Newsgroups: comp.parallel.pvm
From: manchek@rudolph.cs.utk.edu (Bob Manchek)
Subject: Re: Hopes for 3.3.11
Organization: Computer Science Dept, University of Tennessee, Knoxville
Date: 28 Nov 1995 19:29:13 GMT
Message-ID: <49fnuaINNp96@CS.UTK.EDU>

In article <JMARTIN.95Nov19234450@tractor.cs.uiowa.edu>, jmartin@tractor.cs.uiowa.edu (Jeffrey Martin) writes:

|> 1) The damn tar extracted to "pvm3", not "pvm3.3.10", thus extracting
|> across a symlink I had made (for simplicity) to the current pvm (3.3.9
|> at that time), thus mumbling the files together. Ack. Please always
|> stick to the full name for the archives.

Sorry about that.  I've always done it that way, though, and I don't
want to change it in the middle of a version.

|> 2) Still no consting of constant things. How come pvm_packf can be
|> consted and none of the other were?

The format string is declared "const", but the data isn't.  PVM does
sometimes touch the data, f.e. with InPlace packing, so we shouldn't
declare those parameters as const.  We could probably declare a few
parameters in other functions const, but I've never bothered to do it.

Does anyone have an opinion on this?  It seems to me that const
declarations are becoming more of a problem as compilers enforce
passing const data only to const functions.  My thought is to in
the future just declare everything const, to avoid compiler hassles.
Chances are no smart compiler is ever going to come along and really
care.  Man, I hate ansi C.

-b

-- 
/ Robert Manchek                University of Tennessee     /
/                               Computer Science Department /
/ (423)974-8295                 Ayres Hall #104             /
/ manchek@CS.UTK.EDU.           Knoxville TN  37996-1301    /
/     http://www.netlib.org/utk/people/BobManchek.html      /


