jim@crom2.uucp (James P. H. Fuller) (02/08/91)
I recently brought up a public access Unix site from scratch, never had
Unix before, and the worst problem I've blundered into so far is the fact
that all the geniuses who write the additional software that make a Unix
system burn brightly (i.e. the news distribution software, the news readers,
the smart mailers, the non-vi editors, etc. etc.) insist on passing out
their man pages and other docs in files that are full of embedded nroff
commands. Well, I don't *have* nroff -- it was many, many extra $$$ from
my vendor and I didn't get it, not realizing that it might be crucial.
Now, some of these nroff'ed files are more or less human-readable as
they are, especially if they contain large blocks of text. But where they
start to define the cryptic little commands that make the programs go you
can't (or anyway I can't) distinguish between cryptic-little-command-
being-defined and nroff-commands-embedded-in-definition.
I'm running SysVr3.2 on a 486 AT-clone and the monthly FAQ from
comp.unix.sysv386 answers the question "Where can I get troff?" ( problem
being that regular troff only outputs to phototypsetters whereas everyone
wants to output to laser printers.) But it seems to me that WHERE CAN I GET
NROFF is a MUCH more burning question, and the FAQ doesn't mention that one.
Here's what I tried before hollering for help on the net:
1) man
I saw a one-line description of this ("A compiled 'man' program
for SysV") and obtained it from NSF.mod.sources. Turns out that it's
a utility to help manage and update the entries under /usr/catman,
but they have to be put there using nroff in the first place. Nice,
but not what I need.
2) psroff
Nope. This one takes troff output and sends it to postscript or
other less-expensive printers. A great program and I'll keep it for
later but it doesn't help me now.
3) Coherent nroff
I've got a 286 running Coherent as close as the other end of a
serial cable. Coherent comes with nroff (and troff and laserjet
support) all included in the $99 price. So, thinks I, I'll just
shoot all this stuff over to Coherent and nroff it there and then
uucp it back
Nope. Turns out Coherent's nroff isn't anything like a complete
implementation -- lots of "xxx command not implemented yet" and
the output contains missing fragments of text, frequently with no
holes to show that something dropped out. To do it justice it *did*
do the job, didn't crash, but, well, it did a $99 job. By printing
out Coherent's nroff output *and also* printing out the unprocessed
files and comparing the two a line at a time by hand I managed to get
C news and nn up. But I HOPE that's not the final solution! There's
gotta be a better way....
4) groff
Ah yes, the legendary FSF! groff, oh tantalizing groff! To
get groff you first have to obtain and compile gcc, and then obtain
and compile g++, and then you can try to build FSF groff. KIDS,
DON'T TRY THIS AT HOME! I'm a brand-new DOS retread, I program in
Turbo Pascal and MS Fortran, my knowledge of C is still strictly at
the "hello, world" level (I only recently found out that printf
doesn't mean PRINT, YOU STUBBORN #%$@))*!!) For me, building
gcc-->g++-->groff is an option for the rosy but distant future; it
sho' ain't going to happen this week.
A fellow posted a complete set of groff 0.4 binaries to comp.
unix.sysv386 not long ago ready-compiled for the exact flavor of
Unix I'm running (ISC) (and he got thoroughly flamed for it, though
not by me) but one part of his 16-part post didn't make it to a
number of sites including this one. (GRRRRR! Oh near miss! Oh
close but no cigar!) I sent mail to the poster requesting the
missing part but either my mail didn't reach him or his reply
didn't make it back to me. How can I explain to my CPU that it
really OUGHT to be able to run with 15/16ths of a program? C'mon
little chip, you can do it baby, IM-provise! That's the way, you
got it ....
In summary, I've really run into a blank wall on this. I'm trying to in-
stall a smart mailer and I want to read Larry Wall's paper on paths and
pathalias, and the paper is widely available but it's so thoroughly nroff'ed
(among other things) that it's actually broken up into a bunch of ittybitty
pieces with a makefile! As far as I'm concerned it might as well be written
in Babylonian.
WHERE CAN I GET NROFF? (I can't FTP, I'm a uucp-only site.)
Or where can I get fake nroff, or enough of nroff to read all these
&^*%$$# heiroglyphic man pages? Is it all a conspiricy to make me hand over
another 400 bucks to Interactive? I know, money solves a lot of problems.
On the other hand, if I had wings I'd fly south for the winter, if I had wings.
THANKS VERY MUCH!
James P. H. Fuller
jim%crom2@nstar.rn.comdavisp@skybridge.SCL.CWRU.Edu (Palmer Davis) (02/08/91)
In article <1991Feb07.195547.22813@crom2.uucp> jim@crom2.uucp (James P. H. Fuller) writes: > > WHERE CAN I GET NROFF? (I can't FTP, I'm a uucp-only site.) > > Or where can I get fake nroff, or enough of nroff to read all these >&^*%$$# heiroglyphic man pages? Is it all a conspiricy to make me hand over >another 400 bucks to Interactive? I know, money solves a lot of problems. > Henry Spencer wrote an implementation of nroff based on awk; it was posted to alt.sources a few months ago. It handles -man and a subset of -ms and should be just fine for the purposes you described. Just find an anonymous UUCP site that carries the alt.sources archive and you're all set. -- PTD -- -- Palmer T. Davis | davisp@scl.cwru.edu -OR- ptd2@po.cwru.edu Case Western Reserve University | {att,sun,decvax,uunet}!cwjcc!skybridge!davisp --------------------------------+---------------------+------------------------ Wake up and smell the cat food in your bank account. | Life is short.
lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall) (02/09/91)
In article <1991Feb07.195547.22813@crom2.uucp> jim@crom2.uucp (James P. H. Fuller) writes:
: In summary, I've really run into a blank wall on this. I'm trying to in-
: stall a smart mailer and I want to read Larry Wall's paper on paths and
: pathalias, and the paper is widely available but it's so thoroughly nroff'ed
: (among other things) that it's actually broken up into a bunch of ittybitty
: pieces with a makefile! As far as I'm concerned it might as well be written
: in Babylonian.
I don't know if this counts as the handwriting on the blank wall you ran
into, but I don't recall having had anything to do with a paper on paths
and pathalias.
Larry Wall
lwall@jpl-devvax.jpl.nasa.govjim@crom2.uucp (James P. H. Fuller) (02/09/91)
In <11358@jpl-devvax.JPL.NASA.GOV> lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall) writes: > I don't know if this counts as the handwriting on the blank wall you ran > into, but I don't recall having had anything to do with a paper on paths > and pathalias. 'Scuse me, you're right of course, it was Honeyman/Bellovin. (It's not hard for newcomers to Unix to get you famous guys confused at first.... one sees the luminiferous glow over the horizon before one can distinguish features....) I should have examined the paper more closely but since in its present form large chunks of it look like "\f1\s10\&``seismo''\f1\s0" at 123,240 ljust "\f1\s10\&\f1\s0" at 123,240 ljust "\f3\s10\&node\f1\s0" at 67,256 box ht 32 wid 48 with .nw at 43,248 "\f1\s10\&\(bu\f1\s0" at 67,224 line from 43,232 to 91,232 "\f1\s10\&\(bu\f1\s0" at 67,240 I did just glance at it and put it aside until I had the software to recon- struct it properly (yeah, I can tell this one's gonna take more than groff.) To make this message of more general interest to this group, let me mention that Mauro DePalma has offered to send me groff 1.0 binaries for ISC Unix, and I intend to make these available via anon-uucp for anybody who's having the same problem I did -- especially those who are trapped as I am out here in uucp-only-land and whose hearts sink to their shoes whenever they hear "The exact file you need right now can be FTP'd from....) Will post PEP number and login/passwdwhen everything's ready. James P. H. Fuller jim%crom2@nstar.rn.com
rbraun@spdcc.COM (Rich Braun) (02/10/91)
jim@crom2.uucp (James P. H. Fuller) writes: >In summary, I've really run into a blank wall on this. > > WHERE CAN I GET NROFF? (I can't FTP, I'm a uucp-only site.) > > Or where can I get fake nroff, or enough of nroff to read ... This frustrated me too, but there is a simple solution. Get flex from uunet.uu.net in directory comp.sources.unix/volume19/flex2 (files part01-07), and get xfmt from the same system (file comp.sources.unix/volume16/xfmt.Z). This will fix your problem; you can then type 'xfmt -mu foo.man > foo.1' to create readable man pages. It uses termcap to translate certain escape sequences. I also had to patch a couple of places in xfmt source, because it assumes 'printf ("%s", NULL)' prints nothing whereas the Microsoft library prints '(null)' for that case. That's the only source bug I can recall. I know nothing about how to use uucp, but I know uunet is easy enough to get to. -rich