home *** CD-ROM | disk | FTP | other *** search
- Last revised Fri Jun 23 02:05:20 EDT 1995
- -----------------------------------------------------------------------
-
- > I have problems with lock files under SCO when I use ports
- > not conforming to SCO naming conventions.
-
- Read README.SCOTTY.
-
- -----------------------------------------------------------------------
- Why do I get "Device busy" on trying to open a line I know
- no one is using?
-
- If you are using FAS or a driver of a similar discipline,
- EBUSY can be returned if (from fas.c):
- "If the device is already open and another open uses a different
- open mode or if a getty open waits for carrier and doesn't allow
- parallel dialout opens, return with EBUSY error."
-
- Getty can disallow parallel dialouts if it opens a line whose
- "magic number" contains 11 in the the 0xC0 position.
- From the fas README:
-
- Bitmap: m m f f x x x x
- `m m' are the mode bits as follows:
- ...
- 1 0 The device is carrier controlled. It blocks on open if
- there is no carrier.
- 1 1 Same as mode `1 0', but a parallel non-blocking open
- is possible while waiting for carrier.
-
- Using 11 in the mm bit positions are the correct choice for ISC
- to mimic the ISC asy driver behavior. On SCO, you should set
- these bits to 10 for the "modem" or "getty" (uppercase) name to
- mimic the sio driver behavior. There are other mechanisms for
- handling use collisions.
-
- If you are using some other driver, the ISC behavior may
- be the only option available.
- -----------------------------------------------------------------------
- One thing that I would like to see is some sort of side utility
- for changing colors in a shell like ecu does when I run it...it
- leaves the colors permanent for me. Love the blues and yellows
- that I have set up for trn...makes things actually easier on the
- eyes to read.
-
- Color is supported fully on SCO and partially on ISC. Code for
- SVR4 has been turned off (I had reports it was "broken"). Since
- xterms are the only things that make sense on the Sun (who uses
- the console? shelltool/cmdtool? barf/vomit/just-say-no), color
- has not been attempted there. Does anyone know about a working
- "colxterm." I tried long and relatively hard a couple of years
- ago to port it to SCO with no joy.
-
- SCO has an ioctl(screen_file_descriptor,GIO_ATTR,&character) that
- returns a 6845-style attribute byte (foreground and background
- nibble pair). I use that at initialization time (ecutty.c
- get_initial_colors()) by asserting reverse video and reading the
- colors, then setting normal and reading those colors. (On SCO,
- "reverse" and "normal" can be set to arbitrary pairs of colors,
- not just flipped from each other). These values safely
- squirrelled away, ECU or the user can do what they want. When
- ECU exits (ecutty.c restore_initial_colors()).
-
- This save/restore feature is turned on only for SCO (#ifdef M_SYSV).
- Perhaps it should be #ifdef GIO_ATTR, but SCO only included the
- identifier in sys/console.h recently. I have no idea who else
- has it (oh yes I do ... ISC sys/kd.h has it).
-
- If anybody wants to look into this, I'm sure there are many who
- would appreciate it.
- -----------------------------------------------------------------------
- > Can a finish and/or bye command(s)
- > be added to the kermit transfer section? We've found that we needed it
- > many times, and it would have been nice to be able to abort out of the
- > kermit server instead of hanging up.
- >
-
- With C-Kermit 5 available to ecu's $PATH, to finish a remote
- server active on the attached comm line, try a one line procedure :
-
-
- system -s 'ckermit -l '+%itos(%conn)+' -b '+%itos(%baud)+' -f'
- ^^^^^^^ you may have named it kermit or whatever.
-
- You may have to play around with your .kermrc file if
- you have checksum/block check type problems. This is
- unlikely if you have been using Kermit with ecu since
- you must already have a .kermrc
-
- Or look at the following procedure. (Does it inspire other
- things you could do by plugging a temporary .kermrc :-> ?! )
-
- There may be syntax errors since I didn't test this. What example
- ever has been tested:-/ ?
-
- #
- # build tmp file with proper checksum in it.
- #
- $s_rc = '/tmp/'+&itos(%pid)+'.tmp' # choose name
- fdel $s_rc # make sure it does not exist
- fclose 1 # make sure file 1 is closed
- fopen 1 $s_rc # open the tmp file
- # if error opening file, complain and exit
- ifi $i0 != 0 echo 'Cannot open '+$s_rc+': '+%errstr($i0); return
- # !!!!!!!!!!!
- fputs 1 'set block 3' # or whatever
- # !!!!!!!!!!!
- fclose 1 # close the file
-
- #
- # build command:
- # ckermit -l <ttyfd> -b <baud> -y <tmp .kermrc>
- #
- $s_cmd = 'ckermit -l '+%itos(%conn)+' -b '+%itos(%baud)
- $s_cmd = $s_cmd+' -y '+$s_rc+' -f'
- echo 'executing "'+$s_cmd+'"'
- system -s $s_cmd
- fdel $s_rc # remove tmp file
-
- -----------------------------------------------------------------------
-
- > I'm running on an Esix SVR4 machine and I am using ttymon.
- > Whenever I attempt to open my "default" serial port
- > "/dev/term/i1A" it will fail, telling me to readjust the
- > permissions (a chmod 666 /dev/term/i1A will solve this). However
- > I was wondering if you have any plans on fixing this item?
-
- ECU is incompletely ported from SCO in that the ecuungetty
- feature only works on SCO. Ecuungetty, a setuid program, looks
- through utmp and at lock files to determine the availability of a
- line. If a line is free, ecuungetty chown/chmod's the line so
- ECU can get at it. When the line is close, ecuungetty is again
- called to re chown/chmod the line to uucp. This makes ecuungetty
- exceptionally system dependent. There is a small amount of
- similar code in ECU too. Any volunteers?
-
-
- > I've heard about your ECU program for SCO Unix. I would like to
- > compile it under BSD (namely Ultrix). Is there a version for Ultrix
- > available somewhere in the Internet ?
-
- Alas, ECU requires System V termio. Adding BSD line I/O is
- unanticipated at this time. I need to revive tipx, which
- provides X/Y/ZMODEM support to 4.3 BSD tip.
- -----------------------------------------------------------------------
- > When I get half finished with a download of a meg in size and it dumps,
- > I get wishing that I could tell the zmodem part of ecu that I
- > want it to tell the other end to start sending at a later point.
- > More specifically, I want it to find the file the other end is
- > sending, discover the filesize%1024 and tell the distant end to
- > begin sending at that point, not at 0. Is this available already,
- > or should I look at (and hack) your zmodem code?
-
- The sz function supports resume transfer, but I know of no way
- for you the rz user to force the resume from your end.
- There is a resume feature present... the sender must support it
- and you must instruct him to do it.
-
- For instance:
- sz -r big.file
-
- -r Resume/Recover interrupted file transfer (Z)
-
- This merely makes the receiver send an initial ZRPOS of local
- file size minus 1024 instead of the usual ZRPOS 0.
-
-
-
- ------------------------------------------------------------------------------
- Warren Tucker n4hgf!wht or wht@n4hgf.atl.ga.us
- All readers cannot be leaders, but all leaders must be readers. - Harry Truman
-