OK, to start with, I only had a couple hundred Megs free on my system, so I
decided to do each main category separately, so that I could keep tabs on
disk space with df after each run. I went to each category & ran make &&
make install. After it completed its run, I did a quick "ls", checked the
ls with the output of pkg_info, & then just typed the executable's name to
see if it'd either activate (like bonnie) or give me the synopsis line,
like anything that requires arguments. If it did either of those two, I
counted it as "working". If it gave a lib.whatever.so.#, then I noted that.
If it gave command not found, I tried doing a deinstall & reinstall of just
that package, & paying attention to the output. I forgot to do a capture of
the sessions to file, & often did them unwatched, so I did miss 99% of the
warnings that did NOT result in stops. If you like, I'll go through again
to let you know what should be cleaned up a bit, especially the 64-bit
stuff. I'm not a programming expert, nor a un*x expert as you may have
noticed, but I have done a fair amount of work on MUDs and talkers (mostly
Merc/ROM and customized MUDs, and Ew-too and Ew-3 talkers), so I'm somewhat
familiar with adjusting the source code of others, much more than with
trying to do things from scratch. If you could walk me through what to look
for, I could try helping in a more pro-active way to get the packages ready
for prime time.

Anyway, here's my findings:

==========
Archivers - Everything except "rar" works (rar is for 386 only it tells me).
==========


==========
(skipped audio since we don't yet support DEC audio)
==========


==========
Benchmarks - Bonnie worked OOTB
==========


==========
CAD - fastcap had lotsa the cast to pointer from integer of different size
crap (i.e. not 64-bit optimized), & fasthenry had the same, as well as the
type mismatch in implicit declaration for built-in function strcpy and
strlen. Is that related to the use of 64-bit processor as well, or what?
Spice gave Shared object "libXaw.so.6" not found. BTW, the package is
called spice, the man page is for spice, but the executable is spice3. It
took me a bit to figure that out.
==========


==========
Comms - Everything seemed to work OK except minicomm, but that was a
configuration glitch...it looks for a /dev/modem, & when it didn't find
one, tried to link to it from a tty#, while we use ttyC# for our ports. At
least that's my take on it. I _did_ create a symlink from ttyC1 to modem, &
then reinstalled minicom. It then went through OK. When I tried starting
it, it first warned against starting as root, then warned that there was no
config file yet, & then exited saying that it cannot open /dev/modem:
Operation not permitted. I knew that, of course, because I was already
connected to the 'Net via pppd on ttyC1. I'll assume that it would work
fine with an open port. I only tested 1 each of binaries from the pilot-*
and the lr*s* packages, which worked OK.
==========


==========
Converters - kdesupport gave "/usr/lib/c++rt0.o: file not recognized: File
format not recognized" (BTW, I had previously tried to compile the kde
meta-packge in x11 & it failed...when I get to that part, I'll let you know
how it failed).

mpack & uudeview both resulted in executables.
p5-MIME-Base64 and uulib I'm assuming neither of those are s'posed to
result in an executable, as 1 is a perl extension & the other is library
routines. At any rate, both went through make && make install just fine.

Xdeview is dependent on tk8.0p2, which I'm assuming is the package
tcl8.0p2. The make for xdeview did the whole make thing for tk8.0p2, which
resulted in the following
: libtk80.so.1.0
cc  tkAppInit.o -L/usr/pkgsrc/x11/tk80/work/tk8.0/unix -ltk80
-L/usr/pkg/lib -lt
cl80  -L/usr/X11R6/lib -lX11  -lm -lc   -o wish
ld: cannot open -ltk80: No such file or directory
Since I already had tcl installed, I dunno what is up with this?
==========


==========
Databases - gdbm installed OK. mysql failed...it used gmake, dunno if that
makes any difference, but I have no experience with gmake...

postgresql gave me the same tk8.0p2 error that xdeview did.

py-gdbm told me===>  py-gdbm-1.5.1 Needs dynamic loading...so it didn't
even start the make.

py-mysql failed because it depends on mysql...here's the tail of the gmake
stuff:
Making all in client
gmake[2]: Entering directory
`/usr/pkgsrc/databases/mysql/work/mysql-3.21.31/cli
ent'
ld -x -Bshareable -Bforcearchive libmysqlclient_pic.a -o
libmysqlclient.so.321.3
0 -lm   -lcrypt
ld: unrecognized option `-Bforcearchive'
gmake[2]: *** [libmysqlclient.so.321.30] Error 1
gmake[2]: Leaving directory
`/usr/pkgsrc/databases/mysql/work/mysql-3.21.31/clie
nt'
gmake[1]: *** [all-recursive] Error 1
gmake[1]: Leaving directory `/usr/pkgsrc/databases/mysql/work/mysql-3.21.31'
gmake: *** [all-recursive-am] Error 2
*** Error code 2
Stop.
*** Error code 1
(the stop/EC1 repeated a few times)
==========


==========
Devel - Autoconf worked, automake worked, bison worked

boehm failed with
^^^^Starting command^^^^
./alpha_mach_dep.s: Assembler messages:
./alpha_mach_dep.s:4: Error: syntax error
./alpha_mach_dep.s:4: Error: Rest of line ignored. First ignored character
is `\
'.
./alpha_mach_dep.s:5: Error: Rest of line ignored. First ignored character
is `\
'.
./alpha_mach_dep.s:30: Error: unknown opcode `call_push'
./alpha_mach_dep.s:31: Error: unknown opcode `call_push'
./alpha_mach_dep.s:32: Error: unknown opcode `call_push'
./alpha_mach_dep.s:33: Error: unknown opcode `call_push'
./alpha_mach_dep.s:34: Error: unknown opcode `call_push'
./alpha_mach_dep.s:35: Error: unknown opcode `call_push'
./alpha_mach_dep.s:37: Error: unknown opcode `call_push'
./alpha_mach_dep.s:54: Error: unknown opcode `call_push'
*** Error code 1

cvs worked, & I only did the main one...not -current.

ddd gave this:
===>  Patching for ddd-2.2.3
===>  Applying NetBSD patches for ddd-2.2.3
Syntax error: expecting "in"
*** Error code 2

gindent worked fine, gmake worked fine, gperf worked fine, idutils gave me mkid

libg++ gives me:
===>  libg++-2.8.1.1a is marked as broken: Designed for NetBSD 1.4 -
1.3(.x) com
es with this in-tree...that means that we have an older version as part of
our dist because we haven't gotten current to work, right?

libscsi went through OK...being a lib, I'm assuming there's no executable

While doing libslang, I happened to notice the following:
Configuration complete.  You may need to edit src/Makefile.
See also src/sl-feat.h for Kanji support.
Type 'make' to build normal library.
On Linux ELF systems, type 'make elf' to create ELF shared library.
===>  Building for libslang-0.99.38
I then saw it do a "make all" instead of a "make elf"...is that wrong? In
the same manner, I later noticed:
libslang.so.0.9938 created in
/usr/pkgsrc/devel/libslang/work/slang/src/elfobjs.
Use make install-elf to install it.
===>  Installing for libslang-0.99.38
cd src; make install
Should that have done install-elf instead of make install?

m4 is definitely not 64-bit ready...lotsa cast from/to crap, but worked OK.

mit-pthreads got a LONG way through, then gave:
/usr/pkgsrc/devel/mit-pthreads/work/pthreads-1_60_beta6/pthreads/process.c:
In f
unction `execl':
/usr/pkgsrc/devel/mit-pthreads/work/pthreads-1_60_beta6/pthreads/process.c:148:
warning: type mismatch in implicit declaration for built-in function `alloca'
*** Error code 1

ncurses of course works (so many depend on it), noweb, p5-Data-Dumper,
portlint, prcs, readline (no exec. associated with this I assume)

tkcvs brought that stupid ld: cannot open -ltk80: No such file or directory
error

xxgdb works OK
==========


==========
Editors - I already had xemacs installed, so I skipped "regular" emacs.

Joe wasn't 64-bit ready...& then quit with:
cc -O -DTERMINFO  -c ublock.c
cc: Internal compiler error: program cc1 got fatal signal 11
*** Error code 1

nedit also wasn't 64-bit conscious, & ended with
ld: cannot open -lXm: No such file or directory. Locate shows libXm stuff
in /usr/X11

ssam worked fine, as did vile,& I was already using xemacs

vim & xvile gave Shared object "libXaw.so.6" not found
==========


==========
Emulators - The compat11 & 12 weren't for the Alpha processor, presumably
because the Alpha wasn't supported during those versions of NetBSD. Linux
is only for i386, as is wine. Spim worked OK. Xmame gave a sh*tload of the
following before the stop:
src/cheat.c:2694: incompatible type for argument 3 of `xprintf'
src/cheat.c:2696: incompatible type for argument 3 of `xprintf'
src/cheat.c:2698: incompatible type for argument 3 of `xprintf'
gmake: *** [obj/cheat.o] Error 1
*** Error code 2
==========







I ended up here so far...I'll continue tomorrow & email an update with
(hopefully) the rest.

Hope this helps someone,
Mike




PS Reference system...AlphaStation 200 4/166. 80 MB RAM, 1 GB HD
partitioned roughly as 800 MB root&usr, 200 swap, running 1.3F from the
6/28 snapshot (everything updated), with the pkgsrc.tar.gz grabbed before
the Fourth.

Bikers don't *DO* taglines.