Sun-Spots-Request@RICE.EDU (William LeFebvre) (05/16/88)
SUN-SPOTS DIGEST Monday, 16 May 1988 Volume 6 : Issue 88 Today's Topics: Re: Experience with DECNET Re: Need help writing in SunView Re: My favorite shell file Re: Default route problem Re: Bug in initgroups Re: Looking for APL for Sun NFS and executing binaries including rasterfiles in TeX documents 2351 Eagle on Interphase? make bug? lint library for Sunview? Sun MCP equivalent for Sun386i? Send contributions to: sun-spots@rice.edu Send subscription add/delete requests to: sun-spots-request@rice.edu Bitnet readers can subscribe directly with the CMS command: TELL LISTSERV AT RICE SUBSCRIBE SUNSPOTS My Full Name Recent backissues are available via anonymous FTP from "titan.rice.edu". For volume X, issue Y, "get sun-spots/vXnY". They are also accessible through the archive server: mail the request "send sun-spots vXnY" to "archive-server@rice.edu" or mail the word "help" to the same address for more information. ---------------------------------------------------------------------- Date: 11 May 88 18:24:48 GMT From: laic!darin@decwrl.dec.com (Darin Johnson) Subject: Re: Experience with DECNET >From: gfr%wolfgang@gateway.mitre.org (Glenn Roberts) > >With regard to peripheral sharing, the DNI product does not really give >you any utilities for doing this (it would have been nice, for example, if >Sun had provided a device driver to allow a VAX printer to simply look >like the sun's lp - I know of no way to do this, short of writing your own >device driver using the routines they provide). I have had more success without writing a device driver and using printcap. I haven't really finished it (probably never will since I cannot get docs/sources/help with how printing works on UNIX), but it works fine in most cases. Details: Create an input (or output, I forget) print filter. This filter ignores its normal output, instead it opens a DECnet link and sends the data over this. At the VAX end, I have a program from a DECUS tape that retrieves this data into a temp file and then prints that (with no header, so that you don't get the temp file name printed - remote side should send a 'fake' header). Of course, e-mail support is available, just wait for my posting to show up. [[ I had problems unpacking his e-mail support package for dnamail. Hopefully we will soon have it all straightened out. --wnl ]] >This product is a bit disappointing in terms of breadth of capabilities. >I have worked with DEC's DECnet DOS (for the PC), and feel that it offers >a much nicer set of capabilites, including sending E-mail, copying and >submitting batch files, and disk and printer peripheral sharing. ?? Gosh, you mean that you can do more with DECnet DOS than you can with the DECNET on VMS??? Without resorting to clusters, how can you submit batch files and share disks and printers between vaxen? (especially sharing printers, I am supposed to be working on that at the moment) Darin Johnson (...lll-lcc.arpa!leadsv!laic!darin) (...ucbvax!sun!sunncal!leadsv!laic!darin) ------------------------------ Date: 11 May 88 19:06:30 GMT From: nomi%phri@uunet.uu.net (Nomi Voroba) Subject: Re: Need help writing in SunView Reference: v6n76 > Specifically, I would like to see how, for instance, view surface of > SunCore is assigned to a subwindow of SunView so that SunCore application > can run under SunView as if it ran alone. How's about the following: Set up a canvas as a subwindow of your (SunView) frame canvas=window_create (frame, CANVAS, ......,0); Get the file descriptor associated with this canvas dev_fd = (int ) window_get(canvas, WIN_FD); Get the name of the window device associated with this file descriptor win_fdtoname(dev_fd, canvas_name); Set the graphics window to be used by SunCore to be this window we_setgfxwindow(canvas_name); Initialize core, vsurf, devices, etc. Get the whole thing running window_main_loop(frame) P.S. I don't think temporary segments will work. Nomi Voroba...Guberman {allegra, philabs}!phri!nomi Public Health Research Institute 455 First Avenue, New York, NY 10016 ------------------------------ Date: Thu, 12 May 88 11:56:24 BST From: cmc%informatics.rutherford.ac.uk@nss.cs.ucl.ac.uk Subject: Re: My favorite shell file I also like to use menus for as much of my interaction with a Sun as posible and to this end I use a menu of remote hosts for starting rlogin windows but, like wnl, also find that the number of possible hosts is too large for a menu. Often the name of the host I want to log into is on the screen somewhere so I take advantage of this with the following additional entry in the menu file which reads: "rlogin @" sh -c "exec shelltool rlogin `winselection`" This causes the current (shelltool) text selection to be used as the target host for an rlogin. (I use the "@" symbol in my menu files where ever the current text selection is used: e.g. "perfmeter @") The command winselection is a simple shell script: #!/bin/sh if [ -f /tmp/winselection ] then exec tail +2 /tmp/winselection fi exit 1 # no winselection file If you tend to use command tool type windows then the get_selection command could be used instead. Chris. Chris M Crampton UK JANET: cmc@uk.ac.rl.inf Rutherford Appleton Labs, ARPA: cmc%inf.rl.ac.uk@nss.cs.ucl.ac.uk Didcot, OXON, U.K. UUCP: ..!mcvax!ukc!rlvd!cmc +44 235 21900 ext. 6756 cmc@rlvd.uucp ------------------------------ Date: Thu, 12 May 88 09:24:50 CDT From: "Matt Crawford" <matt@oddjob.uchicago.edu> Subject: Re: Default route problem Reference: v6n79 Phone: +1 312 702 8207 The final argument to /usr/etc/route add ... is NOT a hop count. It is not a metric at all, in spite of the fact that TFSource calls it one. The only use of the last argument is that if it is > 0, the route is marked as being a route through a gateway. If the kernel decided that the default net 0 matched the address of an unconfigured interface, that's a bug. Matt Crawford ------------------------------ Date: Thu, 19 May 88 00:00:47 EDT From: Rayan Zachariassen <rayan@ai.toronto.edu> Subject: Re: Bug in initgroups > Even before my posting finally(!) appeared in comp.sys.sun, I got the > answer, from Chris Maio, right here at Columbia... > I also received a response from Rayan Zachariassen <rayan@ai.toronto.edu>, > with the same fix, but I think his reasoning as to the nature of the bug > was less accurate. Ahem, Chris and I are giving different views of the problem. Chris is explaining why you are observing this behavior by explaining what actually goes on; I was saying that this behaviour is broken (not only in SunOS, but in 4.3bsd too). The process rgid/egid values are conceptually orthogonal to the group access list. The documentation makes no link between the two. We ran into this problem here when we got 4.3, but it wasn't important enough to spend time fixing. That doesn't mean it is supposed to work that way. (Alas, I'm told many 4.3bsd programs depend on this bad behaviour). I suggest that Sun either fix the code to work properly, or properly document the behaviour in the manual pages. I prefer the former. rayan ------------------------------ Date: Wed, 11 May 88 23:25:17 EDT From: Tom Molnar <molnar@gpu.utcs.toronto.edu> Subject: Re: Looking for APL for Sun Reference: v6n78 IP Sharpe sells a version of APL for Suns that seems quite reasonable (though I am not an APL expert). I know there is an IPSA shareware APL available for the PC. I don't know if it is available in source form or not. Tom Molnar Unix Systems Group, University of Toronto Computing Services. ------------------------------ Date: Thu, 12 May 88 13:00:55 EDT From: Chuck Musciano <chuck@trantor.harris-atd.com> Subject: NFS and executing binaries A while ago I posted a message regarding spurious "MC68881 missing" messages. Since then, we have tracked the problem to a hole in NFS. The scenario goes something like this: We have 24 machines (50s and 60s) connected to two servers. The servers have a /usr/local/bin, which everyone mounts via NFS to get copies of local executables. Further, we have about 15 local windowed tools which are combined into one giant load file to save space. This combined binary is in /usr/local/bin. When you execute any of our tools, you actually are executing (via a link) the combined binary. The combined binary also includes custom local versions of suntools and shelltool. Normally, a program is demand paged from the file which holds its binary as it executes. Unix remembers if a file is associated with an executing process, and won't let you copy over it. You can force its removal, but only the directory entry is removed. The actual disk space for the file is not released until all processes using the file have terminated. Unfortunately, this behavior only occurs if the binary is on a local, physical disk volume. NFS !does not! retain such usage information over the net. If you execute a binary accessed via NFS, you still demand page from the binary, but the binary file can be removed or copied over while you are executing. If this happens, and the process goes to get another page, all sorts of problems can arise. Each time we had the MC68881 missing problem, I had just installed a new version of our shared tools file. As each machine went to page in some- thing new, problems would occur. Some people who had been running a while had no problems, since they had completely paged the binary in, and sub- sequently paged from their swap space instead. In addition to the missing MC68881 message, we also machines panic with a "text size changed" message, and other people simply had windows disappear, one at a time, as new windows were opened. A coworker told me that NFS does not guarantee full Unix semantics across the net, and I guess this is the problem here. I don't know what a good fix is. As usual, caveat user is probably the best advice. Chuck Musciano Advanced Technology Department Harris Corporation (407) 727-6131 ARPA: chuck@trantor.harris-atd.com ------------------------------ Date: Thu, 12 May 88 15:36:13 BST From: Richard Tobin <richard%aiai.edinburgh.ac.uk@nss.cs.ucl.ac.uk> Subject: including rasterfiles in TeX documents Here is a program which takes a rasterfile (screendump) image, and embeds it in postscript in a way suitable for inclusion in TeX files using the \special command. It's only useful if you use dvi2ps or something compatible. It's called rf2spec, because it converts from rasterfile format to \special format. [[ Sorry, but it does no such thing. It converts a rasterfile to Postscript and then generates a \special command that is very likely understood only by one piece of software: dvi2ps (admittedly, a very popular piece of software). --wnl ]] I'm sorry about the manual page; I have no plans to learn troff. -- Richard Richard Tobin, JANET: R.Tobin@uk.ac.ed AI Applications Institute, ARPA: R.Tobin%uk.ac.ed@nss.cs.ucl.ac.uk Edinburgh University. UUCP: ...!ukc!ed.ac.uk!R.Tobin [[ You all will probably hate me for this, but I placed the shar file in the archives. It is only 5652 bytes long, but that's 34% of a typiccal digest. It is stored as "sun-source/rf2spec.shar". It can be retrieved via anonymous FTP from the host "titan.rice.edu" or via the archive server with the request "send sun-source rf2spec.shar". For more information about the archive server, send a mail message containing the word "help" to the address "archive-server@rice.edu". --wnl ]] ------------------------------ Date: 11 May 88 13:54 -0400 From: Phil Trubey <ptrubey%sce.carleton.cdn@ean.ubc.ca> Subject: 2351 Eagle on Interphase? Does anyone know if you can put a 2351 Fujitsu Eagle disk drive on a 2180 Interphase controller? We have an old Sun 2, and we'd like to do this. Some old documentation states that this cannot be done - it does not say why, and I would have thought it possible since they are all SMD devices. Thanks for any info, Phil Trubey ptrubey@sce.uucp (...watmath!sce!ptrubey) ptrubey@sce.carleton.ca ------------------------------ Date: Thu, 12 May 88 10:36:44 EDT From: spencer@crim.eecs.umich.edu (Spencer W. Thomas) Subject: make bug? I have detected what appears to be a bug in /bin/make. Start with this makefile: makefile: makefile.src cp makefile.src makefile Then copy it to makefile.src (so that makefile.src is newer than makefile). Execute % /bin/make makefile The result is `makefile' is up to date. I have tried this on two different systems, here is their version info: Sun UNIX 4.2 Release 3.4 (IPANDRIM) #15: Wed Apr 27 14:41:12 EDT 1988 Sun UNIX 4.2 Release 3.5 (GENERIC.ip) #1: Mon Apr 4 16:54:06 EDT 1988 I couldn't figure out how to get version info from make, but I noticed that this version of make has quite different debugging output than others I've used. Here is a sample: % /bin/make -dd makefile MAKEFLAGS value: stat(default.mk) doname(default.mk) Looking for % rule for default.mk find_double_suffix_rule(default.mk) find_suffix_rule(default.mk,default.mk,) ... Any ideas? =Spencer (spencer@crim.eecs.umich.edu) ------------------------------ Date: Thu, 12 May 88 09:39:29 PDT From: alan@cogswell.jpl.nasa.gov (Alan S. Mazer) Subject: lint library for Sunview? Does anyone know where I can get a lint library for Sunview? They seem to have pixrect and cgi and everything else covered except for their main windowing package. Thanks! -- Alan ..!cit-vax!elroy!alan elroy!alan@csvax.caltech.edu ------------------------------ Date: Thu May 12 09:51:36 1988 From: portal!cup.portal.com!jxh@sun.com Subject: Sun MCP equivalent for Sun386i? Having heard rumors of the existence of the Sun Multiprotocol Communications Processor (MCP) which, I presume, is a VME board that does BOP (SDLC) at relatively high speeds (I need at least 1Mbit/s; 2.4Mb/s would be nice), I have the following problem: I intend to use a Sun386i, which is not a VME machine. It has PC-AT-like slots. Does anyone know of a PC I/O Channel card that performs this function? I am using an AST CC-232-E, but it needs an interrupt per character, which is limited by the interrupt latency of the machine. This means about 9.6Kb/s on an AT&T PC 6300, perhaps 19.2Kb/s on a 6MHz AT clone. Here are my requirements: - Ability to send and receive SDLC frames at 2.4Mb/s (b=bit, B=byte). This entails: - On-board processor - Dual-port RAM - DMA from the serial chip into the first stage of memory - Ability to pass frames to the host processor using I/O Channel DMA -or- On-board memory mapped into I/O Channel memory space (DMA preferred) - RS-232 or RS-422/423 (232 preferred) I have heard of SeaLevel, AST, and Tecmar. SeaLevel makes such a board with RS-423; AST doesn't do DMA; Tecmar doesn't appear to have anything fast either. Where are the rest of them? Does anyone on the net have any experience with one? How much did it cost? (just curiousity: cost is virtually no object, compared to the rest of the system it's going into) I saw one from some outfit in Georgia that might fit the bill, but I'd like to specify a stable vendor if it is to be the sole source, and I can't remember their name. Please email me any information you have. If interest is high, I'll summarize to the net. Thanks very much. P.S. This is going to comp.sys.ibm.pc, as well as Sun-Spots. Jim Hickstein, VSAT Systems, San Jose, CA (408) 435-8016 jxh@cup.portal.com ...!sun!portal!cup.portal.com!jxh ------------------------------ End of SUN-Spots Digest ***********************