sims@cam.nist.gov (Jim Sims) (10/11/90)
I have recently discovered the -t option to man, for typesetting
man pages and sending to a printer. I am running 3.3.1 and using
lpr rather than lp, hence the problem. man -t sends to lp; I
want to be able to send to lpr. Anyone know how to do this?
According to the man page, *.z files are unpacked using pcat
and then sent to the named printer. Presumably there is more
processing than that going on. If I knew exactly what man does,
I could write a small script to do the job.
Thanks in advance.
-----------------------------------------------------------------------
NAME: James S. Sims TELE: (301) 975-2710
USMAIL: National Institute of Standards and Technology
(formerly National Bureau of Standards) ARPA,BITNET: sims@enh.nist.gov
Bldg. 225 Room B-146
Gaithersburg, MD 20899karron@MCIRPS2.MED.NYU.EDU (10/13/90)
Jim Simms (sims@enh.nist.gov) writes: >I have recently discovered the -t option to man, for typesetting >man pages and sending to a printer. I am running 3.3.1 and using >lpr rather than lp, hence the problem. man -t sends to lp; I >want to be able to send to lpr. Anyone know how to do this? >According to the man page, *.z files are unpacked using pcat >and then sent to the named printer. Presumably there is more >processing than that going on. If I knew exactly what man does, >I could write a small script to do the job. I don't know how to fix the man binary. I also don't know why sgi (or ATT) made the man program into a compiled binary, leaving us without the source. In the old days the man command was a script that could be fixed/hacked to local taste. I think that the lack of man page sources, and the inability of the man system to format man pages without nroff, and the whole catman compiled man page system is an attempt on the part of ATT to limit access of users to unix and protect its distribution from local disturbance. This is entirely contrary to the sprit of unix, and I have complained to sgi to look into ways it can open up the man system so that you can directly use it for what it was intended, not as a barrier to protect the ATT distribution. If there was an additional material that sgi should put in the 4Dgifts, AWF, TeX (binarys), and the "Man" system (even a local shell script) are my nominations. +-----------------------------------------------------------------------------+ | karron@nyu.edu Dan Karron | | . . . . . . . . . . . . . . New York University Medical Center | | 560 First Avenue \ \ Pager <1> (212) 397 9330 | | New York, New York 10016 \**\ <2> 10896 <3> <your-number-here> | | (212) 340 5210 \**\__________________________________________ | | Please Note :Soon to move to dan@karron.med.nyu.edu 128.122.135.3 (Mid Oct)| +-----------------------------------------------------------------------------+
ciemo@bananapc.wpd.sgi.com (Dave Ciemiewicz) (10/13/90)
In article <9010121941.AA00328@mcirps2.med.nyu.edu>, karron@MCIRPS2.MED.NYU.EDU writes: |> Jim Simms (sims@enh.nist.gov) writes: |> |> >I have recently discovered the -t option to man, for typesetting |> >man pages and sending to a printer. I am running 3.3.1 and using |> >lpr rather than lp, hence the problem. man -t sends to lp; I |> >want to be able to send to lpr. Anyone know how to do this? |> >According to the man page, *.z files are unpacked using pcat |> >and then sent to the named printer. Presumably there is more |> >processing than that going on. If I knew exactly what man does, |> >I could write a small script to do the job. |> |> I don't know how to fix the man binary. |> |> I also don't know why sgi (or ATT) made the man program into a compiled |> binary, leaving us without the source. In the old days the man command was a |> script that could be fixed/hacked to local taste. The old man command was a very slow AT&T shell script which could take up to 12 seconds to find a manual page (or no man page). Customers and internal users alike complained that this was way too slow. We replaced the shell script with binary modelled on the features of the AT&T, Sun and BSD man commands. (BTW, the Sun and BSD man commands are also binaries, not shell scripts.) Performance increased to a typical 2 seconds to lookup a man page (or tell you none were found). From the perspective of browsing online manual pages, you could browse many more manual pages with the new command than the old script, especially if you mistyped the name of the manual page you wanted. The reason for the significant speed up is that the man script was based on find(1) which is rather slow at finding files. The man command uses a more sophisticated algorithm for scanning directories for manual pages. In case of NFS mounted manual page directories, the new binary man command far exceeds the capabilities of the old man script. In some situations, the man script could take minutes to find a manual page or exhaustively search the directories to inform you that none were found. The new man executable takes much less time because of the search algorithm. New executable programs like xman can provide even faster browsing of manual pages as they can cache lookup information and provide pretty display of the manual pages. Of course, it's not a shell script either though sources may be available. With large libraries of online documentation like X man pages, GL man pages, and libc man pages, not to mention all of the add on user level man pages, I personally appreciate being able to rapidly browse online manual pages. --- Ciemo
slevy@klein.geom.umn.edu (Stuart Levy) (10/14/90)
I also am glad that SGI replaced the sluggish "man" shell script with a
speedy "man" binary. Anyone wanting more control can still get it:
with the "-w" option man prints just the pathnames of the man-pages it finds,
so you could easily write a script to process those files as desired.
Stuart Levy, Geometry Group, University of Minnesota
slevy@geom.umn.edusims@cam.nist.gov (Jim Sims) (10/14/90)
First of all, thanks to all who have responded. Since all responses have been to this newsgroup, I don't think I need to summarize. I still haven't found the solution I'm looking for, so let me try to restate what I want to do: IF I had lp set up, man -t would give me very beautiful troff formatted man pages output to a postscript printer on the network. I have all of the requisite nroff/troff/psroff etc software (Documentor's Workbench Option and LaserWriter Option). However, I don't want to use lp. It seems to me that the whole point of the 3.3 lpr implementation is that I should be able to use lpr instead of lp. However, I think I have discovered a "fly-in-the-ointment". man -t outputs to the default lp printer, with no apparent way to change this to have it output to the default lpr printer. I have been looking at man on our Suns and the Suns have an environment variable, TCAT, which would let me do exactly what I want to do. The TCAT environment variable specifies "The name of the program to use to display troffed manual pages. If not set, `lpr -t' is used. If SGI had this envrionment variable, I could do a setenv TCAT 'lpr -t' and this would give me what I want: man -t's troff binaries would be piped to lpr with the appropriate flag set so that everything would magically come out right on the postscript printer (I think). The TCAT environment variable seems like a good idea, since it adds flexibility, so, SGI, please add in a future release. Without this environment variable, can anyone see how to do the same thing? Incidentally, I've also discovered a neat feature in catman on the suns. I realize that catman isn't really relevant here, since we are shipped the results of catman. However, catman has a -p option for "tell me what you are going to do, rather than doing i"t. One line of the output of catman -p l is nroff -Tlpr -man manl/plt.l > catl/plt.l If the man command on the SGI's had a similar -p option, telling me what it does to those *.z files (or if someone else can tell me), then perhaps I could write a simple script. ----------------------------------------------------------------------- NAME: James S. Sims TELE: (301) 975-2710 USMAIL: National Institute of Standards and Technology (formerly National Bureau of Standards) ARPA,BITNET: sims@enh.nist.gov
jim@baroque.Stanford.EDU (James Helman) (10/14/90)
IF I had lp set up, man -t would give me very beautiful troff formatted man pages output to a postscript printer on the network. I don't believe it would be all THAT beautiful (unless lp(1) is smarter than I think). By the time the nroff source has been formatted into /usr/catman/*/cat1/foo.z, it has lost all the nice formatting information: bold has been converted to <character> <backspace> <character>, emphasis has been converted to <underline> <backspace> <character>. more(1) is nice enough to display this gibberish as highlighted text on most terminals or emulators. If you want nice bold and italics in your hardcopy man pages, your best bet would probably be xeroxing pages from the hardcopy manual. ;-} However, doing so could violate the copyright notice in the manuals. SIDE NOTE: SGI's copyright notice not only forbids duplication but also calls the manuals' contents "proprietary and confidential information of SGI" and bars you from dislosing it to third parties!!!! SHHH!!!! Don't tell anyone about lp(1) ;-). Is that silly, or what? Another AT&Tism? Jim Helman Department of Applied Physics Durand 012 Stanford University FAX: (415) 725-3377 (jim@KAOS.stanford.edu) Work: (415) 723-9127 "We break windows right." ctrl-shift-f12-/
sims@cam.nist.gov (Jim Sims) (10/15/90)
Jim Helman writes: >>In-reply-to: sims@cam.nist.gov's message of 13 Oct 90 23:29:08 GMT >> >> IF I had lp set up, man -t would give me very beautiful troff >> formatted man pages output to a postscript printer on the network. >> >>I don't believe it would be all THAT beautiful (unless lp(1) is >>smarter than I think). By the time the nroff source has been >>formatted into /usr/catman/*/cat1/foo.z, it has lost all the nice >>formatting information: bold has been converted to <character> >><backspace> <character>, emphasis has been converted to <underline> >><backspace> <character>. more(1) is nice enough to display this >>gibberish as highlighted text on most terminals or emulators. I see your point, and thanks for educating me. I'm now using man topic | ul -t printer | lpr for getting printed copies of on-line man pages. Scott Presnell had suggested man topic | col -b | lpr but that removes the underlining which ul with the -t printer option preserves (at least to the postscript printer I'm using). As Jim Helman has pointed out, the formatting that man topic | ul -t printer | lpr produces is the best that one can get with preformatted man pages. But I also have local unformatted additions to the man pages, and all of the nroff/troff/psroff packages, so man -t should give me very beautiful troff formatted man pages output to a postscript printer on the network (It does on the Suns!) SO my original question still stands: How do I send the output of man -t to lpr? :=) The answer to that question may very well be an SGI implementation (I can wait) of something like the Sun TCAT environment variable mentioned in my last posting. Again, thanks to everyone who has responded. ----------------------------------------------------------------------- NAME: James S. Sims TELE: (301) 975-2710 USMAIL: National Institute of Standards and Technology (formerly National Bureau of Standards) ARPA,BITNET: sims@enh.nist.gov
marinell@Iris1.UCIS.Dal.Ca (Kevin Marinelli) (10/16/90)
I have encountered the same problem with 'man -t' when trying to use lpr. The simple solution that I was able to use at our site (because I do not support lp), is to replace the lp application with the lpr application. I do not know if SGI would recomend this solution, but it has worked well for our site during the last 2 weeks. Kevin Marinelli Academic Computing Services Dalhousie University
karron@MCIRPS2.MED.NYU.EDU (10/19/90)
...lots o' stuff left out... > >With large libraries of online documentation like X man pages, GL man pages, >and libc man pages, not to mention all of the add on user level man pages, !!!!!!!!!!!!!!!!!!!!!!!!!!! >I personally appreciate being able to rapidly browse online manual pages. > > --- Ciemo My point about the man page system is that it is made artifically difficult for a user/programmer to add their own man pages without buying old nroff software, or knowing about awf. (woof woof bark bark awf awf arf arf) Man pages should by dynamic, read/write, not only read/browse. They should be much more adaptable (printers,terminals,formatters(TeX), spoolers, and indexable (keyword)) I like the increased speed, and I look forward to hashed man page stuff. Hope you don't ignore speeding up my local man pages. dan. +-----------------------------------------------------------------------------+ | karron@nyu.edu Dan Karron | | . . . . . . . . . . . . . . New York University Medical Center | | 560 First Avenue \ \ Pager <1> (212) 397 9330 | | New York, New York 10016 \**\ <2> 10896 <3> <your-number-here> | | (212) 340 5210 \**\__________________________________________ | | Please Note :Soon to move to dan@karron.med.nyu.edu 128.122.135.3 (Mid Oct)| +-----------------------------------------------------------------------------+