293 new of 306 responses total.
and, as STeve points out, the big time killer with lookups is when the resolver waits for a failed lookup to time out..
I guess you can't do !w -n|more like on M-net where the -n
does no new lookups. just returns numbers.
tried, w: bad flag -n
Maybe a local fix?
I'm still dropping some text, though
Pine's "-i" flag doesn't seem to be working; it used to take you directly to the screen which lists the contents of your mailbox, but now it doesn't seem to do anything.
hmmm.. perhaps it's being overridden by a default value for the 'initial-keystroke' setting? in any case you can miming the same behavior by going to the <M>ain menu, then <S>etup, <C>onfig, and changing the value for 'initial-keystroke' to "i" (for <I>ndex)
Interesting - didn't know about that. THanks, Mike.
This response has been erased.
If I don't want a list of users *and* their addys, I just type !users. It also keeps me from having to scroll back three screens.
I didn't know about the -i flag. I tried it and it works fine but when I used config to set it I got "bad initial keystroke (missing comma?)". What am I doing wrong?
Thanks for the fix, Valerie.
I tried to abort (cancel) a message that I was writing using mail, by hitting control-C twice, and what happened was that ther appeared on Well, the backspace is not working for me, but that may be something else related to Jim's putting a mini version of z zmodem on instead of procomm The problem in mail, even before he changed to zmodem was: ^C@ ^C@ appeared and there was no way to abort the message. This happened twice already with the Ctl Ctl-C. Also, it took 25 sec to ket get a response to kermit -r, and a minute and a half for zmodem sz, which I think used to come up quite a bit faster. Please excuse any other odd things that Jim's zmodem did here, I cannot get pico to work eiether and hope you don't see what Isee. Let's see if I can abort this response now. Well, I could have aborted the responses here, but not in mail.
My last two sessions have featured grex dropping massive amounts of text in unprecedented fashion. Alas, I believe that someone will actually have to witness how bad it is to truly understand my difficulties.
Re Item #24: I recently encountered the problem with windows 3.1 (on a lousy PC) that it will not pass several pine command codes but instead interprets them as windows commands. Is this what you are seeing?
Procmail appears to be broken. If any of you are using procmail, have someone send you a message to see if it bounces. Procmail seems to expect your mailbox to be in the old spot, and die horribly when it isn't.
hence the message in the motd about having to change your procmail setup.. I suspect that the suspicion on the parts of staff members was that anyone using procmail would probably be knowledgable enough to make the changes themselves but if you really need help I can probably provide it..
Senna, I can probably check out your problem sometime this weekend...
re #26/24, I am using DOS, not Windows. Still can't abort a MAIL message. And Kermit is still taking about 25 sec to come up.
This response has been erased.
kermit should use the "tty" program to determine your tty, not who
When I'm in the 'mail' mailer and respond to a piece of mail I've gotten, a copy of the response gets sent back to me, as well as to the indended recipient. (yes, I'm using 'r' and not 'R') Any idea what's causing this?
I saw some mail on the staff mailing list that indicated that the extra mail problem was due to a global config file that had not been copied over during the changeover. If that is true, it should be fixed now. Is it?
Grace (gracel) reports that Grex hung on her yesterday, somewhere around 2 PM. She read some conference items, ran elm, and went to email someone. It took her into vi as usual. She got into insert mode, & it stopped responding. She finally disconnected. She says that she waited/tried for maybe 5 minutes before giving up. The session was still visible when I looked at the screen last night; I can attest that there was nothing like a timeout warning or anything like that, & she says it was still connected until she disconnected. She has vi configured so that it displays the mode when you're in a text-entry mode, so she is not likely to have been wrong about that.
Could someone please link this item to helpers? I had to join agora just to enter a system problem, & I prefer to maintain my non-involved status.
re kermit: It can't use tty because the tty you're on has no corelation to where you come in from. kermit needs to know if you are in from groupie, or if you're in over the link. Tty can't tell it that.
Dave, is there any chance that she hit ^S, such that things were waiting for a ^Q? In general its a good idea to try that when things appear dead.
re #24 (?). Kermit now comes up in 3 sec, and zmodem in about 2. ~x works to abort a response in MAIL. Ctl-C still does not work there.
Re #24 and #39 - mail does that because you told it to -- what you report is *exactly* the purpose of the "set ignore" that you have in .mailrc.
Sindi, did you run change and do your keys? I had to do that twice plus Valerie remembered an extra command that sometimes is needed. Valerie, could that be her problem?
I've linked this from agora to helpers.
If I recall correctly (and judging from traces of Valerie in my file) I have copied Valerie's file in order not to get the long form of where my mail messages were coming from. In that file is the command set ignore. Valerie, would you try out mail and see if you also cannot use Ctl-C to abort? If you know the fix, let me know it, I am afraid I cannot understand the previous two responses in full. The Ctl-C command worked until the new Sun changeover. It works everywhere else.
Putting a # in front of set ignore in .mailrc fixed the problem, I can now do Ctl-C to abort a message using MAIL.
That would be extremely helpful, Scott, but I'm afraid my next two weekends are absolutely jammed with theater work. (This one included, that is)
Re resp:44: The # turned the line into a comment. I imagine that removing the "set ignore" line altogether would also have worked.
According to Valerie's comments on set ignore, this was supposed to make the system ignore line noise. Before the changeover, Ctl-C was not interpreted as line noise, now it seems to be. Why?
"mail" on the old & new systems is identical. Both would treat "set ignore" as meaning the same thing - to ignore the "interrupt" character.
Valerie, if you still remember the commands that tell "mail" *not* to display those huge mail-routing headers, could you post them in this item? Thanx!!
Valerie gave me instructions how to copy her .mailrc file, where they are all listed with comments as to what they do. (Be sure if you simply use her file to get rid of set ignore).
(You want the lines that say "discard XXX". You don't want the line at the end that says "alternates valerie@cyberspace.org", and you may or may not have a preference about the other things Valerie does, such as "set ignore".)
Just now is the second time today that I typed Pine Return, got the Pine screen, and was immediately disconnected before doing anything else. (Last time I might have hit i first.)
And when I exited Pine this time, the top 2/3 of the screen continued to show the pine menu, and the bbs scrolled by underneath it. (Cured this by giving the clearing command in my commun. program (Ctl-C) followed by Ctl-L). I don't recall this happening before.
I still can't read the top two thirds of the bbs response, above 'Item 3 System problems. Previous commands and Browse (item list), Read (new items) and on there. SOmething is wrong. This is new. But it could have something to do with my having switched from Procomm to Procomm Plus today. (This did not happen the last few times i logged on today, thought). Will try existing bbs and then grex to fix the problem. Odd.
Hanging up and redialing fixed the problem. (I meant 'exiting' grex. and 'are on there'. Hard to type with stuff scrolling away like that).
It very much sounds to me like you've had some terminal problems, such that addressing the cursor around the screen didn't work right. Now, which end that was (yours or grex's) I'm not sure. Have you changed anything on your side of things? Is this happening often? How long? What kind of system and what terminal software are you using?
Problem with BackTalk: From the list of conferences if "Do It Yourself" is selected, BackTalk reports that there is no such conference.
Do-it-Yourself or DIY, must be a single word without spaces. Or Fixit. Now that your 'backtalk problem' is fixed, please join us at DIY. My scrolling problem occurred after changing fromProcomm to Procomm Plus. COuld it be related to the setting term dialup? Happened only yesterday, and only once.
No, that seems to be a backtalk bug. It doesn't seem to like dashes in conference names. Need to investigate that.
Thanx for the mail instructions!
re 56 or so, I had a similair problem when exiting lynx and trying to use the rest of GREX on some occassions in the past. I would get a short screen effectivly.
Minor problem or major for the ladies..but a user not on my .yeswrite file was able to send me a tel...twice...and I hadn't sent her one prior to it. I DID send the userid then she was on before a tel though.. so..my question is..when you exit Grex and get to the login screen again does grex then remember that the prior user was ok to get past the .yeswrite file? Am I making sense?
did you have mesg ne set?
The answer is yes. Grex doesn't know or care who is on the other end of a tel you send, it just marks that tty as being able to send to you for x amount of time. If you tel someone on ttyh4, for example, and then they log off and someone else logs on to ttyh4, they can tel you until the timeout runs out.
Yes mesg -ne is set and always has been. Dang do you know what you are talking about? Thought not.
Yes, I do know what I'm talking about. Since you ask.
Backtalk looks kind of scary this morning, since hvcn.org seems either down or exceedingly slow. The icons are (mostly) not showing up.
Re 65: olddraco, can *you* read? scg, not dang, asked about mesg -ne. But before you ask, *he* knows what he's talking about, and that is *definitely* the first question to ask if someone has a problem with this.
Steve's question is sensible, Dan's explanation is correct. Fixing this in write would add a lot more complexity to an already amazingly complex algorithm to decide if user A can write user B. It's pretty rare for it to be a problem, and it can be a problem for at most four minutes. Steve Weiss had rebooted HVCN, so Backtalk images should be zipping again. I'm wondering if we should relax Grex's ban on images enough to put Backtalk button images on Grex itself.
Today when I dialed in to check my mail with pine, and hit i, I got inbox opened with 18 then a division sign and I was disconnected. Seems to be a bit crowded at the moment, as when I redialed I had to wait for a free conference (conference? does this have two meanings?).
Typical. .yeswrite files and its opposite .nowrite don't pay attention to people you haven't already sent a tel to. I didn't send a tel to the user therefore the .yeswrite file shouldn't have passed it. But also typical is the arrogance of the staff. I will remember and not make the same mistake twice.
Backtalk is working again. The backtalk images are small, and wouldn't be much of a space or bandwidth problem for Grex. I occasionally see annoying delays caused by having the images on hvcn.org. If Grex is down, then it's down and neither images nor text will load. If Grex is up, but hvcn.org is down, slow or having problems, then there are annoying little problems. I don't see a downside to having the images load from Grex, instead of hvcn.org. It would be an improvement for Backtalk users.
Which response are you unhappy with? Steve's double-checking that you really had set "mesg ne"? Lots of people make the mistake of thinking that creating a .nowrite file alone is all that is needed. It's not even a stupid mistake. It's certainly the most common reason for problems like this. Or are you unhappy with my saying that I didn't plan to fix the bug? There are probably hundreds of things on my list of things to do for Grex. I have to prioritize them. This one requires about a days work, fixes a rather small bug, slows all write connects down noticably, and doesn't strike me as any fun. That puts it pretty far down the priority list. Far enough so that the only way it is ever going to happen is if I'm working on some other bug or adding some other feature into the same part of the program. I don't anticipate anything like that. There probably aren't going to be many changes to write in the near future. The only reasonably high priority thing on the To Do list for write is integrating then .nowrite logic and such like stuff into talk/ntalk. That's likely to be a real nightmare, and isn't likely to happen any time soon. My priorities right now are 501(c)3 paperwork and Backtalk. What kind of non-arrogant response would you like to see from us?
Jep slipped in. One thing I'd like to do someday is let users set their own image host. That way they could download a package of images onto their own machines, unpack it into a directory on their machine (say /btimage) and set the image host to "file:/btimage". That way all button images would be stored locally and would have to be fetched across the net at all. The only catch is that it becomes hard for me to chagne the images - everyone has to download a new set.
Has the "more" program been changed recently? It no longer works on multiple files - only the last one is more'd. What should happen is that if there are multiple files in the file spec, each one is preceded by its name, and more pauses at the end of each file. At least that's how it works on other UNIX systems. I thought it worked that way on grex too. Anyone know what's up? The current more is pretty useless, worse than cat... :-(
This response has been erased.
This response has been erased.
Jan, is there some appropriate forum for talking about Backtalk issues? I haven't seen anything in the conferences I read, but I'd be willing to add another conference to my list. From time to time I have ideas, which I'd be happy to post somewhere. In the past I've e-mailed you and srw, or (more often) forgotten them.
This response has been erased.
(I think the Backtalk conference is fairly appropriate. I asked a question there once, and received an answer. it's also fairly low traffic, from what I could tell.)
Just now, in the middle of an entry, i was bumped to the login screen. After logging in twice i was dumped. Any ideas why.
The question mark doesn't work properly while entering agora: Waiting for a free conference (? for help) ...5618 ...5617 ...5616 ...5615 ?...5614
This time I got into Pine okay, including the inbox, but two times out of three as soon as I hit Ctl-X to send a message, I was disconnected. ANd the other tiem (second of three) I was only allowed to write three lines and then got the message a:/write failed, file system is full. I noticed that I had 20 messages (most just arrived). Could that be creating a problem? When I try to answer them I get dumped. Or is the April fools programming spilling over into odd places?
Where have all the flowers gone? Make that three times out of four, this time was when I pressed y to continued an interruped composed message. Do I have to delete all my messages before answering them?
This time I was reading and deleting messages, and in the middle of reading the third one, I got disconnected. If this is a joke, please desist.
This time I got disconnected as soon as I hit C for compose, and I am down to 9 messagesin the mail file, but one of them is 14 K and binary. WOuld this cause all these problems? It was working earlier with the same file there.
I got what seemed to be very bad line noise calling -3000 earlier (the connection died before i could log in). No problem calling back.
This response has been erased.
Grex's net connection is experiencing really bad packet loss. It stinks. We hope this will get fixed soon, but it probably isn't anything with our equipment.
I was going to say. Much better on a dialin. Is the mail getting in and out?
Pine is working perfectly now, I think. But who, after listing the first five or six names, disconnected me. I logged on again and tried who again after a half hour, and the same thing happened. From the grex prompt. Is there some way to reread the first half of the login screen, which is extra long today? If not, could the newer information be put at the end? I already knew about the t-shirts, but the ISDN stuff went by too fast.
Pine is working, but with something broken in the ISDN link, and packets being dropped, mail transmission would involve a lot of repeats, which might cause problems. Logged in now on a dialin everything seems hunky-dory, but it was pretty bad when connected from the net. I was unable to do much as things hung, and then I got timed out.
Re #91: Type "!more /etc/motd" to re-display the login screen. (motd = Message Of The Day)
I've been on the phone with Ameritech several times today about they problem. They're still working on it. At the moment, the line seems to be generating one CRC error every minute or so, which is much better than it was, but still not good. We should not be getting any CRC errors.
re 93, re 91: !motd works on both Grex and M-net to redisplay the message of the day.
I dialed in on 5041 three times getting the "it may take a few minutes for a connection" then within 10 seconds, getting disconnected. The fourth time I tried I used 3000 and have had no problems.
One of the terminal server ports was configured to try to connect to the wrong place. It's fixed.
With the new terminal server and current crop of modems, there is no longer any point of dialing -5041 when things are working correctly. You should be able to connect at any speed up to 14,400 on any dial in number.
Just dialed into -3000, connection okay for a few minutes, then a couple ~10 second "no key echo" pauses, then a longer one (when i noticed my modem signaling "not receiving good carrier/retraining"), then disconnected.
idle killer seems to be busted
System performance/response is horrible right now, 1pm Mon 6 April. Even character echo sux...
Weird, I'm in on a dialin line, same time, and everything is zippy. Karma, my son.
Merit had some severe problems this morning, and seems to be having some less severe problems yet this afternoon. Maybe that's affecting all who are logging in through the Internet.
See item #2, resp #93.
Grex is not connected through Merit. Grex's ISDN line is still screwed up. We are still trying to get Ameritech to resolve the problem.
Connection seems OK right now.
I'm doing Backtalk via Michnet, and everything seems pretty fast.
Karma.
merit is having problems with mci sonet gear last i heard. it started about 5am this morning.
/a: write failed, file system is full Got error 28 (No space left on device) in writing participation file
As of 6:56 AM /a has 49 MB free ... not really an awful lot, but *something* got cleaned up.
I heard MCI was having a lot of problems yesterday due to daylight savings time, of all things. Amazing.
Maybe it was just a warmup for the year 2000.
I got the same error message that 110 got. Tried again, same result. I exited the item, came back in and it was fine.
Re #110,#111. A vandal filled up /a. Appropriate messages have been sent to his ISP's.
I contacted MCI, they're not aware of any issues related to time, etc..
I don't know if this is a problem, but I just now had three pieces of mail dated april 3 show up in my spool. they're all from m-net....
m-net had a mail problem.
I just got one piece of mail forwarded from m-net that someone told me had been sent there a few days ago.
I'm still getting the got error, but it occurs immediately upon attempting to paste-and-clip a block of text (code, actually) while trying to enter an item.
I had to log out and in to get here. WHile writing a response (in DIY, I think it was while wrapping to the second line) I got: /a: write failed, file system is full I think I next type q (not sure) and got Got error 28 (No space left on device) in writing participation file. Leave called twice -- try LEAVE cmd I tried to join agora, to the error 28 and LEAVE messages again. Finally logged out, seems to be okay now.
When I logged in again: Bad participation file. Now all conferences and all responses are new. I hope this is not a serious problem.
I tried a few other conferences. DIY nothing new. Kitchen 154 new. Cars and consumers 1 new each (these are possibly correct, Kitchen is not). Joined agora three times - 50 new, 17 new, 50 new. (Was 51 new before entering #121 here). I will let you know of any other surprises. Let's hope the only messup is my very own participation file, which I can live with.
This isn't really a problem, but with Grex's speed increase, the motd flies by too fast to read all of it. Perhaps a "press a key" option can be added to the motd somehow?
This response has been erased.
I use page up or down to see the motd.
If I do PgUp or PgDn, it is a Procomm command to send or receive files.
Alt-F6 redisplays lines that have scrolled off the screen. PgUp and PgDwn work in this mode. Home and End also work. There is a find function. However you cannot enter new text in this mode, and have to Esc back to do so.
This response has been erased.
I have finally gotten Zmodem working with Procomm Plus, so will say no thanks on MS-Kermit. Jim can figure out how to scroll for me some day, in the meantime we are making people grex e-mail computers. (His sister will try hers out for a while before joining.)
If I recall, the scrollback buffer in DOS Procomm is invoked by alt-F6. I might be wrong. I think alt-Z may bring up a menu (actually just a help screen).
This response has been erased.
In my old version Procomm (2.4.2?) Alt-F6 advances the screen one line, and you have to hit esc to go back to something usable. Alt-F4 puts you back into DOS, and Alt-F4 from DOS puts you back to Procomm. I am about to switch to a 'new' computer and will wait to switch to Procomm Plus at the same time, as it is in an inconvenient directory on this computer. I don't need to scroll back, there are ways to get around it with several steps, but thanks.
All the versions of Procomm I've used had Alt-F6 put the screen into scrollback mode, where the page and arrow keys were used to scroll the entire (pitifully small) scrollback buffer.
/a is very close (13K) to being full again...
This response has been erased.
re 133: Right, hit alt-f6 and the image on the screen moves up one line as it goes into scroll back mode. Then you can use the arrow keys or page up and page down keys to move around.
The scrollback buffer in the DOS versions is however pretty primitive - if you correct a typo, or if you use something that repaints portions of the screen, you get it all there in the order in which it originally appeared, when you scroll back.
Scrollback does work fine with the up and down arrows, thanks. (I will go clean out a lot of dowsing related stuff from my home directory now.)
Running out of disk space like that is a sign that maybe we should have pointed nwewuser at the new /c partition sooner. Not that any of that would help if people are going to pig out on disk space.
Could a disk space limit be implemented automatically? I gather the policy (?) is pretty liberal, so temporary overruns could be allowed, etc. I have gotten personal requests to clean out my directory, but doing that requires more staff time - and is it effective?
There is software that would impose limits on how much people could have in their directories. The problem is that quota software is pretty resource intensive in terms of CPU, and the excess CPU power to handle it would cost more than the disk space it would save us.
I was thinking more in terms of automatic notification - something that runs in .login, which checks du (or whatever) and sends a note if over the quota (and perhaps appends something to a file for staff to watch). Would this be resource intensive?
Well, we do get quite a few logins every day.
Granted, but .login is doing *lots* of other things. How much more would this 'check' increase the load (sending the message will die off as more people adhere to their limits, so that will not be a major load)?
ummm, diff topic... when i come in via telnet, now and then, i runa split screen from netscape navigator which is about 40 cols wide and 34 rows deep. from the csh i set stty rows and cols but nothing seems to pay much attention to that.... like maybe *nothing* seems to pay attention... sure would like this to work as advertised.
the world revolves around 80 column screens. you can conform.
...can i be assimilated too? would that be fair to the assimilee?
This response has been erased.
try a horizontal split screen instead. i have used that technique with some success. i remember having multiple phone sessions ten years ago with eight or nine screen splits with different conversations in each one.
Last night, and again today, when attempting to reply to an e-mail, after about 10 lines everything froze up, no keys worked. I rebooted and my modem would not dial. I turned the computer on and off and it works again. I had sent one e-mail before replying to any. Has this happened to anyone else? If not, it is related to our switching hardware and software (a later version of 4DOS as an operating system, and an expanded memory card and shrom software). Any ideas how to attack this problem? It may have something to do with shelling via the shrom program. I am not shelled now and will check e-mail replying again and report back if it still does not work.
Apparently I was not disconnected, which is why rebooting did not work. The e-mail worked fine now without the shell program involved.
I logged in to grex just now without using the shell program and had the same problem with e-mail freezing up (while I was replying to a webmaster froma website using lynx). This was the second e-mail I sent, the first worked okay. Actually, the first three worked okay, and they were all replies.
I've been told some e's sent me have been bounced back. Is it in the settings or a glitch?
This response has been erased.
Grex cannot be reached from Michnet. Party has been dead since about 07:30, which is usually a good sign that the net link is down. But there still seem to be some users telnetted in from far away places. ???
This response has been erased.
For whatever it may be worth, here is a traceroute from Michigan State: 2 serv3 /home/serv2/krj # traceroute grex.cyberspace.org traceroute to grex.cyberspace.org (204.212.46.130), 30 hops max, 40 byte packets 1 cc2-gw-fa41.msu.edu (35.9.2.1) 2 ms 2 ms 2 ms 2 fdd0-0.msu2.mich.net (35.9.6.3) 8 ms 5 ms 5 ms 3 fdd0-0.msu2.mich.net (35.9.6.3) 5 ms 8 ms 6 ms 4 hssi2-0.michnet1.mich.net (198.108.22.101) 8 ms um-fddi4-0.ann-arbor.cic.net (192.203.195.3) 6 ms 6 ms 5 border1-fa4-0-0-2.arb.qual.net (198.87.18.1) 6 ms 8 ms 6 ms 6 blackrose-gw.arb.qual.net (198.87.206.42) 9 ms !H 9 ms !H 8 ms !H 3 serv3 /home/serv2/krj # I don't know what "qual.net" is.
re #157 msen comes in via a different path than michnet re #158 qual.net is what became of cic.net+iagnet.net+cyberdrive.net+ greatplains.net there was an as seg issue between some hops in the network, i've stoped the router at hop #6 so it won't go on until it becomes fixed.
Looks like it is fixed now, I am connected via Michnet.
I am no longer having any problems with the screen freezing up while sending e-mail, and I have not changed anything at this end.
Actually, once in a whiile I cannot get past a certain spot in a line but I can go to the next line and resume typings. I have noticed that when using MAIL, I can type n to get the next message, except when I know I have just one more message left, and type n, it tells me EOF, and I have to do h for a list of headers and then enteer the message number in order to read it. Am I doing something wrong or is it mail?
usually where there's an "n" there's a "p"..
The getting to a certain point in the line and then having to do a hard return to continue typing usually means that you've been forgetting to press Return at the end of every line and that you've filled up your buffer. Happens to me all the time. ;)
telnetted in using grex.org and had sopme severe weirdness, including but not limited to... password echo, incomplete prompts which could be 'completed' pressing the space bar for each additional character i wanted to display. grex.cyb was all that appeared automagically, the rest ofthe csh prompt came a chaaracter at a time, per space-bar-hit, until the % arrived.. then typing got nowhere but several returns later, the command finally echoed...and failed.... killed the connection, reconeected to cyberspace.org and things were fine. <wtf?> re: previous problem... ok, it's an 80 col world, yippieyippiefsck! the system problem is : why won't grex read&conform to the stty settings? stty -a give all the correct settings.. adn grex ignores them magnificently!
..while i am here.... when and WHY was my .plan file permissions not only CHANGED. but also assigned to cfadmg group? i sure as hello didnt ask for that, either of the two "that's."
My impression is that the .plan change happened to quite a few people.
I tried to connect to www.cyberspace.org with my lynx and was unable to access. I tried pinging cyberspace.org and it worked, so I used cyberspace.org in my browser and it worked. Missing A record?
Hmm... All Grex's name servers seem to be responding correctly to queries for both the grex.org and cyberspace.org domains right now. TS's problem doesn't look at all DNS related. The name server is going from the same zone file for both domains, so they are identical. I'm guessing it was some other intermittant problem, and that it working with one hostname and not the other was coincidental.
If your .plan is owned by group "cfadmg" then that is because someone (presumably you) editted it via Backtalk. Nobody is in group "cfadmg" except the Backtalk program itself, and Backtalk only lets people edit files if they have logged in as the owners. The same thing happens with your conference participation files. It is all a bit odd, but has to be done to be compatible with Picospan. You are welcome, of course, to change to permissions/group on your .plan back to anything you like, but I can't see why you'd want to do so. If you edit it again with Backtalk, Backtalk will reset the permissions it needs (it has extremely limited root access to enable it to get access to participation files created by Picospan by putting them into the cfadmg group).
(ah! I'd wondered about the cfadmg group. thanks for the explanation, Jan!)
Was grex rebooted this afternoon? I haven't seen anything in the MOTD explaining why it was down.
Ok: !uptime 5:13pm up 2 days, 23:36, 57 users, load average: 81.76, 74.14, 71.93 What's up with that?!!!
Marcus changed some settings that makes us more resistant to certain denial-of-service attacks. What you saw there might have crashed us before.
This response has been erased.
Grex Board Meeting Wed May 22 - Jan, is this supposed to be May 27? (See motd).
(Well, the motd now says "May 27".)
It is the 27. Thanks to whomever fixed it.
My screen keeps freezing up when I am typing something in Pine (yesterday)
or in a conference (just now). The phone line stays connected, but I have
to reboot, which takes 5-10 minutes (because Jim set me up with an expanded
memory manager, a virtual drive, a compressed hard disk, and probably a few
other clever tricks), then hang up the phone line with Procomm before I can
dial to get back on to grex. Is there some way to at least not have to
disconnect and reconnect the phone line?
This one cannot be related to line wrap, as Misti suggested,since I
am not using mail and there is automatic line wrap (I think). Am I doing
something wrong, or is the freeze-up just normal?
Probably your setup, especially if recently changed. I've never had that problem, although I never use Pine, either. Conference freezes never happen.
What flow control are you using (that is, what flow control is Procomm using)? <Dave tries to remember what was available in DOS Procomm>
Hardware flow control, Scott fixed that for me. I just had it freeze up when trying to download a file, too. Could it possibly have something to do with the shell room (shroom) program, or the virtual disk, or the expanded memory card and software? My other question was is there some way to get back to grex without disconnecting and reconnecting.
Actually, if your modem hasn't hung up, you *may* be able to just proceed. Make sure Procomm is set to the appropriate settings. (I used to do this, when I used DOS Procomm.)
I can't 'just proceed', as there is nothing on my screen resembling grex after I reboot. What do I do next? What do you mean by 'appropriate settings'?
'Proceed' by pressing the Enter key to see if you have a connection. If entering a conference response, try using Control-L to redraw the screen.
On my computer (MacOS) the comm program will think the serial modem port is occupied, if I reboot without shutting down (i.e. resetting) the modem. This is because the comm program does not overrule the serial interface extension when it is doing its thing.
I tried Ctl-L a few times with no luck, probably also tried Enter. I will try them both again next time this happens. I don't have a Mac and have no idea how to reset the modem. Thanks, this is not serious, just annoying. (I am still hoping a helper will explain to me how to decode binary files that were sent to me as attachments, that is more urgent. It seems I have to know whether unix decodes mime or uuencode so they know what to send me, as well as how to decode the file after saving the attachment. HELP! I will soon have a large binary file coming my way that I won't be able to read.)
You can interrupt the connection by turning off the modem. Of course, you don't keep the connection that way. You donwnload the binary file and open it in the application for which it was written. Saving the file in your directory decodes the MIME. You have do uudecode, or whatever is required. Usually people just send me documents in the original binary form (MS_WORD, e.g.) and once I save them and download them, they open fine.
It worked to hit Ctl-L from Procomm (not from DOS). Obviously this document was not sent to me in MIME, I will ask them to resend it in MIME. I assume PINE automatically codes things in MIME. How do I decode uuencode, in case I ever need to? Typing uudecode followed by the filename did not work.
I downloaded the file which I think was sent to me MIMEd, and when I read it in WP, instead of gibberish consisting of lower-level ASCII characters (7-bit) this time I also got the upper level characters (8-bit). I suggested that they resend the file in binary form without encoding it, but they have probably gone home for the day by now.
I found that instead of rebooting I can just exit Procomm, disconnect (it did not work to do Ctl-L the last two times it froze up, once while typing in Pine, once in Pico in a conference response), and redial. This saves about 5 minutes, and I can read a book for a few minutes while redialing. One time it froze up when I was doing right arrow to the end of a line (buffer problem) a few minutes ago, in Pico editor editing a response of mine. I received another MIME copy of a file, which they suggested might work better since they were sending just one attachment instead of three at a time. Time to download it and try again to read it.
The single attachment gave the same upper-ASCII file as the three attachments did (when I saved them individually). MIME files give upper ASCII, uuencode lower ASCII only. Saving to my home directory (V for View, S for Save) does not seem to automatically decode files. Am I autmoatically encoding when I send binary files with PINE as attachments? I asked them again to send an uncoded binary file.
The freezeups generall seem to occur about one third of the way through a line, this time on the second line of a paragraph of response to an item. I have not had this problem in this response so far, despite not using a hard return at the ends of lines. Any ideas what causes the problem? It is at least three times today, maybe four.
Has anyone else had people apparently attempting to log in to their accounts? (There were a few failures listed when I logged in today, and they don't make sense as far as my time constraints are concerned...)
The MIME encoding and decoding is automatic in pine. Did you specify a *binary* download? Do you have the same application (and same or newer version) that wrote the file they sent? This can be a very frustrating exercise. I am trying to help someone send a file to me so that I can open it, but they haven't yet told me what the application and communications programs are. They may be using a different protocol for the attachment, or encoding it in some fashion they haven't told you, etc. You have to know *exactly* what they did before you can solve these file transfer problems (except by hit or miss or by good luck).
re 194: yes. i apparently had someone try to login to my account while i was logged in a couple days ago...
Re 195, they never even exactly told me what coding program they used the first time. The second time they told me that they could also send it in Mime format or hex something or other as well as uuencode (which led me to believe they used uuencode the first time and mime the second). I typed set file type binary when using Kermit to download. What looks like all letters in grex looks like upper ASCII in WP5.1. I sent them WP5.1, it got edited at their end, they did not tell me if it remained in WP5.1, though I asked twice about this. I will write again and ask for specific answers. I would hope the communications programs do not have to match. I doubt that the person who sent me the files knows what was done to them first. We will experiment by sending a WP5.1 binary file from one of us to the other.
This time my screen froze up at the very end of a line, on about my fourth (possibly fifth) response in Agora, end of the third line. I had also read two e-mail messages and responded to one first. I was letting line wrap work in the response. From now on I may do hard returns and see if that helps any. I cannot get back to grex while still connected on the phone by typing enter or Ctl-L. (It worked once, for some reason, but usually does not).
If the file you received is all letters, it is MIME or uuencode or binhex (others?). Pine decodes MIME automatically (it is "MIME compliant" - the free e-mail service "juno" is not, by the way - you can't send someone on juno an attachment from pine and have it decoded for them). You can tell if it is uuencoded by the first lines of the file, which reads like "begin 644 newname". You uuencode a file with the command uuencode newname <file> uuname.uu newname is the name the file will have when it is decoded, file is the file to be encoded (the brackets must be written in the command), uuname.uu is the name to give the uuencoded file (it is convention to add the .uu to identify the file type). You decode it with uudecode uuname.uu and the file newname will be created with the contents of the original file. (The instructions in man uuencode are not correct.)
I just printed out your instructions and will try them on the first three files that were sent to me. Thanks. I always assume it is me rather than the instructions which have the problem.
Rane, I am afraid I did not understand your instructions. The file sent to me, and saved from my e-mail attachment to my home directory, is called (by the sender, and I did not change the name) Ge9495a.cec. I tried uudecode Ge9495a.cec and got 'No 'end' line. I tried uudecode Ge9495a.uu, Ge9495a.cec.uu, and uuGe9495a.cec and got no such file or directory. Do I need more info from the sender, or am I just completely off track in following your instructions? Hopefully this file is accessible to you as well, do you want to give it a try decoding it? Or let me know what to do to make it accessible? I also saved what I am hoping are mimed files, one of which is mimea (the same file), and it did not automatically decode (I checked in WP5.1).
BTW, MIME is really just a scheme for telling your mailer what kinds of documents are attached and where they begin and end -- there's isn't a specific "MIME" representation scheme for turning binary files into text for transmission.. MIME-compliant mailers typically use uuencode, base64, or (less commonly) binhex to do that..
The uuencode file started with Begin 600 filename. The mimed file, when downloaded, was readable in DOS but not in WP5.1, where it shows up as upper level ASCII. Jim will see if he can remove some of the stuff at the beginning to make it readable. I will e-mail the sender and ask if they sent WP5.1 or perhaps Word for Windows instead.
I downloaded Ge9495a.cec to my machine, uudecoded it there, and it opens as a WORD document. It is quite readable. I don't know why it did not decode on grex. Maybe their uudecode is obsolete? Would you like me to e-mail it to you as plain text?
Rane, I cannot read WORD text. They just assured me that it was sent as WP5.1, and that they have to encode by one of three systems when sending an attachment, or else add it to the body of the text. Jim says their file starts with WPCG, and my WP5.1 files start with WPC7. I will write again and ask if they might possibly have sent me a WORD file by error. I still don't know how to use uudecode. COuld you possibly take the time to see if you can get uudecode working on grex, and if it is obsolete maybe someone could upgrade it? Jim tried to convert the demimed file from Word 4.0 to WP5.1 and this did not work. What version of Word did you get? Thanks again.
As far as I know, the uuencode and uudecode commands on Grex are working. According to the man page, Grex's uuencode even claims to support base64 encoding. Re #202 (technical aside): uuencode isn't part of the MIME standard, so a MIME-compliant mail user agent isn't required to support it (although some do, I'm sure).
The person who sent me the files claims they are in WP5.1, that this is the only wordprocessor that particular editor owns, and that she just opened them in WP5.1 to check. Rane, did you also try to read the file in WP5.1? Is it possible that by opening it in WORD, WORD automatically converted from WP5.1? Does anyone who has WP5.1 want to try reading my (hopefully decoded) file mimea, which reads okay in DOS but is gibberish when I look at it in WP5.1 after donwloading it as a binary file? If Jim ever recovers enough from his viral infection to wake up, we can try sending the original WP5.1 file to his account, then e-mailing it from there to my account, and downloading to see what it looks like after that.
Progress 1. We uploaded a WP5.1 file to my home directory, then e-mailed it to Jim's account using Pine (automatically encoded as MIME?), then Viewed and Saved it to his home directory, then downloaded (upload and download with Kermit, set file type binary) it to the computer here, and it came out the same as the files someone sent me (mimea is one of them, in my home directory). In DOS it was readable, in WP51. is was lower and extended ASCII characters (8-bit). It does not seem that saving from an attachment to the home directory with V and S automatically converts a file from its MIME coded format. Can someone explain how to decode MIME using grex (and please try it out on my mimea file to be sure it works, or I can send you a WP51 file to your account, encoded using PINE, and you can try to find some way to decode it on grex.) 2. We then used the WP5.1 convert program to convert to 7-bit transfer format, did the same upload, mail, download with it, used the convert program to convert back from 7-bit format to WP5.1, and it worked perfectly. So I can ask people to send me 7-bit transfer format, but it is 50% longer than the original file (maybe all coded files are that much longer) and therefore takes a lot omore of grex's time to download. 3. I would still like to know if there is some way to decode uuencode using grex, and please make sure it works on file Ge9495a.cec, which I am told was sent to me uuencoded. What is a missing 'end' line? I hope it is appropriate to be asking these questions in this item, but there is a chance that it is a system problem. Probably nobody else has ever tried to send binary files via grex, before we joined it was a 7-bit system, and a week after we asked Valerie how to send 8-bit files it became an 8-bit system, for which we are very grateful (and for which reason we joined). I will gladly donate $25 to grex (general purpose operating expenses) in honor of Valerie, Scott, Rane and other highly helpful grexers if someone can teach me how to decode uuencode, and another $25 for decoding mime, using only the resources available in grex. (Valerie responded to our request for some way to send binary files, Scott came over and pointed out that we needed hardware flow control to use anything but Kermit, and Rane has been attempting to fix my decoding problems.) It would be much appreciated ASAP, who knows when I will really need the ability to decode such stuff. (Thanks also to Omni for his extended memory card and Kent for steering us to shroom software, it can be in their honor as well.)
One obscure-but-possibly-related point: to do a binary transfer, kermit needs
*both* ends configured as binary. You're doing "set file type binary" (on
Grex I presume), but is your other end (Procomm?) configured for kermit to
do binary file transfers?
The "end" line is what tells uudecode that you've reached the end of the
(encoded) file. Let me go off & create & uuencode a (small) file as an
example:
$ echo this is a test > asdf
$ uuencode asdf < asdf > asdf1
$ cat asdf1
begin 644 asdf
/=&AI<R!I<R!A('1E<W0*
`
end
The "begin" line says that when the file is uudecoded it's to produce a file
named "asdf" with 644 (rw-r--r--) permissions. The following lines (in this
case, 2 of them) up to but not including the "end" line are the encoded date.
The end tells uudecode that this is the end of the file. You can run
uudecode on a file containing stuff before and after any of them, and
it will use on the begin & end lines to determine where the uuencoded
part begins & ends.
So look at your (saved) message. Does it include a line with "begin"
at the very beginning of the line, having the structure of my example's
"begin" line, followed by even-length lines of (apparent) garbage?
Look at the end of it for a line consisting of "end" at the beginning
of the line.
Jim looked at the uuencode file and instead of 600 the first line read begin 600 filename and the last line read end (all by itself). Is the 600 something that grex's uudecode program cannot deal with? And why does it tell us No 'end' line? The file is 60 characters long. Why was Rane able to uudecode it on his system, but it does not seem to work on grex? Davel, could you look at thefile in question, Ge9496a.cec I think it was, in my home directory, and see if you can decode it while on grex? And if you can, what should I type in other than uudecode filename? Rane said something about grex's uuencode manual instructions being wrong, someone (I am losing track) said they were right, but could someone simply try following the instructions to see if they work, then explain them for an idiot? The other problem remains what to do with the automatically mimed file (it is called mimea in my home directory, they sent it in a different coded format). Thanks to everyone for your time, and apologies for my ignorance, but I always seem to need to learn things in a hurry and don't have time to read up on it.
600 is the file permissions. That means that the owner of the file (you) can read and write to it, but that nobody else can access it. Having a number there is part of the UUencode format. I believe the end at the end is what's supposed to be there.
My WORD (Mac) 6.0 reads WP5 files. Anyway - do you want this file e-mailed to you as text - perfectly readable - or don't you? I don't know why uudecode on grex doesn't work. I used UUTool (Mac) on my machine. The man page for uuencode shows commands that don't work on grex: something is out of wack.
This response has been erased.
OK, I think I see what your problem is. This file came from a DOS source, and has extra carriage-return characters at the ends of all the lines. That's why uudecode isn't recognizing the "end" line - it's "end^M" not just "end". You're also going to have a problem trying to uudecode it, since the name of your file and the name to be given the uudecoded file (from the "begin" line) are the same. It will overwrite before it's done reading, I think, leaving you with a partial file. The simplest way to avoid this is to rename the original before uudecoding. Here are instructions for dealing with this, geared to this particular file. (There's a program, flip, intended for stripping out or adding carriage returns, but it doesn't seem to have survived the transition to the new system. Thus the sed & pipe in what follows.) At a shell prompt: mv -i Ge9495a.cec Ge9495a.dos sed 's/^V^M$//' Ge9495a.dos | uudecode (The "^V^M" means to press control-V followed by control-M. This should actually echo as "^M".) This will create the file Ge9495a.cec, which should be what you want. You can then do a file transfer (binary) on that.
Valerie slipped in. No, redirecting the output won't help; it's going to use the filename from the "begin" line (unless there's some kind of option to let you specify a different name). And redirecting input doesn't matter. And the missing "end" is definitely the carriage-returns at the ends of the lines. I've tested this to make sure it all works, if you follow my instructions.
UUTool on my computer is apparently not bothered by extra ^M, since it decoded the file just fine. There must be an equivalent free utility for a PC. It is even easier to use the command Valerie gave, but chose a different destination file name. I did experiment with uuencode and uudecode here on a file in my directory, and they worked fine. So the problem is that grex's version is flummoxed by carriage returns.
Or, alternatively, that something else is sticking carriage returns in where they don't belong.. Usually you get those pesky extra carriage returns when someone uses a binary transfer mode to move a text file (ever wonder *why* most transfer protocols have a binary mode and a text mode? it's to take care of things like those CR/LF problems..) If you expected to be doing a lot of uudecoding of such files you could just alias your uudecode command so it would do sed with the right arguments piped to uudecode..
Thanks to everyone, I have not tried any of this yet because I am waiting for Jim to wake up. We both still have some sort of virus and Jim's response is to sleep a lot. Rane, thanks, I don't need this particular file as much as I need to know how to receive WP5.1 files via e-mail and decode them, so I will try to produce it on my own, but many thanks for proving that the file was in fact decodable with uudecode. re #211, if 600 means nobody else can access my file, how did Rane manage to decode it? re #213, we are trying one other solution to the problem of freezing up the screen (it happens all over the place, not just in MAIL, and even when I do hard returns every line, and may be worse in warmer weather). We replaced an internal modem, which may be overheating, with an external one, and so far so good. I have had no such problems except while online to grex. After we get uudecode working, it would also be nice to be able to read the MIMEd files, which could conceivably also have a hard return problem. Hopefully with all these hints Jim can figure out that part. I don't understand 'alias' or 'piped', perhaps Jim knows their meaning. If any of you superhelpful people ever need any help reading a Slavic language, please call me first.
The freeze-up problem may have been the modem, as it did not happen in 40 min
or so with the external modem. But then I got disconnected, which I am
guessing had to do with grex being overloaded. I got disconnected a second
time in 2 minutes, so now have a 2400 bps modem hooked up, Jim was convinced
the disconnect was due to the other 14.4 modem.
Instead of the mv command, I could, I think, simply rename the file
to be uudecoded while saving it to my home directory. We will now try to
figure out what sed means, before applying this command.
We looked up mv -i, which means to rename a file and have an interactive promprt, which asks us whether to overwrite any existing file by that name. We looked up sed, with is the stream editor, and the $ addresses the last input file. We have not found '/^V//, but presume this means to delete all those ^M's at the line ends. We are about to try out these commands. See you soon!
The file was permitted 644 in your directory. I think 600 would be the permission of the unix file after it is decoded.
yes, the 600 perms would apply to the *output* of uudecode, not the input.. ^V means "take the next character literally" (i.e. instead of treating it as a special character like a carriage return or backspace..) so s/^V^M$// would "s"ubstitute a literal control-M at the end of the line with nothing, i.e. would delete it..
Well, we followed the directions to use uudecode and delete the control M's,
and, once I figured out that I could not send a file to a directory that
already had a file with that same name (something to do with the Procomm
settings, it will not overwrite but aborts, I deleted the original files
there), we succeeded in sending the uudecoded three files (Ge9495a.cec, and
the b and d versions) to my computer. They can be read in DOS, just like the
automatically decoded MIMEd files, but are still gibberish in WP5.1, in fact
the uudecoded file looks identical to the mime-decoded file. The first line
reads something like:
Nm'hn{^S.)=5<h k^Q,%4#;"0{9',ik*/<2^\.
This is what I get trying to view the file in WP5.1, but if I try to download
I get the message 'incompatible file format'. SO now I have two ways to
decode to the point where I cannot read the file in WP5.1. What next?
What is now in my home directory is the uudecoded versions Ge9495a.cec etc.
and the decoded mimed versions mimea, etc., which seem to be identical.
I don't understand the 600 and 644 stuff, am getting pretty confused by now,
but does anyone still have WP5.1 for DOS, and could you possibly download oneo
f these decoded files and see if it is readable in WP5.1, and tell me where
I am going wrong? MANY, MANY THANKS. If the files are no longer accessible,
tell me how to make them accessible.
The first line of the files viewed in DOS is not actually text, but is stuff
like happy faces, hearts, and other upper ASCII characters, and the command
Prestige 12.
While sending via Procomm, with Superkermit, after typing set file type binary
on a separate line, then return, we were told that the file was being
transmitted as TEXT. Could the problem be with a Procomm setting? We have
had no problem uploading binary files with Kermit. I may try downloading the
same files with zmodem, but so far I have not been able to figure out where
the files go when I use zmodem, they seem to be sending but just disappear.
(Too much to learn all at once - new computer, new wordprocessor, new version
of Procomm, unix, e-mail, grex.... I am having trouble keeping the unix and
dos and wp commmands separate and tend to remove instead of delete, etc.)
When I recoded the file on my computer it appeared as a WORD text document. Are you sure it is from WP5.1? My system usually tells me it is translating a WP file.
We had the same problem when we sent ourselves a known WP5.1 file (upload,
e-mail to Jim, view and save, download), so it cannot be WORD, we don't have
WORD in theh ouse anywhere.
We are attempting to download with zmodem, but when I type sz filename,
give the send command (PgDn), specify parameters (drive name and file name),
the system freezes up and I have to log off and on again. Whatever we are
getting after decoding certainly is not acting like a WP5.1 file at our end.
Jim thinks that Word may automatically convert WP documents to Word.
Well, back to the zmodem manual. We seem to be pushing grex (and a few other
programs) to beyond their normal limits. We are downloading a binary (8-bit)
file, but can't read it in WP5.1, just in DOS. Maybe we have to change the
settings on KERMIT? It is downloading 8-bit, but says it is sending text.
In the middle of this I got a chat request from Turkey, while using
the external 9600 baud modem, and the screen froze up, so it does not seem
to be a modem problem that freezes up the screen.
Davel, did you download the decoded file and then view it with WP5.1
for DOS? If so, how did you download (Procomm, Kermit, or how?).
Jim will set up a second computer the same as this one to see if that
also has the freeze-up problem. It is annoying but tolerable. The inability
to read decoded files may not be tolerable if people are not willing to both
convert their WOrd files to WP for DOS and then also convert to 7-bit transfer
format. I am probably driving quite a lot of translation agencies crazy by
using outdated software, but then there are not a lot of Slovene chemistry
translators around who know English as wellas I do, so they are patient.
We changed the kermit settings from TEXT to BINARY and downloaded the file, and it is now readable in WP5.1. (I had thought the command Set file type binary would have taken care of that. Do I continue to set file type binary? Do I set it binary for text files?). So now I can recognize and decode and read both uuencoded and mimed binary files, as well as sending Mimed files. (I will not worry about learning how to encode with uuencode.) We still don't know how to download with zmodem, but will read the manual for hints. I will now write out a check for $50 to the grex general fund, in honor of DaveL for teaching me how to uudecode, Rane for proving that the text was uuencoded and could be decoded, and everyone else who has helped me to make some sense of unix and other computer stuff. I would include here Marcus for his magnificent explanation of how binary logic works, Mike McNally for trying to explain a few other things (never sure whether I understand or not, but after a while it gets clearer), Jan for his extract program, Valerie for her immediate and detailed answers to a lot of beginner's questions, and anyone else I forgot to mention above. I will probably continue to come up with questions, some challenging, some rather basic (read stupid?), and be grateful for the answers. I wonder why Kermit was uploading files as binary, but downloading as text.
I don't think I had used Kermit to upload anything since we switched from Procomm to Procomm Plus (too many switches all at once -computer, wordprocessor, communications program and protocol - had been uploading with zmodem, and downloading with kermit but until now only text). The check is signed and in a stamped addressed envelope. My screen froze up this time while trying to access a website. Is there some version of uudecode that could be innstalled on grex that would take care of the line-end Control-M's, such as UUtool?
The dos2unix program is available for doing conversions. Use that in combination with uudecode.
Ratheer than try to figure out how to use one more program, I will stick with Davel's one line ctl-M remover. I compared the first of the three files sent me with what I had sent to the translation company. I had asked to see the copy after the chemistry editor finished with it. I put my original file on screen 2 and the edited file on screen 2 and kept switching back and forth every line. I found only about three changes, such as steam for water vapor and a couple of typos. This did not seem worth the work to compare. Now a friend just emailed me a uuencoded zipped program for comparing files. I exported it from the message to my home directory under a slightly different name than what was in the header, and typed in the line with the ctl-M in it, and got out a file of the original name (it may still have the one line text intro, which I hope I can manage to delete) which looks decoded. For some reason, I also got the message: uudecode: stdin: Short file What does this mean? Anyway, I have now been able to use uudecode on another binary file, in this case a zipped program. (Still have to learn to use the program). The check is in the mailbox waiting for the mailman to pick it up. Jim also figured out that you could transfer the printer files from a 3.5 inch floppy to two 5 1/4 floppies by zipping them, and we got me a printer definition working. All that is left to do is to get zmodem to download, figure out why my screen freezes up so often, and it would be nice to have a monitor that can display graphics (Cyrillic alphabet) but first Jim wants to use the Hercules graphics plus card to fix another broken one, so I may get a VGA temporarily instead. The end is sort of in sight, and at least I am prepared to send and receive binary files, which is what I originally joined grex for. Many many thanks again to everyone.
"short file" means the file got truncated. The "uuencode" format works
something like this:
first line: "begin" mode filename
N lines: each containing a non-zero character encoded byte count,
followed by character encoded binary data.
A line with a zero byte count, and no data.
A line that contains the word "end", and nothing else.
Binary values from 0 to 63 are encoded as follows:
0 ` 1 ! 2 " 3 # 4 $ 5 % 6 & 7 '
8 ( 9 ) 10 * 11 + 12 , 13 = 14 . 15 /
16 0 17 1 18 2 19 3 20 4 21 5 22 6 23 7
24 8 25 9 26 : 27 ; 28 < 29 = 30 > 31 ?
32 @ 33 A 34 B 35 C 36 D 37 E 38 F 39 G
40 H 41 I 42 J 43 K 44 L 45 M 46 N 47 O
48 P 49 Q 50 R 51 S 52 T 53 U 54 V 55 W
56 X 57 Y 58 Z 59 [ 60 \ 61 ] 62 ^ 63 ^
Note there aren't any lower-case characters here. When uuencode was written,
there were plenty of systems around that automatically translated text from
lower-case into upper-case.
Since a binary byte contains 8 bits, but the above only encodes 6 bits,
it takes 4 of these printable characters to store the information from 3
bytes. The encoding looks something like this:
Binary data(hex): 4D 44 57
(in binary): 010011010100010001010111
aaaaaabbbbbbccccccdddddd
6-bit numeric values: 19 20 17 23
character-codes: 3 4 1 7
(This was probably not a particularly good example, as it is
"coincidence" that the 4 characters, 3417, all happen to be digits. All
those other characters, including punctuation and letters, are also
valid data.
Each of these N lines must contain at least 1 character of data, but
generally speaking, most lines will contain 45 bytes of data. These 45
bytes will expand out to 60 printable characters, and, with the byte
count, that means most lines in a uuencoded file are 61 characters long.
Each of these lines will also start with M, because M is the character
that encodes the byte value of 45. The line with the zero data count
will contain one printable character, `, because ` is the character that
encodes the zero byte count. If the file isn't a multiple of 45 bytes,
the last data line will contain fewer than 45 bytes. If the file isn't
a multiple of 3 in size, then the last byte of the file doesn't line up
on a nice character boundary. Now, it doesn't *really* matter what's
done here, because uudecode will ignore whatever's put there, but what
uuencode does is to pad the last character triplet out with "garbage"
(actually, data from the previous data line), then write out 4 character
values. So, most lines in a uuencoded file will contain 1 mod 4
characters. Also, because uudecode ignores extra trash on most lines,
there is only 1 line where a trailing control-M is bad -- the *last*
line! So, rather than running the whole file through dos2unix, you can
just use vi (or whatever) to delete the last control-M in the file.
uudecode can complain about the following:
uudecode: <file>: no "begin" line -if the first line
doesn't
start with "begin".
uudecode: <file>: illegal ~user. if the filename starts
with ~, but doesn't
have a /.
uudecode: <file>: no user <user>. if ~so-and-so isn't on
this system.
uudecode: <file>: path too long. # if the expanded
filename is
too long...
uudecode: <file>: <operating-system-error> if there's an error
creating
the file.
uudecode: <file>: short file. EOF before reading a
data line
with a zero byte count
(ie,
the file dones't contain the
line (`)
uudecode: <file>: no "end" line. the next line in the
file
after ` doesn't say
"end".
Sindi,
(1) kermit usually needs to be set to binary on *both* ends - otherwise when
the two ends negotiate they choose the least-common-denominator setting & go
with text, & garbage your file. This is dumb, but is the way it is. So you
need to set Procomm to use binary - which you should be able to do as a
configuration setting & thereafter not need to do again, I think - and also
do "set file type binary" on Grex's kermit. (Or, to just send a file,
invoke it as "kermit -is yourfilename", I think. I don't use kermit much.)
(2) What all Marcus said is (I presume) correct, & is interesting & useful,
besides being way more than I ever knew about uuencode/uudecode. But in
your particular case I think you have one likely cause for the file truncation
which yielded the "short file" error. (I mentioned it earlier, but didn't
stress it enough, I think.) If the filename on the "begin" line in the
uuencoded file is the same file name as the uuencoded file itself - which was
the case with the one I looked at - the file will get overwritten before
uudecode is done reading it. That's why I put in the step to rename the file
to a different name.
(3) I'd recommend using the dos2unix program Kent suggested, instead of the
sed command I gave you. I was going to suggest flip (or its alias toix),
but it apparently got lost in coming to the Sun4, and I didn't know about
the (hopefully equivalent) dos2unix program. You should probably read
the man page on dos2unix (I just did so, & think you probably want the
-ascii option).
Better yet, you could write a script to rename your file, convert it, &
send it all with one command. Assuming that (as with the sample) the name
your file was saved to is always the same as the name on the "begin" line,
the following should work:
#!/bin/sh
# for each parameter, renames & uudecodes. Then sends the entire batch
# using kermit
for filen in "$@"
do
echo Going to convert ${filen} now
mv -i "${filen}" "${filen}.dos"
dos2unix -ascii "${filen}.dos" | uudecode
done
echo
echo "... Going to send the following files, kermit binary mode:"
echo "$@"
kermit -is "$@"
(You would still need to make sure that Procomm is set for binary kermit
first. Also, I'm pretty sure that kermit would send all the files on the
command line, but I haven't tried it to make sure.)
Thanks Dave, I did have to change the Procomm settings for Kermit from text
to binary before I got readable flie (I think I mentioned this, but I dont
write all that clearly), and I had saved the file from my e-mail message to
a new file name instead of the one it decoded as, rather than using your
mv command. What I did not do was to remove the first line of my e-mail
message saying something like : Here is a file I am sending you. This may
have confused uudecode. I will edit the message and reexport it (I had better
delete it from my home directory first). We will do that and see if we get
the short file message.
Is there any problem with uudecoding a file 140,000 bytes long?
We will also attempt to look at the manual and learn to write our
own unix programs, some day, or at least enough to understand what you wrote
for us above. (First we are attempting to figure out why my screen freezes
up by putting me back on my old commputer, which I bought new in 1985 and
which is known good, with the same new software, minus the shroom program,
and if it all works, add the shroom program and if that works it was a flaky
computer and I will use another one, we have a large choice of them).
Marcus, as always, many thanks for the detailed explanation (which must have
taken you at least an hour to write up typo-free, even if you knew this all),
which we will extract and print out for further reference. I vote that we
start a separate archive item for this explanation and for a few of Marcus's
other contributions, including the one on binary logic. (In fact I would be
happy to save all that Marcus writes, Marcus not only knows a lot about all
sorts of things but writes about them so even a beginner can understand, but
others probably don't find it boring).
Unfortunately we do not have time to read and understand all the above
right now, as we are supposed to be at Kiwanis at 9:00 fixing stuff.
The inability to download with zmodem (it freezes up the screen) may
be related to the other freezeup problem, we will try it on this computer.
WOuld it be possible to at least add to the uudecode part of the manual a
suggestion to use Dave's program (script) above if you get the error message
no end of line (or whatever it was)? Some day another person may try to use
uudecode on grex and start this same discussion again, otherwise. (And can
Marcus' explanation be part of the manual?)
Will report back as to whether we get this file to uudecode okay.
Re 231: I looked for flip, which I would also recommend (even over dos2unix for most uses, since flip will do in-place conversions where with dos2unix you need to redirect the output...although piping the output through uudecode might work better with dos2unix) but I couldn't find flip on grex, either. In general, unix is all about learning new programs since huge monolithic multi-functional apps are not welcomed, usually. You get by by stringing two or more small commands together (and generally get more flexibility that way, too). I wouldn't worry about learning the dos2unix command. It's just a drop in the bucket of all the commands you could learn, and if it helps you, do it. You will be ahead in the long run especially if you use a unix system frequently.
One wonders what item it is I'm in.
Yes. Sindi, I really don't think these are Grex system problems. The Info conference has been pretty near dead for quite a while now, but that's been a really good place, in the past, for trying to debug problems of this sort. I'd suggest entering a few new items there. (Or maybe I'm the only one who still has that in a .cflist. <sigh>)
I have it in my .cflist too.
Thank you all, still trying to decode DELTA.ZIP, but I got a book on UNIX (for dummies) from the library. Will ask further questions in Info.
I am happy to report that uudecode is working fine now, and that I may have solved the problem of the screen freezing up by switching modems. (The second modem kept disconnecting, but this one works okay). Still cannot get zmodem to download, or even upload now, but Ymodem batch works, and Kermit, in both directions. Time out to learn Unix for Idiots. Thanks to all of you for all the lessons.
You'll probably find that you already know more unix than does Dummies.
My terminal's word wrap isn't working at the moment. Any idea what's going on?
Wish I could help, Senna. Could it be some setting on your communication program that needs changing? Rane, is a dummy or an idiotsmarter? This book is for Idiots, not Dummies.
Reasonable, if not good, question. A dummy just doesn't know much; an idiot has incorrect ideas. Or, at least that's how I see it (at this moment). It is harder dealing with idiots than dummies. Idiots probably don't read many books; dummies might want to learn. So, what sterling cutting-edge Unix knowledge is imparted to idiots?
"Dummy" and "idiot" used to have fairly specific meanings that are only partly related to their current usage (especially "dummy", which referred to someone who could not speak, "idiot" at least still connotes feeblemindedness, though generally of a different magnitude) A similar generalization has taken hold of other related phrases (for example "moron".) Does anyone know at what time the technical use of terms like "idiot", "moron", etc. to describe levels of mental and developmental retardation fell into disuse and whether it was as a result of their being coopted as insults by the general population or for some other reason?
I think single terms like that cannot describe any complex mental deficits. Medical terms today more and more refer to a syndrome that can display a variety of features, or even more particularly to specific genetic or congenital defects that result in a syndrome. So the improvements in medical specificity obsoletes broad categorical terms. However I don't think there is a specific syndrome or genetic defect that makes learning Unix difficult.
Wrap is working today...
I decided I would not learn unix simply from reading agora. This book starts by explaining shell, editor, etc.
The problem with zmodem was not a system problem or a problem of incompatibility between two versions of zmodem. Thanks to Mike McNally's explanation of what a UART is, I switched back from the external 9600 bps to the internal 1.4K modem (with a newer UART), set at 9600 bps, and now can both upload and download with zmodem. His explanation of buffers may have something to do with the fact that at 1.4k my 8088 on grex freezes up (when I type too fast), so both problems are solved with the 1.4K modem at 9600. Many thanks to everyone for their ideas. It was not a software problem.
'idiot' and 'moron' have specific definitions in terms of IQ range. The specifics I did not deem important enough to commit to memory...
That's now all obsolete. Even "IQ" is an obsolete measure, and no longer has any scientific use.
However, last I heard that was *not* a Grex system proplem.
You think "dummies" and "idiots" are not Grex system problems?
Around 5:00 this afternoon, and again just now, I dialed in to 761-3000 and got "modem answers but no Grex." Each time I re-dialed 761-4931 and got in with no problem. What's going on?
Similar. Twice today i dialed in and got the CONNECT 19200 message i always get, but no welcome and no login prompt. the first time, i tried a couple times more and then got in. this time i tried a couple times more and then dialed my isp and telnetted in. on all occasions i dialed 761-3000.
re #251: Nope, it's nice to have someone to talk to. (-:
Grex was crashed last night, not sure when it started.
I think there is a built-in filter against anyone who cannot figure out how to use the system, so therefore a shortage of dummies online. I have noticed how eager people have been to help me, so I must be a real rarity.
I just tried to telnet in and was #73 in the que! (Im still in the que ten minutes later at #55 as I type this via backtalk) Has grex been cutting telnet ports or something? Why is there any reason to allow so many into the que? Wouldnt it be better to cap the que at a max of 35 or 40? Anybody above 35 or 40 in the que, grex could give a: "sorry, Grex is full at the moment. Try again later or feel free to visit us via backtalk at www.cyberspace.org" and then disconnect. I imagine many who find themselves in the 60's or 70's in the que are just ghosts by the time they reach the login prompt. Any statistics on what percentage reach logon after being in the que but then dont logon?
So basically what you want is a way to get into the queue ahead of others who you figure probably aren't paying attention by the time their names come up? If they're gone by the time they get to the head of the queue the system will take care of it within a minute or two and move on to the next person. The whole point of having a queue is so users can get a connection in a bounded amount of time with connections assigned in an equitable manner -- if the queue "filled up" and you had to compete for a spot in the queue you could get screwed, with people who started behind you entering the queue ahead of you -- in fact there'd be little point in *having* the queue because you'd be right back to the situation where people coming in would have to "attack dial" (or "attack telnet") for a limited number of spots and hope to get lucky. The current system has the advantage of being as fair as it can reasonably be made and is still quite efficient. I don't think anyone can argue with the fairness of the queue -- what you're really arguing for is the chance to have special priveleges and jump ahead of people who have been waiting longer..
It would be useful to know the "telnet attack" frequency after the queue is full, and also the frequency of timeouts when users reach the head of the queue, during the course of the day(s). No logical decision on the optimum length of the queue can be made without knowing its dynamics.
Or, of course, have a queue length of zero, right Richard?
The queue is definitely a Good Thing, and I think something that Mnet is sorely lacking. There's been talk of increasing the number of ptys to allow 80 remote users to log in at once (as opposed to the current setting of 64). After the mail machine is up, that'll take even more load off the sun, and could quite possibly permit even more simultaneous users (I'm looking forward to the day when Grex can support 100 users at once). I think that increasing the number of simultaenous users allowed (while being careful to make sure system performance has not been sacraficed) actually has two positive effects: Not only is the queue going to be shorter (on the average), but since more users will be able to log in, check their mail, conferences, etc., and log off, there will also be a higher turnover of users, and thus the queue will move faster.
(In fact The Queue is one of the main reasons which persuaded me to shift to Grex from M-Net.)
I haven't received a notice from M-Net that no ports were available for many moons. I also tried to telnet in to Grex earlier, and was shocked to find myself at a position over 70. The highest I'd seen before was in the mid-40s. Congrats to Grex on it's success. However, richard is not entirely off base here. I have telnetted to Grex and not noticed my position in the queue came up until I had been disconnected. I don't think many coming here during the day can sit at their computer and wait until their number comes up, when that can take 15-20 minutes or even much longer. My favorite solution would be to increase the number of allowed telnet sessions until there are some complaints about connection speed, but barring that, a limit to length of the queue doesn't seem out of line to me.
Unlimited queue length gives greatest flexibility to folks, so I wouldn't want to see that change. If I connect to Grex and find myself way back in the queue, I have the choice of either (a) waiting it out, or (b) disconnecting and trying again later. If I couldn't even get in the queue, my option would be limited to (b). How does that help anyone? When I telnet to Grex, I'm using a windowing environment in which I can do other things on the interent while I'm waiting in the queue. Waiting 15-20 minutes or longer is no usually no problem. I use an ISP which M-Net doesn't always recognize as "local" (i.e. Michigan-based). The non-local-guest telnet ports on M-Net are frequently filled up.
When I see that the queue is obnoxiously high, I just disconnect and wait till later. *shrug*
If the queue is too long, you can always come back later, which is what would would have to do without the queue, or with a limit to its length. M-Net hasn't given me a notice about no ports being available in ages either, but they consider the network I come in from to be "local", so I have access to a lot of ports that people coming from other places don't.
Could people far back in the queue just dial in instead? (If local).
Depends. Some people have ubiquitous Internet (at work or school, usually) but no modem.
How does one reach the Internet without a modem?
Easy. By turning on my work computer at my desk and clicking on Netscape. Or by telnetting, or lynxing, etc. It's all connected all the time.
Modems are like bicycles. Businesses and colleges have trucks.
Jim is trying to explain how there are other types of connections, such as direct digital to the Internet. I prefer bicycles, less wasteful.
Yeah, when I telnet in and get stuck in the queue I just do something in another window until I hear the beep.. BTW, very good idea putting the bell character in the login message -- it helps a *lot* when I'm not paying direct attention. About the only time I miss the login prompt now is when I step away from the computer or when I'm on a computer that doesn't beep..
I'm not sure I like the truck/bicycle anology, but I guess it fits. For transporting just you a relatively short distance, your bicycle is far more efficient than a truck. If you needed to move several hundred pieces of furniture to another city, the truck would be quite efficient while your bicycle would take several days to do what the truck could do in an hour. Imagine an office where you've got 40 people who sometimes need to do stuff on the Internet. To do that with individual modems, you would need 40 modems, at around $100 each, for a one time cost of $4,000. You would then need 40 phone lines, at a one time cost of $40 each, for a total of $1,600, and then around $20 each per month, so $800 per month. Add to that 40 dial-up Internet access accounts, at around $15 each, for a total of $600 per month. So you're talking around $5,600 for a startup cost, and then $1,400 per month. Even that doesn't touch on the personell costs required to support a setup like that. Very likely, a single dual channel ISDN connection would be more than adequate for everything those users are doing. For that sort of connection, which connects the office ethernet to the Internet, there would be around $800 in startup costs, including line installation, the router, and the ISP startup fees, and then around $300 per month for the line and the connection from the ISP. To top it off, the connection will probably feel a lot faster than the individual modem connections did. You're saying the individual modems are less wasteful? For an example on a smaller scale, there's my home network, which has three computers on it. One of them has to be on the Internet all the time, because it's my mail server, so I leave it dialed up over a 33.6 modem link. The other computers connect to the Internet through that computer, which means I can run a few network cables in the apartment rather than ordering another phone line. Again, you're saying that's more wasteful than connecting via multiple phone lines?
I don't mind waiting in line, and it is many worlds better than having to constantly bombard the system with telnet requests until I find a port, which is what I used to do. I don't have a telnet attack program, so I just clicked on it a lot. It was a pain.
re #274: Your numbers are a bit fudged -- if you're going to charge each modem for the 40 computers you sould also charge for the ethernet cards.. And if you're really using 40 $100 modems on 40 different lines you should, even without steady usage throughout the day, have periods where you're getting much more peak throughput than a single ISDN line could provide. I agree with your basic premise, though, that there is an economy of scale which makes "trucks" more efficient than "bicycles" for some uses..
I think the assumption is that there are no 40 person offices that don't already have ethernet. If there were, you wouldn't catch me working there. As to the queue, the number of ports has already been raised, and there are plans in action to raise it steadily until we get too heavily loaded, and then back off a bit. The 'problem' is that in *very* short order we have gone from 18000 users to 21000. That trend is not slackening.
As the question goes..."where is everybody"? They are not here in the conferences. What's party like these days? Or are they mostly just using e-mail?
Yes, I was making an assumption that an office with 40 computers would have an ethernet already. It's been years since I've seen one that didn't. I was also assuming that most of those modems would be used occasionally to check mail, but that having fewer phone lines than that and making sure the modems didn't conflict with eachother would require some sort of PBX system (or a terminal server that people could dial out through), both of which would cost a significant amount. Anyhow, I was trying to show a cost savings. Had I been trying to show less of a cost savings, I would have pointed out that it would still have been somewhat cheaper to get a T1 than to do all those modems, and all those phone lines, and all those PPP accounts. That doesn't even begin to cover the support costs whenever there is a problem with the PPP stack on a Win95 box.
I don't have a "modem" to connect to the internet. I use this fancy thing called ethernet. It goes to a hub to a router to I guess what could be clasified as a modem, but it's called a csu/dsu.
I have received an e-mail with another message forwarded as an attachment. I could read the attachment in pine, but I could not figure out how to respond to the attachment and include it in my response, except by copy/paste in a new message. Is there a way in pine to do this? Is there a way in IE e-mail to include the forwarded message in the message forwarding it (I could tell my correspondent how to do that if it can be done).
I always just export the message and then read it to the response...but that doesn't help if you'd like pine to put quote arrows on the attachment.....
Re the modems, if anyone would rather dial in to grex from at home instead of waiting in a queue at work, we can supply modems for $2 (about $10 total if you also need a cable and adapter) and really basic computers for $25. Come into Kiwanis Sat am or email me if you are in A2 but busy then.
re #281: "include-attachments-in-reply" is a configurable option in the pine setup menu.. or, when you get a message with attachments you should be able to use the "V" command to go to the "Attachment Index" screen, from which you can "S"ave an attachment. You can then attach it to a reply you have composed with ^J
I will look for "include attachments....", but with some fear of the *very large* attachments I don't want to include. Too bad there isn't an item in the menus in the reply window to do this. I did try "V" and "S" but got either a complete message with HTML coding, or the message without the header. Yes, copying and pasting didn't give me the quote arrows. Well, I'll try "include" - but could the sender change his preference to include fowarded message in message, in IE?
re 283:
Will you throw in a phone line? How about free long distance? ;)
I need help. I've got all these wacky files in my home directory, and I don't know how to get rid of them. All of them are zero-byte files with screwed up names like "ikg%", "K", and "1c7m06..." with extended characters in them. Any Ideas on what could've made these files appear? Also, bash seems to zip along fine for me, perhaps the message in the shell change program could be changed. Yes, after three(!) years of using the menu shell, I've switched over to bash. This is due to my recent introduction to Linux, and over the past week or so, I've been learning a great deal and have gotten used to the bash shell. I just installed Pine on my box this morning, and it runs great.
re #287:
If you want to remove files with funny characters in them, the main
thing to watch out for is to make sure that the shell isn't intepreting
those special characters in some way (i.e. as a wildcard character) and
coming up with some *other* filename that you don't want deleted..
There's probably a FAQ on how to delete files with funny names floating
around somewhere but given the names of your files I'd recommend trying
to get them with a wildcarded rm. Being very careful to pay attention
to what you're doing, I would:
bash$ cd
bash$ rm -i *
The "-i" flag is very important, as it will ask you for each file if you
really want to delete that file. It's possible that you already have -i
turned on by default but you really don't want to leave it off accidentally.
If you don't want to use wildcards, you can put the *exact* filename
(control characters and all) within single quotes (') and the shell will
do no expansion on the names (e.g. : rm 'bad?filename') but that may be
tricky with some of the control characters you seem to have in your
filenames.
Thanks. I discovered that I could use cuteFTP to get rid of them. I'm still new to using Linux in this way, and I'm being careful not to screw anything up. I still wonder what created them in the first place.
Those files with funny names were on your linux machine? If so, you may want to check the details on your hardware and check that against the kernel info. There are some buggy IDE chipsets that corrupt disk data. One of there is the RZ1000 for instance. Thse can be 'fixed' with compiled in workarounds when you build a kernel. If the files were here, well, then, line noise maybe?
re #286, we can throw in a phone cord, but not a phone line, Kiwanis is not in that business. The offer is to local grex members only (paying members will have the required software put onto the machine if wanted).
At about the fifteenth character on the fourth line of a 'mail' message, I could type no further but could hit a hard return and then continue. This has happened a few times before. Is it the system or something in my hardware or software?
Sounds like mail doesn't do auto end of line. It could be allowing a certain number of characters per an internal data structure that it calls a line. When you reach the end of it input halts until you hit enter to stick that 'line' into the output file and start over with a new 'line buffer'. This is something that is internal to many programs. Even though your terminal software shows you wrapping to the next line, the program at the host end doesn't see it that way. I always put the returns in myself unless I know for sure that that particular program doesn't need me to. (the 'tel' command (similar to write) has a limited message length. You must hit enter and send that telegram and start another if you have a message longer than about 2.5 'lines'.)
I am pretty sure that I hit return at the end of the previous line, but perhaps I forgot this time, thanks.
You'd have to forget for more than one line.. I'm not sure which "mail" program you're talking about (there can be several similarly-named and similar-appearing "mail" commands on a Unix system ["mail", "Mail", "mailx", etc..]) but I doubt any of them has a hardcoded line length less than 250 characters..
plain 'mail', had not heard of the others and I barely know how to use mail, just to read and reply and delete. I doubt that I forgot for three lines. This has happened before, on the third or fourth lines. Not important, just curious. Not like when it was freezing up before, I can delete a letter, just not type forwards. I got into the habit of using mail when pine took 5 minutes to open the inbox.
Yep, the files were here. Don't know what made them appear, but almost all of them had a date of 04/11/98. Can't remember what I was doing then, I don't think I was transferring any files. I tried to use the "rm -i *" command, but it choked on the first file that it encountered that was named "-B". Oh well they're gone now, I was curious if anyone else had a similar problem.
Re #292ff: 3*80 + 15 = 255. I suspect mail has a 255 character max line length. Can anyone say why Grex was down today?
Most Unix programs have a 255 character line length.
This response has been erased.
telnet driver?
I'm sure Valerie meant to say the tty driver.
It did not beep, but backspacing did cure the problem. I will make extra sure to press enter at all line ends in mail, or just use pine.
elm is a lot less of a system hog than pine, I'm told.
This response has been erased.
(terminal server seems to be down.)
You have several choices: