[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

19990108: more on compilers for solaris7



>From: address@hidden
>Organization: St. Cloud State
>Keywords: 199901081824.LAA00686

Alan-

>I appreciate the offer of binaries, but am not totally against our
>purchasing the compilers.  I guess that would require a little more
>time,   to order and receive, but probably not too much.

It will cost about $400.

>I believe you offer binary versions of ldm, and perhaps other software.
>Do you think your office 
>will be moving toward providing binaries of mcidasX in the future, at 
>least for the operating systems that are most commonly used?   If that
>is the direction your office is tilting, then compilers would not be 
>very important for us and we would wait for binaries.

For LDM, the binaries are kept up to date but then again, LDM software
doesn't change that often.  For McIDAS and GEMPAK, we
really don't have a procedure set yet for creating binaries every time
code changes are made.  That is the problem with binary releases - 
if a change is made but the binary is not updated, sites using binary
releases are out of luck.  However, OS/2 is a binary release and
when updates are made new binaries have been made also (since that
is the way we distribute McIDAS-OS2).  We just have to get in the
habit of doing this.

>I have an email off to sun for directions on getting the compilers, so
>will at least have the information.  

SSEC is going to move toward using gcc and f2c for all their platforms
in the future.  They plan on only supporting f2c and gcc on Solaris X86
so the appropriate changes to the makefiles should be in place in the
future.  The problem is that the support is not there now.

So, if a binary release can tide you over while you get your feet wet
with McIDAS-X, then you can probably get away without the Sun compilers.

Of course, sending a note to the User Committee requesting that the
UPC support binary releases for Solaris X86 or that we support
gcc under X86 wouldn't hurt either.  I've been trying to nudge Tom
in this direction for a while, but other things seem to take
priority.  If you want to be a guinea pig for testing the use
of gcc and f2c under X86, that would help.

Don