The current handler accepts the format as described in EDOC417.
There exists offline tape and disc versions, and online (fast ethernet) sort spy.
The tape written online uses the IN2P3 format. It should only be necessary to investigate this format if tape reading problems arise. See Ganil web site for further information if necessary.
It requires a daemon running as root. The data blocks are then shared with the sort program via a memory-mapped file.
/MIDAS/bin_SunOS/exogam-spy device event-builder-host tape-server-host nstreams
for example ...
/MIDAS/bin_SunOS/exogam-spy hme1 10.0.0.94 10.0.0.173 1
#Format: group... #Clover[0:15] (ea,eb,ec,ed,ta,tb,tc,td) group 0 0x65 0x66 0x67 0x68 0x165 0x166 0x167 0x168 group 1 0x69 0x6a 0x6b 0x6c 0x169 0x16a 0x16b 0x16c #posn[20:35] (pe0,pe1,pe2,pe3,pe4,pe5,pe6,pe7,pe8) group 20 0x75 0x76 0x77 0x78 0x79 0x7a 0x7b 0x7c 0x7d group 21 0x085 0x185 0x285 0x385 0x485 0x585 0x685 0x785 0x885 #posn[40:55] (pt0,pt1,pt2,pt3,pt4,pt5,pt6,pt7,pt8) group 40 0x175 0x176 0x177 0x178 0x179 0x17a 0x17b 0x17c 0x17d
There are a small number of bad events reported by the online sort program.
These are caused by duplicated ADC parameter addresses within an
event, and seem to be correlated with hardware event number
re-synchronisation in VXI.
These events are suppressed before writing to tape, and so do not appear in the
offline sorts.
The runfile details to be provided for the input handler for tape sorts should be of the form ...
TAPE1 RUN3 TAPE1 RUN4 TAPE2 RUN5 etc.The first word on each line (tape name field) is only checked to see if it changes, to force the program to unload the current tape, and allow the user to load the next tape.