ocker@lan.informatik.tu-muenchen.dbp.de (Wolfgang Ocker) (10/04/88)
Another long mail from Avygdor Moise has arrived. Here my comments
(about UUCP, Mail, OXM, ls, documentation, ... Ulli will write some
comments concerning the csh (again :-)).
~ Thanks for promptly replying to my bug/complaint report. I hope that my
~ comments help improving what is already a very functional and useful
~ software tool kit.
They do! Thank you very much!
~ Questions/Problems/Bugs/Wishes:
~ -------------------------------
~ Well, I got uucp going, easy stuff if you don't use an Atari-ST V2.1,
~ on my QT-Plus-II running OS-9 V2.0. Now that I have uucp running I
~ encountered a few small problems.
~ 1. smail - Does not accept the syntax user@host.
~ I ran pathalias on my map file, created a systems file, defined a
~ host name, edited the aliases file and installed the paths file.
~ The following destination is accepted my smail 'yugas!avy'. Smail
~ complains about 'avy@yugas'.
First hint: Try running smail with the option "-d". Smail then prints
lots of debugging info and doesn't send any mail.
The following problems could occur:
1) The "paths" file is not located at the correct position, default is
/h0/USR/LIB/SMAIL
2) The "paths" file must be public read.
3) The host "yugas" is not included in the paths file.
[4) others ... ]
~ 2. mmon -- uucp
~ Do mmon and uucp support an agreed upon protocol for sharing a MODEM
~ line for login and call-out ?
With the next version of mmon it is possible. For now you have to restart
the mmon after the uucico call with vcron.
~ 3. uucp
~ 3.1 When sending/receiving mail to a host unix system the end of
~ line characters are properly mapped out. How do we handle uucp
~ transmission of text and/or binary files among OS-9 and Unix
~ systems ?
The uucico does a LF/CR mapping (in both directions: LF->CR and CR->LF).
So text files are handled correcty. Binaries must be reconverted with
the "lfcr" utility (with option -u.)
I'm not sure whether the uucico handles file receive/transmit requests
correctly (it seems to work, but I don't know whether it's secure:
uucp mypw recco\!/h0/SYS/password )
~ 3.2 When it attempts to connect to a host system and fails, it
~ retries twice more. However, the second and third pass ignore
~ the ALTERNATE selections as prescribed in the "systems" file e.g.
~
~ ATZ OK-\d+++\dATZ-OK
~
~ will execute the +++ sequence on first pass but not on the
~ second or third pass during the same connect attempt.
Ohhh, that is a bug ... thank you for the hint.
~ 3.3 The option HAYES-20 will set the timeout value to 20 seconds.
~ How do we modify the timeout value on a DIR line ?
You can't set a timeout for a direct line. This piece of code in
DCP is a big hack and I kept going on hacking there.
~ 3.4 When debbuging uucp from the command line viz.
~ uucico -r1 -x4 -syugas
~ uucico does not gracefully abort upon receipt of the SIGINT (^C) or
~ SIGABORT (^E) signals. As a result we have to wait until it
~ decides to timeout and die.
Yes. I will change this in a future release. We here don't have the problem
because with the ST you can switch to another virtual terminal and on my
VME bus system I have four (physical) terminals where I can kill uucico.
~ 3.5 The -x0 option of uucico does not seem to turn off completely
~ the uucico debugging output.
The names of the transmitted/received files are logged, that's true and
intended.
~ 4.0 oxm mailer
~ 4.1 Missing ability to print a short list of the headers of all
~ unread (and possible read) messages. It would be nice to be able
~ not to be forced to read the mail (as in the -e option) when
~ mail is available, and only be inform as to whose mail is
~ pending - very much like the Unix "from" command.
With the next release of the OXM comes the utility "oxmnew" which displays
only the new messages.
~ 4.2 How do we instruct oxm to read other (users) mail files.
Not yet implemented.
~ 4.3 Missing ability to send mail in a non interactive mode and
~ redirect mail from files. e.g.
~ (I wish)
~ $ oxm joe <joe.letter
~ will send the file "joe.letter" to user "joe".
~ 4.4 Missing ability to send mail to more than one destination
~ (without the use of the CC: feature. e.g.
~ $ oxm tom dick harry <file.dat
For this the command "mail" is provided. Try "mail -?".
~ 4.5 Missing ability to send a multi-line closing signature containing
~ address information about the sender (typically stored in
~ ~/.signature). e.g.
~ --------------------------------------------------------------------
~ Avygdor Moise, York University. (C.R.E.S.S.), Petrie Bldg. Rm 344,
~ 4700 Keele St., North York, Ontario, Canada. M3J 1P3
~ UUCP: ...!yunexus!yugas!avy Tel:1 416 736-5359
~ --------------------------------------------------------------------
That is implemented. In the option page you can enter the names of TWO
different signature files.
~ 5.0 Dot files and ls.
~ 5.1 I find the use of file and directory names which begin with a
~ dot "." excessive. This leads to situation where users may run out
~ of file space without a "reasonable" explanation, simply because
~ they forgot all about things like ".LOG" directories. I STRONGLY
~ recommend not to use hidden file names in directories which have
~ been specifically created for the use of application programs
~ and not people, or else design a uucp, cron, and other daemons
~ that clean up the work space periodically.
I like hidden files. For example, a "del *" doesn't complain about
a directory "LOG" in the "SPOOL/UUCP"-Directory. On the other hand it's no
problem to cut down these logfiles. You can tell Vcron (via "crontab") to
shorten them, for example:
05 00 * * * tail /h0/SPOOL/UUCP/.LOG/logfile -o=100 >/dd/TMP/log.temp;
copy -r /dd/TMP/log.temp /h0/SPOOL/UUCP/.LOG/logfile;
del /dd/TMP/log.temp
~ 5.2 Given the fact that system files or directories may be hidden,
~ and now thanks to your efforts there are quite a few of them
~ around, it would be nice if the 'ls' utility would recognize UID
~ 0 when invoked and as a result will automatically turn on the
~ '-a' option to display all files including hidden ones. This may
~ be of importance to system administrators who normally function
~ as non super users and then use the 'su' command or login as a
~ super user to work on the system.
Good idea. I have to rewrite "ls" completely (I don't like the code I've
produced there :-(, and will relealize this feature then.
~ ---------------------- PART II -------------------------------------------
~ >~ 1. The disk format
[...]
~ >1) Start the debugger: debug
~ ~ ....
~ >That's all.
~
~ Just note that Personal OS-9 users do not have access to DEBUG.
Ohhh, hmmm anyway. Then they have to write a BASIC program or so ... (I
don't know another alternative. Is there a public binary patch program
available?)
~ >~ 2. uucp
~ >~ To be more clear about that. On the Atari 1040 ST, whenever uucico
~ >~ gets started, it dials the required number and then the system crashes.
~ >
~ >Thats probably because of a KERNEL(2.1) bug. When the first signal send
~ >to a process is lower 32 (and greater 1, of course) and a intercept
~ >routine is installed (as in uucico), the KERNEL runs into a bus error.
~ >Perhaps Microware should tell us a patch.
~
~ Thats sounds serious. If indeed this is the case, we need a fix very soon.
~ However, I could not reproduce this problem on our OS-9 V2.1 (Atari-ST).
~ Here is what we did (see source listings):
The faulty kernel 2.1 is 17006 bytes in length, the corrected version
(same edition number) ist 17018 bytes long. Perhaps you try sending signals
starting with 5 (or so ...).
~ >~ 5. Documentation
~ >~
~ >~ Some of the documentation is in German (sorry I cannot understand
~ >~ German). The part which is in English needs to be tailored to OS-9 and
~ >~ not just copied from Unix. This can be quite confusing to people who
~ >~ do not have access to unix and cannot mentally map file name
~ >~ conventions like /usr/local/lib into the appropriate domain under
~ >~ OS-9. Also things line "see foo (1)" do not make sense under OS-9.
~ >
~ >We agree completely. But don't be too stern. We do what we can, but please
~ >consider, that it's a lot of stuff and English is not our native language.
~ >The documentation for the OS9LIB is completed now and ready to post.
~
~ As a programmer who wrote documentation for his software and
~ who knows that writing documentation is at time harder that actually
~ writing the application, I fully agree with you. My practice is to
~ provide in the absence of finished documentation a short
~ description (a single paragraph) of what a given program does, what the
~ command line parameters are and how it interacts with the user's
~ environment. Typically it takes about 1/2 a page or so...
We are working on the documentation. For now, we have completed the man
pages for the OS9Lib (which we will post soon). Our "little" documentation
is the help text displayed when you use the "-?" option.
~ I have implemented the Suspend capability in my version of microEmacs 3.9(2)e
~ using the following algorithm (Copyright 1988 by A. Moise, permission is
I personally have never missed (at least during the last year) a layered
shell. I have virtual terminals on my ST (max. eight) so I can switch
at every(!) moment to another screen (shell). No programs need to be
adapted. But nevertheless, thank you for the suggestions. Perhaps we
will realize it in the MINIX shell. I talked with Reimer about
that. He has tested your program and he said: "It's nice, let's
implement this feature in the MINIX shell ..."
Another idea: There's a public domain program for UNIX available, called
"screen". It relies on PTYs and simulates virtual terminals on a normal
serial tty line. The KA9Q implementation for OSK contains a kind of PTY
driver. Perhaps we should have a look at this stuff ...
/// Wolfgang Ocker (TOP)
P.S.: Our new e-mail address is "...!altnet!altger!top".
--
| Wolfgang Ocker | ocker@lan.informatik.tu-muenchen.dbp.de |
| Lochhauserstr. 35a | pyramid!tmpmbx!recco!weo (home) |
| D-8039 Puchheim | Technische Universitaet Muenchen |
| Voice: +49 89 80 77 02 | Huh, What? Where am I? |