97/05/05
Porting the new codas and makefile to the Sun.

1) files used in the make system (such as objects.lst) must end
with a newline.  I changed listobj.prl accordingly and reran
it.

2) The root makefile simply has to be different.  Saved in
make/makefile.unx

3) gcc is not installed correctly at SOEST on Solaris, so we
can't use it.  For cc, the -Wall option does not work.

4) Copied dbhost.sgi to dbhost.sol and linked it to dbhost.h

5) cd2u has to be changed.

6) Permissions got all fouled up.

7) cntfiles is hopeless because of the symbolic links.  It
will just have to stay out of transfers from the PC to the Sun.

8) After changing dbhost.h to a copy rather than a link,
the only links are in cntfiles.


>> when ladcp is here, move hunflag to ladcp system--take it
out of codas makefile, cd2u.

>> removed cntfiles, matlab/matlab4, adcp\demo

csh scripts:
   adcptree: replace with something simpler
   renamepng.src: deleted; obsolete
   do_adcp: obsoleted by newday.prl
   replace: check with June
   chtrange: likewise
   autoplot: save for now
   maketree: version of adcptree for analysis phase (?); save
         for now
   tree_all: removed; identical to adcptree

new directory: bin/scripts; not clear whether to split it into
bat, csh, perl sections.  Copied in bin/pc/*.bat,
adcptree, maketree, autoplot.

used find to make a list of files, edited the list, then used
it to change the permissions to 0644

==smoothr.c -> optpos.c: changed to dynamically allocate memory
for segindex array

fix_temp.c: fixed a nasty bug that could corrupt the depth. 
The number of bytes to read/write for the bottom_track variable
was not initialized. 

ubprint.c: BUG: .ags file is missing first and last fixes;
fixed.

rotnav.c: changed so that when the input is (t u v depth) from lstbtrk,
the output file retains the depth column.