[5954] in SIPB bug reports
Andrew K Percey: Latex/dvips
daemon@ATHENA.MIT.EDU (Albert Dvornik)
Mon Apr 29 00:47:19 1996
From: "Albert Dvornik" <bert@MIT.EDU>
To: sipb@MIT.EDU, bug-sipb@MIT.EDU
Date: Sun, 28 Apr 1996 12:23:46 EDT
I didn't know sipb-request *existed* (it is an alias for sipb-acl).
------- Forwarded Message
To: sipb-request@MIT.EDU
Cc: akpercey@MIT.EDU
Subject: Latex/dvips
Date: Sun, 28 Apr 1996 00:31:35 EDT
From: Andrew K Percey <akpercey@MIT.EDU>
The following mail is the response I received from an OLC consultant
about a problem I'm having with figures in Latex. Apparently, the
current Latex/dvips combination does not allow any cross-hatching
created in Xfig to come through to the final product. This problem
also shows up when using xdvi to view the latex'd file.
Are there newer versions of latex or dvips that eliminate this
incompatibility? Any help you can offer would be greatly appreciated.
This is for my Master's thesis.
Thanks,
Andrew Percey
akpercey@mit.edu
------- Forwarded Message
Received: from SOUTH-STATION-ANNEX.MIT.EDU by po8.MIT.EDU (5.61/4.7) id AA19391; Sat, 27 Apr 96 12:49:47 EDT
Received: from COPILOT.MIT.EDU by MIT.EDU with SMTP
id AA04444; Sat, 27 Apr 96 12:49:45 EDT
From: mouse@MIT.EDU
Received: by copilot.MIT.EDU (5.0/4.7) id AA25950; Sat, 27 Apr 1996 12:50:55 -0400
Date: Sat, 27 Apr 1996 12:50:55 -0400
Message-Id: <9604271650.AA25950@copilot.MIT.EDU>
To: "Andrew K Percey" <akpercey@MIT.EDU>
Cc:
Subject: Your OLC question about "latex"
Content-Length: 12061
--------
Hi again!
Another response to your question:
> 2) When latex reads .eps files that I created in xfig, it translates the
> cross-hatchings to solid black. How can I get it to read the "fillings"
> correctly?
And then:
> Hello there. Has my question been lost in limbo?
Well, I dig some digging through old questions in OLC. I thought I
remembered something like yours. And I found it.
Unfortunately, the conclusion of the question was that dvips/latex is
not able to deal with hatch. Actually, it's probably the current
version of dvips/latex in Athena that is the problem.
I've included below most of the last user's question. (Anonymity
preserved.) You might want to read it to see what our suggestions were
to the guy. Especially the stuff at dates/times:
[Wed 03-Jan-96 6:18pm]
[Thu 04-Jan-96 7:10am]
I'll leave your question around for a few more days, but I have a
feeling that this is all that we'll be able to do. Sorry.
--Arthur Housinger, consultant
------------------------------------------------------------------
[13949] root@ATHENA.MIT.EDU OLC-latex 01/10/96 22:30 (455 lines)
Subject: Problems with Xfig 3.1 hatch shading with Athena latex/dvips.
Log Initiated for user
[Wed 03-Jan-96 4:22pm]
Topic: latex
Question:
Hi
I have created a few schematics in XFIG that I would like to include in
latex. I saved the XFIG file in 'postscripy' and included in my document
using psbox.
The strange thing is that, I have a polygons in my schematics that are cross-hatched and they print out fine if i directly send the saved postscript file to theprinter. But when i dvips my document and print it out the same figures have
the polygons filled in black.
I get the follwing message when I dvips the latex document.
dvips: Unknown keyword (.ps) in \special will be ignored
This message follows immediately after the page where the XFIG file is present.
Could it be a problem of dvips / xfig upgrades. A more recent version is
available ....
Machine info:
Processor: Athena Workstation (sun4) Version 7.7U Mon Oct 30 23:42:13 EST 1995
Display : SUNW
Display : SPARCstation-5
Display : cgthree
Memory : 26052 user; 31144 (31 M) total
___________________________________________________________
--- Question grabbed by consultant consultant [0].
[Wed 03-Jan-96 4:34pm]
*** Reply from consultant consultant [0].
[Wed 03-Jan-96 4:35pm]
Hello, User...
I assume that you mean "postscript" when you say "postscripy". How
are you including the PostScript file in your LaTeX file?
--- User read reply.
[Wed 03-Jan-96 4:35pm]
*** Reply from user user [0].
[Wed 03-Jan-96 4:37pm]
I am using psbox to include the figures.
\begin{figure}
\centerline {\PSbox{/vandiver/data/user/jasa/ppr2/mat_ps/schem4lyr.ps hoffset=0voffset=0 hscale = 92.6 vscale = 92.6}{6.02in}{3.72in}}
\vspace{0.1in}
\caption[Schematic of fluid-filled borehole with drillpipe]
{Schematic of fluid-filled borehole with drillpipe \label{fi:schem4lyr}}
\end{figure}
The above is a sample ..
User
*** Reply from consultant consultant [0].
[Wed 03-Jan-96 4:39pm]
Hmmm... can you type "where xdvi" and "where dvips" and tell me
what you get?
--- User read reply.
[Wed 03-Jan-96 4:39pm]
*** Reply from user user [0].
[Wed 03-Jan-96 4:40pm]
Sure..
[~/acoustics/new_ac/array ] 83
% where xdvi
/usr/athena/bin/xdvi
/mit/sipb/sun4bin/xdvi
[~/acoustics/new_ac/array ] 84
% where dvips
/usr/athena/bin/dvips
*** Reply from consultant consultant [0].
[Wed 03-Jan-96 4:41pm]
Okay. Can you save your Xfig drawing as EPS instead, and try using
that in LaTeX with both \PSbox and \epsfbox?
--- User read reply.
[Wed 03-Jan-96 4:42pm]
*** Reply from user user [0].
[Wed 03-Jan-96 4:42pm]
ok...
I'll try that..
*** Reply from user user [0].
[Wed 03-Jan-96 4:46pm]
How do i include a eps file ?
is it just like a ps file ?
*** Reply from consultant consultant [0].
[Wed 03-Jan-96 4:46pm]
Use "epsf" in your documentstyle options, and then
juse use \epsfbox{filename.eps}
--- User read reply.
[Wed 03-Jan-96 4:46pm]
*** Reply from user user [0].
[Wed 03-Jan-96 4:55pm]
no that did not help..
..
*** Reply from consultant consultant [0].
[Wed 03-Jan-96 5:03pm]
It didn't work with \epsfbox either? Same error message from
dvips?
--- User read reply.
[Wed 03-Jan-96 5:03pm]
*** Reply from user user [0].
[Wed 03-Jan-96 5:06pm]
no I did not get the warning message..
I dvips'ed only the pge on which the graphic was.
I did not get a message even when i used .ps when i dvips'ed only the
specific page.
I got the message when I dvips'ed the entire ducument only..
*** Reply from user user [0].
[Wed 03-Jan-96 5:06pm]
And yes.. It did not work with epsf..
same trouble..
*** Reply from consultant consultant [0].
[Wed 03-Jan-96 5:08pm]
Hmmm...
I'm not sure what to suggest then.
--- User read reply.
[Wed 03-Jan-96 5:08pm]
*** Reply from consultant consultant [0].
[Wed 03-Jan-96 5:13pm]
What happens if you use "xfig.old" in the SIPB locker?
*** Reply from user user [0].
[Wed 03-Jan-96 5:15pm]
Some of my officemates use xfig in latex with fill patterns on 'non-athena'
workstations (DEC5000 and DEC alpha) and they said that it could be bug
in dvips that was ficed on a later version. They have used polygons with patterns succesfully on those machines, ehich have later versions of dvips.
I wonder if that could be a problem..
*** Reply from consultant consultant [0].
[Wed 03-Jan-96 5:17pm]
It may or may not be the problem. The version of dvips on Athena
should go with the version of LaTeX, I believe...
--- User read reply.
[Wed 03-Jan-96 5:17pm]
*** Reply from user user [0].
[Wed 03-Jan-96 5:20pm]
They were suggesting an incomaptibility between the versions of dvips and
xfig on athena.. that cuased dvips to misinterpret the patterns..
*** Reply from consultant consultant [0].
[Wed 03-Jan-96 5:23pm]
I don't think dvips actually interprets the patterns, if Xfig is actually
outputting PostScript. Dvips is a post-processor for .dvi files,
and handles such things as "special" insets, of which PS files are
one -- mainly handling the PostScript code required for saving state
before the PS file is inserted (including sizing and rotation), and
then restoring the state after.
--- User read reply.
[Wed 03-Jan-96 5:23pm]
*** Reply from user user [0].
[Wed 03-Jan-96 5:24pm]
ok..
*** Reply from consultant consultant [0].
[Wed 03-Jan-96 5:24pm]
It's possible that newer versions of Xfig may be attempting to
do something tricky, which older versions of dvips might not
know about.
--- User read reply.
[Wed 03-Jan-96 5:24pm]
*** Reply from user user [0].
[Wed 03-Jan-96 5:24pm]
true..
*** Reply from consultant consultant [0].
[Wed 03-Jan-96 5:25pm]
Have you tried using the older version of Xfig and seen if that
works?
--- User read reply.
[Wed 03-Jan-96 5:25pm]
*** Reply from user user [0].
[Wed 03-Jan-96 5:25pm]
how do I access that ?
*** Reply from consultant consultant [0].
[Wed 03-Jan-96 5:28pm]
It should be in the SIPB locker for the DECs, IBMs and Suns (not for
the SGIs)
--- User read reply.
[Wed 03-Jan-96 5:28pm]
*** Reply from user user [0].
[Wed 03-Jan-96 5:28pm]
ok.. i'll try that ..
*** Reply from user user [0].
[Wed 03-Jan-96 5:38pm]
The old version does not have a 'filling with a pattern' option!!
It only allows solid colors of varying intensity , which the current version
also allows.
--- Question from User [0] on "latex" has been forwarded.
[Wed 03-Jan-96 5:38pm]
*** Reply from user user [0].
[Wed 03-Jan-96 5:58pm]
I changed the 'fill with pattern' to fill with solid color' and
it conmes out fine when i dvips my document.. (when i use the current
vesion of xfig from the dash).
User
--- Question grabbed by consultant consultant [0].
[Wed 03-Jan-96 6:15pm]
*** Reply from consultant consultant [0].
[Wed 03-Jan-96 6:18pm]
Sorry about that. Had to logout for a while (looks like I'll have
to logout again though).
It sounds like that "fill with pattern" uses something which isn't
handled in the older versions of dvips but was specifically for the
newer versions. I don't know what to suggest, except perhaps to ask
that SIPB try building a newer version of dvips for their outland
locker (that might take a while, though -- they're a volunteer group,
and of course it's IAP). The supported version on Athena is unlikely
to change until next fall.
--- Question forwarded to pending.
[Wed 03-Jan-96 6:21pm]
--- User read reply.
[Wed 03-Jan-96 6:39pm]
--- User has logged out.
[Wed 03-Jan-96 6:51pm]
--- Comment by consultant boojum@UNCA-DONALD.MIT.EDU [0].
[Thu 04-Jan-96 7:10am]
Something I've seen in DOS apps (no idea if this is the problem here)
is that a fill with pattern which is at maybe 72dpi on the screen gets
output as a fill with pattern at 600dpi when it goes to the printer,
thus making just about all patterns look like a shade of grey...
--- User has logged back in.
[Wed 10-Jan-96 9:39pm]
--- Question grabbed by consultant consultant [2].
[Wed 10-Jan-96 9:40pm]
*** Reply from consultant consultant [2].
[Wed 10-Jan-96 9:40pm]
Howdy... can you copyolc the Xfig file (both .fig and .ps)?
--- User read reply.
[Wed 10-Jan-96 9:40pm]
*** Reply from user user [0].
[Wed 10-Jan-96 9:42pm]
sure..
*** Reply from user user [0].
[Wed 10-Jan-96 9:47pm]
copyolc has copied /vandiver/data/user/jasa/ppr2/mat_ps/olc_f.fig /vandiver/data/user/jasa/ppr2/mat_ps/olc_f.ps to /mit/readable/user/submit.
created the pickup directory. to set right acls, type:
fs sa /mit/readable/user/pickup user read
*** Reply from user user [0].
[Wed 10-Jan-96 9:47pm]
I've copied them
*** Reply from consultant consultant [2].
[Wed 10-Jan-96 9:48pm]
Okay... let me look over it.
--- User read reply.
[Wed 10-Jan-96 9:48pm]
*** Reply from consultant consultant [2].
[Wed 10-Jan-96 10:06pm]
Hmmm... dunno what to suggest, then. It just looks like the
pattern fills (patch) don't work with that version of latex/dvips
(dunno which part is generating the problem).
--- User read reply.
[Wed 10-Jan-96 10:06pm]
*** Reply from user user [0].
[Wed 10-Jan-96 10:11pm]
copyolc has copied /vandiver/data/user/jasa/ppr2/mat_ps/olc_f2.ps to /mit/readable/user/submit.
*** Reply from user user [0].
[Wed 10-Jan-96 10:15pm]
ok..
I;
've just copied another postscript file to olc. It is longer and has
comments like this:
} stopped {
(*** PATstroke Warning: Path is too complex, stroking
with gray) =
cleartomark
etc. I do not know if it is significant, but i thoiught i'd let you know.
I could not get the pattern to print right on the figures copie4d earlier..
and in the current one that i just copied ..
*** Reply from consultant consultant [2].
[Wed 10-Jan-96 10:21pm]
It's a nice figure, but it comes out with solid black also...
*** Reply from consultant consultant [2].
[Wed 10-Jan-96 10:21pm]
Without the warning.
--- User read reply.
[Wed 10-Jan-96 10:21pm]
*** Reply from user user [0].
[Wed 10-Jan-96 10:26pm]
I guess if it is an incompatibility between versions of xfig and dvips nothing can be done about it..
*** Reply from consultant consultant [2].
[Wed 10-Jan-96 10:27pm]
If you have access to the other versions of dvips, you might want to
try them and see what happens
--- User read reply.
[Wed 10-Jan-96 10:27pm]
*** Reply from user user [0].
[Wed 10-Jan-96 10:28pm]
ok.. I'll try that. Apparently it works fine on the latest versions of
xfig and dvips.
--- Conversation terminated at Wed 10-Jan-96 10:30pm
--- Title: Problems with Xfig 3.1 hatch shading with Athena latex/dvips.
--[13949]--
------- End of Forwarded Message
------- End of Forwarded Message
