[2617] in Release_Engineering
"delete" project for release 7.3
daemon@ATHENA.MIT.EDU (Jonathan I. Kamens)
Tue Jun 4 23:10:22 1991
Date: Tue, 4 Jun 91 23:09:58 -0400
From: "Jonathan I. Kamens" <jik@pit-manager.MIT.EDU>
To: rel-eng@ATHENA.MIT.EDU
I have prepared the "delete" project for pickup by release
engineering for release 7.3. The following comments concern the
testing I performed to verify that it compiles on all platforms; they
should be of interest to release enginering.
Vax and RT compiled with no problems.
Pmax, on paris, I had to compile without AFS support enabled, for
two reasons:
1. There are no AFS libraries and include files in the build tree for
pmax_ul4.
2. The pmax_ul3 AFS header files conflict with <sys/ioctl.h>, because
they try to use the old preprocessor method of 'x' and having the x
substituted in a macro, and that won't work. The AFS header files
need to be fixed.
I also had to compile with "ATHLIBDIR=/usr/athena/lib" on the command
line, because the Imake config files seem to think that the libraries
are installed in /usr/local/lib on the pmax, and they're not. I don't
know if the build will succeed with AFS enabled. This should
definitely be fixed; if you have to modify my source code to fix it,
please send changes back to me so I can put them back into my source
tree.
On the rios, on tardis, I was trying to use the _AIX and AIXV
symbols as the document in /source/control/doc says I should be able
to, but they didn't appear to be working properly. I settled on
_IBMR2, which does appear to work. The stuff in /source/control/doc
also says that I should be able to rely on POSIX being set when the
compiler is run, but it doesn't appear to be, so my Imakefile has been
modified to set it explicitly. If you fix things so that it gets set
properly, then please notify me so that I can remove the hack from my
Imakefile.
Once I dealt with the _AIX and AIXV and POSIX problems, everything
compiled fine on the rios. I had to make some modifications to the
sources to account for the fact that malloc(0) returns 0 on the rios.
Ugh.
I have not tested yet on the Mac because I don't know what machine
I'm supposed to use for testing. If you come up with a machine that
is prepared to to Mac testing, let me know and I'll deal. If you end
up deciding to tset on the Mac yourselves, please let me know what
changes (if any) you have to make to the Imakefile or the sources so
that I can incorporate them back into my source tree.
*************************
The following text should appear in the release notes:
Lsdel now defaults to multi-column output if output is to a terminal,
or single-column output otherwise. The "-1" and "-C" options are
provided to force single-column or multi-column output. The lsdel(1)
man page has been updated to reflect these changes. (Fixes changes.73
No. 1083; fix is by dskern and jik)
Lsdel now sorts filenames before printing them. (Fixes changes.73 No.
1082; fix is by jik)
Expunge can now cope with the situation where a symlink pointing to a
nonexistent file and a deleted file with the same name exit in a
directory, and the user asks it to expunge the deleted file. (Fixes
changes.73 No. 1179; fix is by jik)
Undelete now continues prompting for files to undelete in a directory
after a file in the directory fails to be undeleted. (Fix is by jik)
Undelete can now deal with sockets. (Fixes changes.73 No. 989; fix is
by jik)
The expunge(1) man page has been corrected so that its whatis info
comes out correct. (Fixes changes.73 No. 1440; fix is by jik)
*************************
Release engineering can pick up the source code by tarring
/afs/athena/astaff/project/delete/src into
/source/athena/bin.athena/delete. However, they should make sure to
use "xmkmf /source" to rebuild the Makefile right after doing this,
because the Makefile that's there right now is the one I release when
I distribute delete to the network.
Please let me know if you have any problems.
jik
