root@mjbtn.MFEE.TN.US (Mark J. Bailey) (09/13/88)
I am in the process of installing a brand new Xenix *2.3* 386 system for a
client of mine, and I was setting him up with an init replacement that
Sandy Zelkovitz had on his XBBS board, along with modem-ctl (bi-directional
getty replacement - not using ungetty), and, on my system (2.2.2), when
uucico tests a line and/or acquires that line to make a dialout call (I use
the same init and modem-ctl), it creates a LCK..* file for tty1A (the port
which I specify in L-devices as the port to dialout on). This is necessary
as it seems that the creation of LCK..tty1A is necessary for modem-ctl to
go to sleep on tty1A until uucp is through. However, on my client's 2.3
system (with HDB uucp), when uucico is started and is designated to try and
make a call on tty1A (Note *A* not 'a'), it creates a lock file for *tty1a*
even though I have tty1A listed in Devices! No matter what, it always creates
a lock file for tty1a instead of tty1A. This is a problem, as without the
tty1A, modem-ctl cannot determine the port is being requested, and therefore
never goes to sleep...nothing works! My reason for using modem-ctl is that
it took care of that darn modem getting stuck answering at 1200 problem.
My question is, has anyone else experienced this occurance, and if so, (SCO
are you listening? :-) is this done on purpose and why? If not, is there
some way to patch uucico so that it will create the proper lock files. It
seems to me that whatever is listed on the Devices line should be the device
the LCK is created for. I do know that it is getting the right arguments
as the 'mlock' routine in uucico is reporting "mlock tty1A succeeded" and
then the dialers arguments also show tty1A.
Thanks in advance,
Mark.
--
Mark J. Bailey "Y'all com bak naw, ya hear!"
USMAIL: 511 Memorial Blvd., Murfreesboro, TN 37130 ___________________________
VOICE: +1 615 893 4450 / +1 615 896 4153 | JobSoft
UUCP: ...!{ames,mit-eddie}!killer!mjbtn!root | Design & Development Co.
FIDO: Mark Bailey at Net/Node 1:116/12 | Murfreesboro, TN USArosso@sco.COM (Ross Oliver) (09/15/88)
In article <288@mjbtn.MFEE.TN.US> root@mjbtn.MFEE.TN.US (Mark J. Bailey) writes: >..................................................... on my client's 2.3 >system (with HDB uucp), when uucico is started and is designated to try and >make a call on tty1A (Note *A* not 'a'), it creates a lock file for *tty1a* >even though I have tty1A listed in Devices! No matter what, it always creates >a lock file for tty1a instead of tty1A. .... >My question is, has anyone else experienced this occurance, and if so, (SCO >are you listening? :-) is this done on purpose and why? If not, is there >some way to patch uucico so that it will create the proper lock files. This was supposed to be a helpful new feature to prevent conflicts arising from using the modem control and non-modem control devices at the same time. uucico (and cu) now translates the last character of the device name to lower case before creating the lock file. It should be fairly simple to modify modem-ctl to do this also. If you can't modify your program, you can get around this feature by renaming one of the devices so it does not end in a letter, for example, changing /dev/tty1A to /dev/ttyA1. If this feature causes too many problems, we may have to un-do it. However, I think it is an improvement over the previous method, and modifications to existing programs to account for the change should be fairly straightforward. Also, any future programs which access dialout ports should use the dial(S) library function, which has been fixed in the 2.3 Development System to work properly with the current lockfile/dial/getty/ungetty scheme. That way, any future changes to the dial/locking scheme can be incorporated into existing programs by simply linking with the new library. Ross Oliver Technical Support The Santa Cruz Operation, Inc. uunet!sco!rosso