[12971] in Athena Bugs

home help back first fref pref prev next nref lref last post

rsaix 7.7F: Cygnus gcc & Ted

daemon@ATHENA.MIT.EDU (John Hawkinson)
Fri Dec 2 02:07:16 1994

To: bugs@MIT.EDU
Cc: tytso@MIT.EDU
Date: Fri, 02 Dec 1994 02:07:01 EST
From: John Hawkinson <jhawk@MIT.EDU>

System name:		pickled-herring
Type and version:	POWER 7.7F (1 update(s) to same version)
Display type:		POWER_Gt1

What were you trying to do?
	Run gcc from the cygnus locker with a nonstandard PATH.

What's wrong:

1]	It fails to find machtype. /mit/cygnus-94q2/common/scripts/gcc,
	also known as /mit/cygnus/{rsaix,sun4,decmips}bin/gcc. For instance:

% echo $PATH
/usr/bin:/etc:/usr/sbin:/usr/ucb:/usr/bin/X11:/sbin
% /mit/cygnus/rsaixbin/gcc -v
Can't exec "machtype": No such file or directory at /mit/cygnus/rsaixbin/gcc line 41.
GCC_DEFAULT_OPTIONS=
gcc version cygnus-2.5.90-940429

2]	Ted is broken.

What should have happened:

1]	The line that reads:

	  $machtype = `machtype`; chop $machtype;

	should read:

	  $machtype = `/bin/athena/machtype`; chop $machtype;

	resulting in having gcc work. It should be noted that this is a bug
	for all 3 athena platforms. It has been suggested to me by an
	over-the-shoulder reader that it is reasonable to attempt
	to invoke machtype prior to /bin/athena/machtype; I think this
	is unnecessary, but it certainly acceptible.
	
2]	Shoot Ted.
	More specifically, per the comments:

# 94/01/13   tytso    Converted script to perl, so that shell metacharacters
#                     in gcc's command line don't get hosed.

	Ted converted this script to perl. *HOWEVER*, the construction
	he uses in perl:

exec("$0.real", @ARGV) || die "Can't run $0.real: $!\n";

	Can be replaced by the Bourne shell construction:

$0.real {$1-"@$"}

	Which is a less-than-commonly-known, portable, construction.

Please describe any relevant documentation references:

1)	The thread, "mosaic outputs cruft to stdout" in the bug-sipb
	discuss meeting on bloom-picayune, consisting of transactions
	[4984] [4986] [4987] [4988] [4992] [4993] [4994] [4995] [4996]
	[4997] [4998] [4999], where the above was recently demonstrated.

2)	PERL CONSIDERED HARMFUL, a work in progress, by John A. Hawkinson.
	Please send requests to perl-considered-harmful-request@panix.com.

3)	Ted is supposed to be at IETF in San Jose next week; if you don't
	catch him before he goes, you might ask one of the following
	MIT people to shoot him for you:

  55 Berners-Lee, Tim        Massachusetts Institute of Technology
 139 Clark, David            Massachusetts Institute of Technology
 200 Eichin, Mark            Cygnus Support
 322 Horowitz, Marc          OpenVision Security Business Unit
 426 Linn, John              OpenVision Technologies
 638 Schiller, Jeffrey       Massachusetts Institute of Technology
 674 Sollins, Karen          Massachusetts Institute of Technology
 719 Ts'o, Theodore          Massachusetts Institute of Technology
 773 Wroclawski, John        Massachusetts Institute of Technology

	Barring the above, shooting a real-time image of Ted over the
	MBONE real-time is permissable; channels are:

    IETF Channel 1 Video        224.0.1.12      4120    0
    IETF Channel 2 Video        224.0.1.15      4150    0

4)	As far as I know, Ted is not an Athena supported program
	(/mit/tytso/README.athena is nonexistant, as is
	/afs/net/user/tytso/README.athena, as is
	/afs/sipb/user/tytso/README.athena. As an aside,
	/mit/tytso/README.serial does not discuss the serial nature
	of Ted, nor the serial shooting of him); as such, certain
	parts of this bug report, though intend to be humorous, may
	not be whoally appropriate to this forum.

--jhawk

home help back first fref pref prev next nref lref last post