        rdseed/Printers/print_35.c
        rdseed/Printers/print_41.c
        rdseed/Printers/print_43.c
        rdseed/Printers/print_44.c
        rdseed/Printers/print_45.c
        rdseed/Printers/print_46.c
        rdseed/Printers/print_47.c
        rdseed/Printers/print_48.c
        rdseed/Printers/print_50.c
        rdseed/Printers/print_51.c
        rdseed/Printers/print_52.c
        rdseed/Printers/print_59.c
        rdseed/Printers/print_70.c
        rdseed/Printers/print_71.c
        rdseed/Printers/print_72.c
        rdseed/Printers/print_73.c
        rdseed/Printers/print_74.c
        rdseed/Printers/print_abrvd.c

    the result of these changes is that all of the 'station' information
    will be preceded by a blockette and field number prefix, as was the
    case with the response information in the RESP files.  This consistency
    should make it easier to parse the station information from this output,
    since the routines that added a prefix to the response output in versions
    4.1, 4.11, and 4.11a were used in the 'station' information output (so
    the station information output had a mixture of blockette/field number
    prefixes for some of the information and no blockette/field number prefixes
    in other parts of the output).  This fix may make this type of output useable
    with evalresp once more.

    Thomas J. McSweeney ([e-mail]), Feb 20, 1996
    -------------------------------------------------

version 4.14

    changed 'print_type50' so that a network code is always printed.  If it
    is a NULL pointer, the string "??" is printed, otherwise the network code
    is printed.

    bug fix to find_41.c to fix problems with output of RESP files containing
    blockette [41] style FIR filters.  Each was followed by a blank line that
    should have contained a '#' character prefix on it.

    Thomas J. McSweeney ([e-mail]), Apr 29, 1996
    -------------------------------------------------

	Added support for differing logical record sizes within
	the data record.

	Added  support for event processing from weed. rdseed
	will place waveforms for each event as found in the
	summary file into seperate directories. This function is
	activated from the weed interface.

	Alpha port:
	Fixed many problems associated with 64bit compilation. At
	the same time fixed many bugs which showed up with Alpha
	port (byte swapped data).

	Added code to place event/station distance, both degrees and
	Km, azimuth and back azimuth into SAC headers. 

	Version support, looks to see the version of weed which is 
	running it.

	added sip and azimuth in AH output.

	Opps, dip and azimuth needed switching in AH header output.

	added dip and azimuth reversal in output_ah.c

	added event lat/lon information in AH headers.

	Changed some longs to ints in steim.c (64bit port)

	Added support for data blockette swapping if needed. 
	Currently only blockette 100 is located and swapped.

	Small bug fixed in output_css.c where if no description
	entered for instrument type, program would crash.

	Added support for negative BIAS in summary file processing.	

	Took out call to timelocal() and converted to mktime()

	Bug fix in summary file processing where it goofed about 
	what data to grab.

	Changed the output format for event files into a more readable 
	format, as per Ken suggestion. The precision of the floats
	were changed.

	changed decode_geoscope.c so i is insensitive to case when
	scanning for data type code.