This was initially a redhat 7.2 (beta; codename roswell) install. it seemed to work
just as well as rh71, except I got so disgusted with a number of broken features, I was
not going to wait for the real 7.2 and reinstall 7.1, and thus decided to do this
install again. After all, it's only 4-5 hours on a P600....

The install takes about 1.7 GB, although I didn't keep good track what the maximum
amount was at any time. 2GB is probably a good aim (as long as you don't do the
docs)


- downgraded the compiler from the rh71 default gcc-2.96-.... to gcc-2.95.3

$mkdir -p /home/aips2/code
$cd /home/aips2/code

$ncftp aips2.nrao.edu
  cd /pub/code
  mget README VERSION configure install-16.000.tar.gz 
  quit


$chmod +x configure
$./configure --source

   y		(on working on the install-16 file)
   y		(answer some obvious questions)

***   linux_egcs	(there don't seem to be more options, like linux_gnu ??)

   linux_gnu		-- was not accepted
   linux		   was accepted......

			   NO KIDDING, this is for freakin' rh5, why is this not
			   warned for..... so, i have to go back to stupid
			   linux_egcs.... back to step1 ...


$edit /home/new_aips++/linux/umd/aipsrc

		(set silly things about printer and papersize...)
		sorry, this is stupid, reasonable defaults should 
		 be figured out here....

$edit /home/new_aips++/linux/umd/makedefs

     ok, lapack and blas are NOT on powertools anymore, now on regular.
     as well as lesstiff:
	      
	      rpm -i	      lesstif-0.92.32-4.i386.rpm
	      rpm -i	      lesstif-devel-0.92.32-4.i386.rpm
	      rpm -i	      lapack-3.0-11.i386.rpm
	      rpm -i	      blas-3.0-11.i386.rpm
 
     now, as for pgplot. the old pgplot from redhat6.1 that NRAO
     distributes, cannot be found. it doesn't list where it is.
     I had an installation in /usr/local/pgplot anyways, so I will
     keep that.

     PGPLOT has a quirk. Default pgplot installation from the caltech
     distrbution places all pgplot things in one directory ($PGPLOT_DIR)
     and thus

	PGPLOTROOT := /usr/local
	PGPLOTDEFS :=
	PGPLOTINCD  = $(PGPLOTROOT)/include
	PGPLOTLIBD  = $(PGPLOTROOT)/lib

     from the template will then not work. Need to get rid of include and lib

     also do XmPGPLOT in same way

ok, so i finish this and the screen scrolls by many pages with no
way to get back... i wish i had just  set scrolling to 1000 lines,
mine is only 100....  ************

ok, there is a reverify luckily -- this is an important step to 
look at the output and check if all seems ok. During the 
first install there will be a lot of warnings of libraries 
that do not exist yet, since they have to be built :-)
and the big inhale of code has not even started yet.



 GLISHLOLIB =-lgtkpgdrv
/bin/sh: cd: illegal option: -l
cd: usage: cd [-PL] [dir]			<<--- both in rh71 and rh72


Look at the file
     /home/aips2/code/build.log
will take not even a minute to finish this bootstrap build


$ source /home/aips2/aipsinit.csh 


       -b: get the latest stable beta, but 
       -n: do not make it yet, we do that next step

$ inhale -n -b >& inhale.log

started 20.20pm on a 604k DSL line, ended  21.12pm. Download took 30 minutes.


$ ln -s /home/aips++/data /home/aips2/data


$ mkdir ../logs-teuben
$ avers
16.118.00 Tue 2001/07/03 11:30:16 GMT


$ time gmake allsys >& ../logs-teuben/LOG.16.118
13189.810u 1907.530s 4:29:52.06 93.2%	0+0k 0+0io 61554086pf+0w

So, done at 1.50am, took 4.5 hours at 93% (not bad for a laptop!)

.. now, only if it would work again. Although the rh72 version at least worked,
this first attempt on rh71 simply doesn't. 'aips++' just sits there and nothing
is going.. very very very very slow....
o dear, i had given my laptop a local IP, and this particular hostname was now
online as something else... Darell and Joe figured that out around the GB meeting.