changes

解吸SEED格式的源代码

	added scan_for_blk_1000() in process_data.c so it would get
	the word order from this if available.	

	Changed process_data.c to do the byte swapping upfront, instead
	of peppered throughout the code.

	The LOG record output wasn't being done right for "Seed" 
	output. WOuld miss the log records.

	Small bug fix for seed output, blockette 74s seek locations
	weren't being computed properly on subsequent interations.

	Took out some longs and other problem spots for Alpha port.

	This version was never releases, went to 4.15 directly from 4.12

	Added SAC ascii output

-----------

version 4.15

	Added SEEDTOLERANCE - or rather activated it.

	Bug in output_polesNzeros() where if LOG channel it 
	crashed.

	Changed the way rdseed reads stn/chn information, more
	flexible, not dependent on order of info in seed 
	volume.

	Added ability to process data records of varying sizes.

	needed to look at data record size from the blk1000

	added command line processing for summary file.

	bug in output_event_info, date conversion on leap years

	bug in steim.c when day is 366

	changed some things so it compiles and runs on linux.

	Conflict with rdseed.h's abs and stdlib.h's. Made into
	all CAPS.

	Small bug with using one letter network code ala geoscope

	Activated blk 11 and 12 usage. 

	Added SEGY output.

	Bug fix with network codes of one letter, ala geoscope

	implemented blockette 11 and 12 usage, if disk file

	Added alert message processing, if reversals are initiated,
	rdseed will write message to disk and alert the user about this
	file.

	
	Small bug caught in scanning for blockette 1000

	Added to "WEED" station file, station effective dates.

	Small bug in date conversion routine when leap year.

	removed some unnecessary routines from the Snoop directory

	Small bug in output_seed() where if network code is length of one,
	the resulting blockettes needed padding to 2.

	In output_seed(), needed to get the network code from the data record.

	summary.c - small bug in how the channel names were being interfaced to
	the old style holding array (from keyboard).

	ah_resp.c - small bug in the complex math routines.

	steim.c - bug in time computation when leap year.

	Added event info (if there) to output seed volume

	Added stdlib.h for atof()

	solaris compatability

	revision 5 - took out getwd() for getcwd()
		   - also bug fix, fread didn't work on tape drives

	rev 6 - Bug in naming of output files for geoscope data
		bug in getcwd() in log_errors() resulted in 
			I/O error - result too large
		Bug in the 't' option - lrecord_ptr off by one
		
	
	rev 7 - added event hypo info for SAC headers
	      - put some sanity checking for fsdh->bofb field
	      - one last bug fix for finding chn recs when stn
		  recs don't match based one effective time.
		  Default to what is there
	      - added blockette 51, 59 output 'C' option
		   - new command line options for 'C'
	      - bug fix in convert date of event if leap year
		   in output_event_info.c


	rev 8 - changed how the summary file is processed.
	      - took out the old_resp flag option
	      - added a perror() to output_segy.c
	      - added some I/O error messages to output_sac.c
	      - same with output_resp.c output_ah.c and output_css.c
	      - output_css.c : changed elevation from meters to km
		               also fix a calib of zero bug (if no
			       stage 0 (58))


	rev 9 - bug fix for really small (<32k) seed volumes.
	      - bug fixes for summary file processing.	
	      - small apparently harmless overflows fixed (process_data.c)
	      - some needed byteswapping for updated version

version 4.16

	      - output of sub-seed from Geoscope files had 0 in network code
	      - also, mod made for reading small <32k seed volumes
	      - took out innocous warning mess on 'c' option, could be dataless

	      - added station/channel/net prompts to option 'C', the
		station/channel comments option. Also user can specify
		time windows for effective time processing.
	      - bug in output_mini() where it didn't properly write records
		with an existing blk1000 and the rec length !- 4096

	      - bug in output_seed, changed timeadd() to timeadd_double().
		Also in output of blk 55/56, missing % in sprintf()
	      - some errors which showed up in linux compilations.
	      - added in location code processing

	4.16.1 - added toms fix to the IIR instrument error message,
		 poles > 0 in parse_53.c

	4.16.17 - (rdseed 2000)

                prelimenary support for opaque blockette 2000 data type

                added support for decoding 16 bit ints

                added support for multi blk 53s in calc_AH

                another coding of summary file processing.

                bug fixes to print_51.c - changed to old style param
                             get_date.c - same
                             timecnvt.c - paring of day number was wrong
                             output_seed.c  =^?- blk 74 needed loc id
                             parse_53.c - no complaints if poles error > 0

	4.16.2 - small fix to bug that seems to have come back in 
		 extract_time_span() where if the data record size 
		 was < logrec size, it didn't properly look at the sub index
		 value as found in the block 74. Off by one
	       - also add in if num_itms < 0 do perror("") in
		 extract_this_time_span().

	4.16.3 - blk 1001 virtual support, program complained if running
		 under linux
	       - location code processing added to the 't' option

	4.16.8 - added in support for block 42/62.

		 printing blocks 51/59 needed a flag to print the station
			 information.

		 added support for CSS's affiliate, network and
			hypocenter table.

		 bug fix in output_mini.c, pointer to new block needed
			recalculating.

		 sac ascii output needed the numbers reformatted from
			%15.7f to %15.7g, to conform with the SAC manual.
			Maintains a 80 column output with large numbers

		 output_seed.c - changed fopen("a+") to fopen("a"), 
			caused problems with linux, but should be this
			way anyway. Also "w+" to "w".

			Polynomial response was added in.

			get_date() returned pointer to local variable.

			bug in rdseed.c - process_stn_lrecs(), buffer needed
				memset() to spaces in case read failed from 
				EOF.

			summary.c, many changes associated with location codes.
				Also added file version checking. 

			
	4.17    - new summary file processing
		- support for opaque data type
		- added capability for processing multi 53 poles_zeros blks
			when calculating A0
		- took out erroneous message about found a pole > 0
		- added support for 16 bit ints (H2O data)

	4.17.1  - changed prompts to allow for poles and zeros output
		  without asking for 'd' sac output option
		- added in the 'p' option
		- bug fix: poles and zeros file name needed added loc
	    1.1	- type 40s not processing correctly in get_AO - ah_resp.c
		  get_A0() needed error message for blk58 to say stage zero

	4.18  	- t option needed fix for rec < 4k
	 	- added in atc.c but not implemented
		- output_sac.c changed to do the sac_ascii thing
		- small change to output_mini.c to allow for more graceful 
			exit when encounter an unkown code (code_to_english())

		- sac poles and zeros file needed location code added to name
		- incorporated Doug Neuhausers changes (Feb 2001) for
			512 byte data and also extremely low sample rates

			Decoders/process_data.c
				Convert duration from int to double, 
				this_sample_rate from float to double, 
				and time addition computations from integer to double.

			Include/rdseed.h
				Convert last_sample_rate from float to double.

			Main/steim.c
				Added roundoff to time computations to reduce double->integer
				truncation errors.

				Added function my_anint(), my implementation of anint(),
					which is not in the standard Unix math libraries,
					and avoid having to use sunmath library.

				Fixed problem with year rollover in date.
				Handle leapyears properly.

			Main/output_sac.c
				Added asciiflag argument.
				Reintegrated ascii output output_sac_ascii.c into output_sac.c.
				Ascii output now includes event info, network, location code.
	
				NOTE: The file_output_sac_ascii.c is NO LONGER USED.

			Main/output_data.c
				Changed to call output_sac() with asciiflag argument for both
				sac binary and ascii output.	

			Parsers/parse_100.c
				Updated argument for parse_type100.c from *float to *double.

			Utilities/timedif.c
				Change time difference computation from integer to double.
				Not sure this is required, but seemed to be a good idea.

	4.18.1

		decode_16_bit.c - needed to take out debug output file
	4.18.2
		
		added in Jean-Francois Fels' fixes to decode_steim.c for
		linux

		added in Chad Trabant's addins for alt response file
			processing using command line option (g).

		added in command line processing to allow for designation
			of file output directory (-q).

		added in command line processing for seismogram data input
			buffer size specification (-b).

		small bug fix needed when fsdh has zero samples.

		added in Doug Neuhauser's fixes for default lat/lon output
			in the station file (-S) option to have + for
			positive lats.

version 4.19:

		added in the DRQ in interactive and command line prompts.
		Added in DR or Q into file output names.

version 4.19.1  took out Jean-Francois Fels' fixes for decode_steim.c

version 4.19.4  bug in CSS output for origin event time. 
		Function get_secs_since_1970() did not work. Took it out.

version 4.2 	add in more support for DRQ
		bug fix in output_seed_data_file. needed support for multi
		data rec lengths. decode_usnsn had an ugly bug.

		MAC makefile changes. Some fixes for linux.


version 4.3     added in some robustness in process_data.c because
		bad data was creeping into seed volumes.

		ah_resp.c wasn't processing dictionary blockettes
		properly.

		linux bug in ah_resp.c for pointer math.

		ALT_RESPONSE_FILE processing wasn't right for linux

		rdseed.c:
			changed reads to  fread

			needed num_bytes_read returned from skip_to_DRQ

			changed checking for LRECL only if scln changed

		alt_response.c

			changed file numbers to FILE

		process_data.c
			checked for sample rate sanity
			
			miniseed blockettes processing had a bug.
				check for good bod field.

		decode_steim2.c
	
			still had LRECL hardcoded as 4096 instead of
				variable.

		timeqc.c, timetol.c:

			if the blk52 has a zero for a clock tolerance,
				substitute a 1 so math will work.

		VERSION.h updated to version 4.3 so page_top.c would
			print out properly.

version 4.4 log_errors.c for newer gcc
		also, bug in ah_resp.c which only showed up in
		ah_resp.c

version 4.5 fix for GEOSCOPE data, changed makefile to use gcc, took out some
		references to version for easier maintenance    

.....

version 4.6:

Work around for CN which has subarray=0.

Changed reference [ix + ] and parens to work in ternary expression in gcc on linux.

Added 40's logic (exercised by CN data)

Change to calculation of timetol, using sample rate, thanks to Chris.

Set sac idep (units).

Set sac evdp *1000.

Made sac poles and zeros files have a unique (based on channel start time) name.

Fixed calculation of epoch time in timetol - caused a leap year boundary problem.

SRS