Citadel Q&A>
JayDee
puzzlement:
I went out and set up a new Cit last eve, and took a prefetched packet
along that had been incorporated @ LOK for that node (TEH.SEA), so they
would have a few K messies to mess with and go-figure out some Cit...
autoroom for LOK was set on, and the rooms autohalled to Net Hall
the rooms on the prefetch were controlled by a preinstalled roomreq.teh, too
well, the packet incorporated, but most of the rooms (about 200) ended up
being skipped, with only a few of them being created at random
a few more rooms were made when I prefetched a packet for LOK
where am I hosing-it?
we are just being so lazy and don't want to have to key in all the rooms
she wants to carry on her Cit...
chester cheetah
You have to continually prefetch. taht's the only way Iknow of.
JayDee
screw that...not on a 8088
I'll just make the Cit on a 486, go-in and kill, zip it up, and transfer it
make a prefetched packet with its' roomreq.lok file, and incorporate that
it is a question of delivering Citadel PSS efficiently in allocated time
perhaps I can have the new node sysop negotiate a larger timeslice for me
in this particular case, that just may be possible, yepyep
and I can be sneaky, and arrange net-partners concurrently...using
WinTerm as-to conceal the fact that I'm Citting
Windows, Citadel, and Efficency...oh my...
MontyL
You didn't include a NETID.CIT in the setup, did you, JayDee?
John
Why is he saying "Citadel PSS"? Does anyone know what he's talking about?
JayDee
yep, a big fat NETID
and Citadel PSS = Citadel Product Support & Service
Livia
he's talking about helping someone set up a cit
he obviously thinks it's cute to refer to it as PSS
Livia
citadel is not a product
JayDee
cute as your BUTT, Livia
and it only took a little while, too, 102 rooms on a 10Mhz 8088
and the sysop knows her shit, and will soon learn her Cit...
she started with a Trash-80, programming in BASIC, went to a Sanyo 555, and
ported stuff to MS-BASIC 1.31 (real nifty version, btw...beats GW), and
then got into real IBM instead of that mutant MS-DOS Sanyo
didn't use documentation, either, for the most part
got me high, too, while I was hanging-out waiting waiting waiting for
reconfigure/defrag/get strange modem init string figured...
that was nice of her
JayDee
ok, I see that
sorry...confusion due to advanced bastard senility...
Citadel+SS
MontyL
Product Support Services, John...
uncle herb
I still don't get why GremCit has built-in text files on the age of consent.
Saint Bob
Probably to point the wolves toward it when they start drooling over the lambs.
Livia
you're not an official support person
JayDee
ok, Livia
you have a gas-efficient car, and know Citadel over under sideways down
you make the trek over to Queen Anne Hill, ok?
and come-on down to Tukwila, too, when BSA is broken and incoming packets
are piled-up by the hundreds in the cit directory instead of the trans dir
and run out to the Sappho-Beaver area in the dead of night for an install...
and I'm not an official MavenCit Support Person, either
but I've supported that bug-riddled piece of shit, too, in the past...
now, I notice the big fat hidden files MavenCit creates and just leaves
scattered all over a volume, and preventing a good defragging on the disk
there has been a change, apparently, somewheres down-the-line
I just DON'T GIVE A SHIT
Livia
RAPE.HLP is an err head relic
he likes slutty teenage girls
MontyL
I didn't see 'official' in his description...
JayDee
it is real, John
be afraid...
John
And what's that LIVIAN.TXT thing all about? Really strange. Does anyone know if it's for real?
Black Wolf
Besides I think I'd much rather have JayDee for product support than some of the so called official ones...
JayDee
thank you for the vote of "confidence", BW
but Citadel, in the GremCit/Citadel+ incarnations is not a product
neither is MavenCit, but the EFS is, as-is TurboCit
but they all have been produced
semantics is such a jungle...
Blain Nelson
Jaydee is the Cit SS? A room-based storm-trooper. Hmm.
Jack Daniels
it's not there anymore, a friend of mine wanted to know the laws, and i found it has been removed in 65
JayDee
that will happen when I infect the SS with Cit
she hates it now, but I will have my way with her
she is a room-based storm trooper, btw...come too think of it...
eep...
Lab Rat
Well for anybody who wants it, just grab the 64_hlp.zip package or whatever your looney sysop has renamed it to and that takes care of it.
Livia
we decided that it wasn't quite the thing to put in the standard distribution
when err head acquires worldwide fame, we can put it back
StraitJacket
You do have that right.. SS is a room based stomper... But infecting SS with CIT? I am not sure that would have any positive output.
JayDee
heard that, SJ
just exposure to Cit has caused ragwarz
but maybe, just maybe, these warz could be migrated to Citadel
oh
what fun...
Megamol
ooooooh sj i really don't know if giving SS their own hallway would be a good idea.....might cause crashes on the system......not a put down against SS, just an observation.......I've known SS since 4th grade....
StraitJacket
Woo.. give the SS their own hallway... they could keep themselves amused for hours.
JayDee
all a VERY interesting proposition
Just-George
Where and what kind of hidden files does MavenCit produce? I haven't found any that should not be. Just for my own info, please!
JayDee
go over to BSA and have a look-see, J-G
sorry I missed the barbecue...that entire weekend got hosed, as-usual...urf..
only, this-time, I'm radically-pissed, and it won't happen again
I'd give a dissertation on the situation
but I promised someone I wouldn't do that kind of thyang, anymore...
Just-George
BSA? Real board name and number please! :)
JayDee
I ain't going to EVEN get into that-one, J-G
the ratshit that came down a couple-years ago left me burned
you will have to call her yourself and get the name & #...
Counselor
JG: JayDee actually means Rumble Turf, Bypass's Sub-board.
If BSA was generating unwanted hidden files, it would be a GremCit problem.
But, I'm not aware of any problems in that regard...to that extent, I too would be interested in knowing about them.
JayDee
the hidden files won't hose-off without putting a bigger hose on than
I'm willing to-do, and they interfere with the disk being defragged right
I ain't messing with it
I'm so lazy
JayDee
bite the wall, Brent
that is exactly what had happened
and they were strewn about the disk, according to the map defrag displays
and quit your halfassed babbling, yourself...it doesn't become you...
Brent Bottles
I think that you can just ignore JayDee's babbling.
Unless The Maven has completely lost his mind, the only hidden files that MavenCit would be creating would be swap files when it super-shells.
If the computer crashes while super-shelled, then the hidden file sticks around, as Citadel never gets control back, so it doesn't get a chance to delete it.
And if Rumble Turf is indeed a sub-board, then if the computer crashes while super-shelled out of MavenCit, thus leaving the hidden file around, then it would also be super-shelled out of GremCit, which will also leave its hidden file around. (As I would be beyond surprised to learn that MavenCit would fit into memory if you don't super-shell out of GremCit.)
The last I heard, The Maven was using the same super-shell routines that we are, so it will operate exactly the same. (Though he might still be using the version with a slight bug in it that caused the hidden files to be created in other than the #TEMPPATH, thus JayDee's babbling about them being strewn about the disk. Or, as is very likely, this was something that happened once about a year or two ago, before Ralf Brown released a bug-fix for that bug, so of course the bug was still in MavenCit.)
JayDee
I already knew what was happening
I was even in a position to use the tools I had put on that system, to fix it
but JayDee don't do that no-mo
go figure out just why...
I just walk-away...
Counselor
Well, JayDee, at least Brent shed a bit more light on what was happening...why didn't *you* say that if you knew. :[
Livia
we only wish
JayDee
there is that "we" again...
hosed attitudes
Counselor
But, JayDee, whether you "do-that" anymore or not, you could have explained it to those of us inquiring in here...
JayDee
it was a MavenCit thing, Counselor
I didn't think anyone really cared
and, now...hmm...
Just-George
Well I looked at my disk with PCTools and found no hidden files and was able to defrag with either MS-DOS or PCTools defrag utils! :)
JayDee
that is your disk, J-G
on the machine in question, the story is different
and I ain't gonna mess with it, because if I do...
and it goes sour
I get a pile of poop...
and there has been a long history of too many people cookin' on that box...
Livia
he doesn't need you to mess with it
he fixed it himself
learn to read
JayDee
you may know how to read, Livia
but something, like your vast dislikes for me, is altering your perception
I said "on the machine in question", after "that is your disk, J-G"
if you weren't so quick to jump in my shit, maybe you would have .rr and
figured out I was talking about BSA.SEA/Rumble Turf, not the Jolly Time
but, no
you pounced like a raptor
JayDee
be my guest, J-G
and if it still refuses to cooperate, shoot it in the hdd
Just-George
BSA can be fixed the same way as any cit that crashes and then is restarted. Take both boards down and then kill all the junk files that have accumulated! Quite simple! And with any decent system that has the ability to look at hidden/system files and delete them you can too! :)
Black Wolf
*kapow* you are hereby assassinated
JayDee
yeah...
I remember that temp file I looked into when you took Anarchy Park down to
the DOS level...it had the elegia message in it that I thought was odd, at
the time...
but I have never had any problems with temp hidden files, using GCIT, being
left after the system had crashed, so I was figuring that GCIT was smart
enough to get rid of them
and I shouldn't speak of MavenCit in this room
one of you should have assassinated me, instantly
Brent Bottles
Wow. I was giving you some credit when I was interpreting your "strewn all over the disk" comment. I thought that you were talking about logically strewn all over the disk.
But, no. You were talking about physically strewn all over the disk.
The physical location of the files on the disk is something that is completely out of the application program's control: it is an operating-system thing. (Of course, with MS-DOS having no protection built into it, as it runs on a CPU with no protection mechanisms built into it, an application program can easily bypass the operating system. This is how disk defragmenters work.)
What you describe is a completely normal action of a system that has crashed while temporary files exist: the temporary files are left on the disk, and they are "strewn all over it," as the operating system has done. It just happens to be that the temporary files in question were made as hidden, which is again not something that The Maven has control of, as that is done by the swapping library.
And you are incorrect in assuming that it is a MavenCit thing, and therefore Counselor should not be interested. Incorrect in two ways, even: Not only does Counselor run a MavenCit (last I heard, at least), but MavenCit and GremCit share the same super-shelling code, so anything that MavenCit does while super-shelling GremCit will, too. Including making temporary files (that's how super-shelling works if you don't have a lot of XMS or EMS around: all of Citadel ends up in a temporary file on your hard disk) that are hidden (so you don't see them while shelled and think, "gee, this looks like a stupid file, I better delete it") and that do not get deleted if the system crashes or is reset while shelled (as Citadel never comes back to clean itself up).
So what you are talking about is a completely normal operation of Citadel.
JayDee
you are usurping someone-elses' position, Black Wolf
they might take offense
and bite you severely...
MontyL
Heh... Falls into the "oops" category, JayDee? ;)
LITTLE MERMAID
Oh my gosh!!!!....I tripped over my tail and fell in to the wrong room!(sorry)
Bookworm
He would enjoy that! ;)
JayDee
well, I don't think she really bites...
but, I have several styles, personally
ZooKeeper
depends on the who how and where of biting, tho
Torch Song
Oh, yes...
Megamol
and how hard....
MontyL
Sheesh, it spilled over into here... }:>
Brent Bottles
I see the Net ID of this room as "Citadel Q&A"
While your posts might not really be considered questions and answers, so strictly speaking don't belong in here, I see no reason to exclude talk about MavenCit in here. (Unless The Maven has decided to declare that MavenCit is actually not a Citadel. In which case I would ignore him and still think that discussion of it in here is perfectly valid.)
Try this from Loka some day. Press the F6 key. Press the @ key. Once you get to a DOS prompt, press the Control, Alt, and Del keys all at the same time.
When the computer is sitting at another DOS prompt, take a look at the hard disk.
If there are no temporary files, it means that you had enough memory free to super-shell in memory, and therefore would never get temporary files on your disk from super-shelling.
If there are temporary files, then there you go.
(If you didn't get any temporary files, and want to see them, try allocating all of your XMS (a ram disk or disk cache is an easy way to eat memory in a hurry) and then super-shell out of Citadel.)
Brent Bottles
And I can think of no reliable way to have Citadel be smart enough to delete the temporary swap files after a crash.
JayDee
yeah...that is the way I have always set this Cit up, is to have the swapfile
in memory...has taken a bit of jugglin' to do it, too, with the .exe and
all .dat (except msg.dat), .cit, and space for .tabs running on a RAMdisk
if I load the PRINT tsr, there isn't enough free mem to swap to ram anymore
if I reduce the size of the RAMdisk, there isn't enough room for the ect.tab
I guess the only time I ever really dealt with a swapfile on the hdd was when
I was running those experimental sub-boards
I guess Maven could call this ver. something other than a Cit
but it would be rather dumb to do that...
Alpha Wolf
maybe on startup look for old temp files and nuke them?
ZooKeeper
there there Monty, it's all a bad dream. You won't remember it in the morning... ;>
JayDee
a pot-o boiling Cit, eh?
Saint Bob
I delete 16Mb worth of swapfiles the other day.
JayDee
eew
Inchee
The Maven, through limiting the Alternet to only Mavencits, has cut back on many networking problems...
JayDee
just get a half-gig drive and let it fill-up, then clean it off every year...
ShadowStalker
that's the biggest bunch of shit i've ever heard. i've seen the sysop rooms on "alternet," i've seen half the rooms on "alternet"... they are no different than CitNet except for the network is much smaller and full of much more petty people.
MontyL
Twice, even... No time for regular maintenance, St. Bob?
Brent Bottles
Alpha Wolf: And how to you propose that Citadel know what is an old temp file?
You cannot go by the name: It is random. (You cannot make the name fixed: if you did, then you could not super-shell out of a subboard that was super-shelled to.)
You cannot just delete all hidden files in the #TEMPPATH directory: this might be a subboard that was just super-shelled to, and doing so may delete the swap file of the main board, which would cause it to crash when it tries to return (as it needs the temp file, of course).
So what can you do?
Nothing that I can think of. (And, nope, you cannot look at the date: you might have super-shelled out of something several months ago, so the date on an active swap file might be several months old. Or, part of the batch file that calls a sub-board might change the date for some reason.)
Blain Nelson
Brent -- Why can't you just require sub-boards to use their own #temppath. For that matter, why can't you name the swapfiles with the node alias as part of the file name? Do you really need all eleven characters?
I'm confused.
Brent Bottles
Blain: what if your sub-board has the same alias, as you don't net it and were too lazy to come up with a new one?
No we go to phase two of why we cannot use different file names: the file name in use is completely out of our control. It is picked by the shelling library that we use that actually does the swapping. (And, no, I do not feel like figuring out how to do it myself, as there is a perfectly good product out there that does it for us. I don't feel like rewriting the Borland C++ Run Time Library at this moment, either.)
And, as this library is freely available, other programs could use it, too. You have no way of knowing which other programs sitting around on the disk might use it.
Which is why the author of the library chose to use random file names, I am sure.
MontyL
Agreed, Brent, but Blain does have a point: With a slight rewrite of the shelling module, GremCit could be able to clean up after itself, even through a system crash. Which is a good thing.
Twoflower
Brent: you could have citadel create sequential swap files (like, say, SWAPFILE.1, SWAPFILE.2, SWAPFILE.3, etc.)
Of course, this doesn't allow for restarting Citadel with -c and having the BBS possibly destroy swap files that are actually in use, but I think it would be, in fact, fairly rare that someone would try to reconfigure the BBS after doing a supershell. And I have no idea if it would work. Hmmm. Maybe I should try that and see what happens. Probably just creates all new tables and ignores any old swap files, which would mean that you could just have cit delete all sequentially named hidden swapfiles when doing a reconfigure regardless of being shelled or not, except in the case of a subboard, and hey, that's just too fucking bad.
Twoflower
I guess you could, instead, make swap files dependent upon the address, so they would turn out like TFTSWAP.1, TFTSWAP.2... then a subboard reconfiguring would not delete the swap files of the parent BBS unless the addresses were the same (at least, the first 3 characters of the addresses.)
Blain Nelson
Brent -- You don't want to rewrite that just because it's already been done? Well I just don't know what to say about that!
If folks are too lazy to make the program work then guess what? It won't work! Wow! But reason two is adequate for now.
But what about the "keep your #temppath to yourself" idea. Laziness of course is not an adequate excuse because (see prior paragraph). Is there any reason to believe that any other application would be writing its swapfiles in a directory named something like c:\cit\temp\swapfile\just\to\be\sure? I mean, we can get as silly as necessary to make this work.
chester cheetah
Hello Brent, good to see you back in the world again.
Brent Bottles
A solution that is not 100% safe is not acceptable to me.
And this goes beyond mere laziness: it goes into allocation of resources.
I am sure that it would take me at least a week of full-time programming to make a super-shell routine. It would probably be closer to three weeks.
(And by full-time programming, I am thinking of at least 60 hours/week.)
So this is 60-180 hours of work (perhaps more).
As I probably have no more time than 10 hours a week to work on this, this could be four and a half months of work.
That would be quite the wasted four and a half months, considering the potential benefit. (Not having to do a dir /ah/b/s \ when you think you have less free disk space than you should.)
Clue: it took me less than eight hours to write the full EMS/XMS/Virtual message table. Since then, I have spend another two or three to make it more general-purpose so it can be used for the room table (64) and the scroll-back buffer (65).
Do you really think that it makes sense for me to spend that much time for so little benefit? I certainly don't.
Livia
unless blain is willing to pay you the going rate for contract programming
Livia
maybe people could just learn how to find and delete their own damn swapfiles
you can't really all be that stupid
JayDee
don't count on people not being stupid, my love...
Richard Goldfinder
What was wrong with my suggestion (in GremCit Test, responding to Blain) of
erasing extra swapfiles automatically in AUTOEXEC.BAT? Unless you're one of
those types that runs AUTOEXEC manually during a session (in which case maybe
you deserve to potentially crash), you could just throw in lines similar to
these:
attrib -h c:\cit\temp\*.*
echo y | del c:\cit\temp
with reasonable assurance that they won't be executed while shelled out of
something. It doesn't seem necessary to burden the running program with a cleanup from
crashes such as this.
Oh, and note that this will kill all of the files in that temp directory: if
you're one of those types that likes to keep real files in your temp directory
(there are several where I work), just move them to another directory (copy and
move don't act on hidden files) then move them back after erasing the hidden
files.
Blain Nelson
Okay livia, maybe it'd be simpler to let folks know that the swapfiles are there and that if you're at a raw prompt with nothing else running you can kill the sobs. I just don't remember anybody saying that until I ran into by accident my temp path and found about 20 megs tied up in old swapfiles. And it didn't seem to be that difficult a think to fix.
Brent -- That was sarcasm. I'm not asking you to rewrite the whole thing. I'm asking if something short of that can produce adequte results and I'm providing ideas of how it could be done. If my ideas suck, then fine -- say so. If you really want to be nice, tell me why they suck (you're pretty good at doing that, being that you're a nice guy and all).
And I don't think you mean 100% safe. As safe as possible, fine. But 100% safe would be an apparently new and definitely unachievable standard.
Citadel Q&A> _