Changes between Initial Version and Version 1 of v7.5

01/16/12 08:28:17 (10 years ago)



  • v7.5

    v1 v1  
     1=Earthworm Release Notes V7.5= 
     2August 18, 2011 
     4==NEW MODULES== 
     7This is not a module, but an independant binary used by hyp71_mgr. 
     8It is a slightly modified version by Alexandre Nercessian of the widely used USGS location program. 
     9Modifications allow higher number of stations, altitudes to be taken into account, 
     10earthquakes above sea level as well as stations below sea level. 
     11Jean-Marie SAUREL 07/15/2011 
     14Written by Ruben Luis of CVARG. Produces a fancy email with HTML using 
     15embedded images and a google map. Is triggered by a TYPE_HYP2000ARC 
     16message. This is in the reporting section of the source 
     17Ruben Luis 07/07/2011 
     25Written by Jim Luetgert for taking netquakes mseed data and injecting it 
     26into a ring, winston wave server, gif files and peak ground motion calcuations. 
     27Moved from Contrib/Menlo to data_sources in EW7.5 
     28Moved by Paul Friberg 07/07/2011 
     31This is a module to do for NonLinLoc what hyp2000_mgr does for hypoinverse. It 
     32basically runs NonLinLoc and converts the input and output so that it matches 
     33what hyp2000 does so that downstream programs can get a new earthquake solution. 
     34This uses Anthony Lomax's NonLinLoc locator program found here: 
     36The nll_mgr was contributed by INGV but written by Anthony. Formerly this was in 
     37the Contrib bundle, but it is in use in a few places as a production module. 
     38NOTE that we are also including NonLinLoc with each EW release as a convenience, 
     39but if you need the latest, go to Anthony's web site link above. 
     40Loaded into SVN by Paul Friberg 05/15/2011 
     43A Java 1.6 GUI that receives HYP2000ARC messages from import and creates alert 
     44emails and SMS (sms using clickatel interface) with information about the 
     45event. In addition to the origin info, it sends emails with a google map of 
     46the location and a .png display of the waveforms that had picks. 
     47Written by Ruben Luis 
     48Loaded into SVN by Paul Friberg 04/18/2011 
     51This module sends regularly trig2k messages in order for a listening trig2disk 
     52to produce continuous data archiving. It can replace the contrecord Perl and modules 
     53of GlowWorm package (used until v6.3). 
     54Based on carlsubtrig module. 
     55Tested under Linux CentOS. 
     56Jean-Marie Saurel 03/12/2010 
     59This module converts trig2k messages into hyp2000_arc messages. 
     60Only the stations that has been trigged are send in the output. 
     61P arrival time is the trigger time for each station, and the weight is 4. 
     62Tested under Linux CentOS. 
     63Jean-Marie Saurel 03/12/2010 
     66This module is contributed from Richard Luckette and Doug Scott of the BGS 
     67and it takes an HYPO2000ARC message and converts it to SEISAN event format. 
     68Paul Friberg 12/07/2010 
     71This module allows to use Hypo71 at the end of the megamodule instead of hyp2000. 
     72The input/output is the same as hyp2000_mgr module. 
     73It allors also to compute and integrate local magnitude using standalone mode 
     74of localmag. 
     75Tested under several Linux flavours. 
     76Jean-Marie Saurel 10/12/2010 
     79This module takes data from one or more waveservers, computes and processes 
     80their spectra; results can be written either to a file or a ring.  Processing 
     81includes tapering, smoothing, filtering, and peak identification.  Data 
     82acqusition can be initiated from the config file or from a message read from a 
     83specified ring (see 'compute_spectra'). 
     84Scott Hunter 04/26/2011 
     87This module posts a message that, when read by ewspectra, begins processing data 
     88starting at the specified date/time.  (Deprecated w/ the creation of 
     90Scott Hunter 04/26/2011 
     93This module is a sniffwave for spectra messages, as written by ewspectra.  Results 
     94can be directed to stdout or to a different file for each message "sniffed", in 
     95a specified directory. 
     96Scott Hunter 04/26/2011 
     99This module monitors a ring for TRACEBUF2X messages and generates an alarm message 
     100when a SCNL-specific threshold is exceeded. 
     101Scott Hunter 04/26/2011 
     104This module puts an ACTIVATE_MODULE message in the specified ring, the body of 
     105which starts with the module ID of the specified target module, followed by the 
     106remaining command-line arguments. 
     107Scott Hunter 05/05/2011 
     110This module, upon reading an ACTIVATE_MODULE message in the specified ring, 
     111executes a sequence of scripts (making substitutions w/ supplied arguments), 
     112and reports the final return code in a message to the specified ring. 
     113Scott Hunter 05/18/2011 
     116This module monitors a ring for ACTIVATE_MODULE or alarm messages; generates email 
     117and posts a message to a ring for an instance of activated_scripts. 
     118Scott Hunter 05/20/2011 
     121This module monitors a ring for TRACEBUF2 messages, assumed to be from 
     122sensors on different floors of a building.  The sum of the forces for the floors 
     123is computed for each sample time and, if it exceeds a specified threshold, an 
     124alarm message is written to an output ring. 
     125Scott Hunter 06/08/2011 
     128This module is a multi-threaded version of sendfileII that runs as an Earthworm module, 
     129meaning it sends heartbeats to statmgr, and restarts like other well-behaved EW modules. 
     130Running sendfilemt is equivalent to running multiple copies of sendfileII, but 
     131only one only one configuration file is needed, and sendfilemt needs to be mentioned only 
     132once in the startstop.d and statmgr.d files.  The module was tested on Solaris and Windows, 
     133Since sendfilemt uses the standard multithread Earthworm functions, it will probably run 
     134ok on other operating systems too, if the appropriate makefiles are written. 
     135Will Kohler 7/24/2011 
     138Adsendxs is a rewrite the old adsend module that works with newer model DAQ hardware and 
     139GPS receivers. Here are some features of adsendxs: 
     140- Runs on National Instruments X-series DAQ modules, which come in PCI Express and USB 
     141  versions. 
     142- Uses the NI-DAQmx API, rather than the legacy NI-DAQ API. 
     143- Not tested on Linux.  The NI drivers may or may not be Linux compatible. 
     144- Obtains accurate time from an inexpensive Trimble ThunderBolt E GPS Disciplined Clock. 
     145- Seismic amplitudes are scaled the same as the older adsend program, to ensure 
     146  compatibility with existing downstream software. 
     147- Runs at a user-configurable sample rate.  Tested at sample rates up to 2000 sps. 
     148- Periodically updates the O/S clock, using time from Trimble GPS. 
     149Will Kohler 7/24/2011 
     155Added in CoincidentStaTriggers option to squash triggers of telemetry spikes. This was 
     156done for the Azores network but could be applicable elsewhere. 
     157Note that I also added a version number to the program which is stored in the log file 
     158AND also echoed to the command line if no args are provided. All EW modules should do this. 
     159Paul Friberg July 19, 2011 (with testing help from Ruben Luis). 
     160On August 18, 2011, I added IgnoreCoincident to ignore the coincident check if more than a 
     161specified number of subnets triggered coincidentally. The idea being that a telemetry glitch 
     162would not cross more than this number of subnets. 
     164===hypoinverse 1.3 upgrade=== 
     165Note that the docs have changed and one command (ZTR has 2 args 
     166instead of one). This new version is really 1.35 and allows hypoellipse 
     167features which have now been integrated in. 
     168Fred Klein did all the code, Paul Friberg integrated with EW and checked it in. 
     169July 18, 2011 
     172Added a new option no_S_on_Z which is a simple flag to 
     173not allow S phases to be associated with picks on vertical 
     174component channels. This option is off by default. 
     175Thanks to Barry Hirshorn for the idea from binder_agl. 
     176Paul Friberg 07/04/2011 
     179Now accepts multiple pids for restart instead of just one! 
     180I made the configfile option a flag (-c) too. 
     181Usage: restart [-c config_file]  [ ...] 
     182Paul Friberg 06/28/2011 
     185Fixed the -e option, which apparently never worked! 
     186Paul Friberg 06/10/2011 
     189Stefan and I fixed the windows compilation which produced bogus numbers and never got 
     190a magnitude (despite running fine). The issue was a output array was filled with 
     191garbage numbers (not zeroed). While this was easy to solve (zero the array 
     192before calling convertWave()), it was not easy to find! 
     193Also, later we added the new configuration option ChannelNumberMap 
     194to map numbered channels to letter code orientations. 
     195New version is 2.1.4 (DOES NOT EFFECT ANY UNIX build of this) 
     196Paul Friberg & Stefan Lisowski 06/08/2011 
     197Later on added in capability to  for Z channel mags, via optional allowVerticals 
     198New version after allowVerticals is 2.1.5 
     199Paul Friberg 06/30/2011 
     202added in a -n option to allow  NO FLUSHING of messages if you want to see all 
     203the messages in a ring at a given instant you can use this new feature. 
     204Paul Friberg 06/01/2011 
     207Fixed a bug where no heartbeats were issued if the Directory Scanning mode was 
     208used to playback data. 
     209Paul Friberg 04/29/2011 
     212modified main tport_getmsg loop such that messages received with return codes 
     213GET_OK, GET_MISS, and GET_NOTRACK are all processed.  Previously, only msgs 
     214returned with GET_OK were processed. Msgs with GET_MISS and GET_NOTRACK returns 
     215were essentially ignored; only the return code was logged. 
     216Lynn Dietz 04/01/2011 
     219updated nmxptool to version 2.0.0 
     220Matteo Quintiliani 01/24/2011 
     223added in tracebuf2.1 detection capability display (to show a conversion factor instead 
     224of data quality flags). This is fully backward compatible with all prior 
     226Kevin Frechette 01/12/2011 
     229Fixed a case where if multiple wave_serverV's were listed, only the first one 
     230was ever polled for data (the rest were marked as duplicates). 
     231Paul Friberg 01/12/2011 
     234any linux module using PTHREAD_STACK set to 8192 can possibly segfault on 
     23564-bit linux (unreliably). The thread stack size was increased to 81920 
     236for Linux #ifdefined statement and all problems went away. This was not 
     237tested on Solaris or Windows or Mac OS X. 
     238Paul Friberg 12/02/2010 
     241tankplayer tools, wave_probe, dumpwave, cleandir, glass, startstopservice 
     242On Windows, added kom.obj to link so that the new kom dependency to logit 
     243is taken into account. 
     244Stefan Lisowski 11/11/2010 
     246tankplayer tools, ew2mseed, dumpwave, getmenu 
     247On Solaris and Unix makefiles, added kom.o to link so that the new kom 
     248dependency to logit is taken into account. 
     249Matteo Quintiliani 11/4/2010 
     252Added parameter EnableLongFileName in the configuration file of menlo_report. 
     253It enables writing arc/sum files with format DATETIME_ID_VER.SUFFIX 
     254(i.e. 201011040932_0000097658_2.arc) 
     255Matteo Quintiliani 11/4/2010 
     258Increased buffer size to 512 bytes for reading a line from the station file. 
     259Same behaviour of pick_ew. 
     260Matteo Quintiliani 8/25/2010 
     263Fixed a misleading print statement when WaveMsg2MakeLocal() failed on a SCNL, it 
     264showed possible spaces in the SCNL that were not really there. 
     265Paul Friberg 8/19/2010 
     268Added QueueSize optional parameter to increase the default from 100 
     269Added SleepMilliSeconds optional parameter to decrease the default from 500ms 
     270Paul Friberg 8/18/2010 
     273*Bugfix You can now comment out the line that includes an email addres 
     274        so nobody gets emailed by statmgr, and you don't need to put a 
     275        fake email address in to get statmgr to start. This is the 
     276        behavior that is desccribed in the config file. Fix by Paul Friberg 
     277        201006 
     280- Removed hundredths of second from values of latency. 
     281- Changed the SCNL line output format in order to increase the readability. 
     282  Now each field falls nearly at same column. 
     283- Quality fields are now displayed in hexadecimal. 
     284  Before non-ASCII characters were improperly printed to the terminal. 
     285- Distinguished output for gaps and overlaps. 
     286Matteo Quintiliani 6/22/2010 
     289Removed -D_SPARC from CFLAGS line of makefile.sol. This flag prevented 
     290proper byte-swapping on Solaris for little-endian (Intel) platform. 
     291Ilya Dricker 5/05/2010 
     294*Bugfix to prevent overflow in printing of location code 
     295*Fixed a condition of a bogus 'FULLY in THE GAP' return from waverserver. 
     296 Now if waverserver insists that the request is fully in the gap, but the 
     297 requested startTime is more than 12 hours AFTER the current time, we redefine 
     298 the waveServer FLAG is 'REQUEST-AFTER-THE-TANK' 
     299Ilya Dricker 5/19/2010 
     302Postproc external script is now launched with tow arguments : 
     303date (YYYYMMDD) and time (HHMM) of the processed event. 
     304Change by Frederic Randriamora. 
     305Jean-Marie Saurel 27/12/2010 
     308Added ability to recognize ACTIVATE_MODULE messages to initiate processing; also 
     309added ability to override the set of spectral periods to be processed. 
     310Scott Hunter 05/18/2011 
     313Fixed writing of empty error messages 
     314Scott Hunter 5/17/2011 
     317Restored ability of unix versions to control modules by id; made Windows version 
     318of this more robust 
     319Scott Hunter 6/22/2011 
     322Now allows for up to 10 mailservers; when mail is sent, each is tried in order 
     323until one works. 
     324Scott Hunter 6/29/2011 
     327Uses new transport call to clear ring flag requesting termination of a process 
     328Scott Hunter 6/29/2011 
     331Changed to reflect changes to transport library 
     332Scott Hunter 7/8/2011 
     335I have made very significant changes to the first version. Here is a short list of the main: 
     3361.       Windowed or full-screen - This allows larger traces 
     3372.       Multiple and prioritized waveform sources - One may have multiple waveform sources. The app will automatically select one according to the required station and time interval 
     3383.       List of available stations and time intervals for each source - This allows checking if a waveform source carries the data that we need. Similar to get menu but extended to other waveform sources. 
     3394.       Reads data from wave/Winston servers, tank files (SCNL only), and mini-seed files - I included the latter to be able to check data from continuous archives. 
     3405.       On-screen indication of available data for each station within requested time interval - There is a small check after each station to let the user know that there is data available for the requested time interval. 
     3416.       Facilitated automatic setting of the request time interval based on availability of the sources - I included a button to automatically set the start and end times according to the time interval available from the sources. In case of waveservers it sets the end time to NOW-1 minute (I think) and NOW-21 minutes. 
     3427.       Indication of incorrect date/time settings - Characters turn red if you make an obvious mistake or swap start and end times 
     3438.       Zoom-in and pan buttons as well as drag-to-zoom - Swarm like waveform browsing is really cool to find details. 
     3449.       Missing parameters from pickew are included: NoCoda, RestartLength and MaxGap - I had some requests for this, particularly the nocoda 
     34510.   Tooltips with the pick and coda texts  
     34611.   Tooltips with the causes for failed picks - Same as before. 
     3472011-07-06 Ruben S. Luis (commited to svn by Stefan) 
     351Environment Variable expansion in .d files 
     352Added file environment/earthworm_commonvars.d for declaring "Earthworm Common Variables" 
     353It is now possible to declare variables like ${...} within .d configuration files 
     354Matteo Quintiliani 11/4/2010 
     357==KNOWN BUGS== 
     362==LIBRARY CHANGES== 
     364Added a tport_flush call to the transport library; acts as if tport_getmsg were 
     365called repeatedly until no more messages were found, but doesn't incur the 
     366overhead of returning the messages (which would get ignored anyway). 
     367Scott Hunter 5/18/2011 
     369Changed default mail program used by SendMail for MacOSX 
     370Scott Hunter 5/20/2011 
     372Added a tport_detachFromFlag call to support change in flag: instead of 
     373a single value, it now records all individual process IDs put to it 
     374(until TERMINATE is put to it, which effectively replaces the list) -- 
     375this prevents the flag from being changed before the target process 
     376reads it (which is now possible, due to making the stop & restart 
     377commands multi-threaded).  tport_getflag now yields TERMINATE if that 
     378has been put; if it hasn't but the process's own ID is in the list, that 
     379is returned; otherwise, 0 is returned.  tport_detachFromFlag is now used 
     380to remove a value from the list (what used to be done by putting a 0). 
     381Scott Hunter 6/29/2011 
     383In order to support legacy modules not using the new transport library, 
     384SHM_HEAD was restored to its original form, and the new list-of-IDs flag 
     385is kept in a distinct ring which is only used for this purpose.  Processes 
     386that do not use the new library will be communicated with via the old flag 
     387field (which has the original overwriting issue, but now contention is only 
     388between legacy modules). 
     389Scott Hunter 7/8/2011 
     391Modified transport library so if a module using the new library is run with 
     392a startstop built against the transport library using the original flag 
     393representation, module will use the old style flag.  Also modified to allow 
     394key for flag ring to be specified in earthworm.d (named FLAG_RING). 
     395Scott Hunter 7/19/2011 
     397Modified getutils to provide a version of GetKey (GetKeyWithDefault) which 
     398takes a default value to return if the specified key does not exist; modified  
     399transport to use this new function to avoid error messages when FLAG_RING is 
     400not defined (and a default value is used). 
     401Scott Hunter 7/25/2011