usenet@codas.UUCP (04/06/88)
I am running XENIX 386 and would like to use the Elm mail system. I have the latest source code for Elm but cannot get it to run on my system. What can I do to make it work with XENIX 386. I have Rel 2.2.2 with latest 386 development system. Thank you Jim Voytek codas!pbsfwb!jim
billb@amcad.UUCP (Bill Burton) (04/12/88)
In article <10861@codas.att.com> writes: >I am running XENIX 386 and would like to use the Elm >mail system. I have the latest source code for Elm but >cannot get it to run on my system. What can I do to >make it work with XENIX 386. I have Rel 2.2.2 with >latest 386 development system. > > >Thank you > >Jim Voytek >codas!pbsfwb!jim I have been recently attempting to get Elm 1.7 working on SCO Xenix 286 and it appears that I have succeded for the most part. The one problem is that occaisionally I get a "Segment Violation" and sometimes a core dump. This happens if I use the "builtin" pager, or when I try and save the configuration options. Have compiled with the following flags (if I remember correctly): CC: -O -Ml2 -Mt24 -LARGE LD: -Ml2 -F 4000 Would appreciate any input on how to solve this segmentation problem. I am quite sure that that all the modifications made also apply to Xenix 386 with the exception of the compiling and linking flags. Here is a brief list of some of the changes I made: o hdrs/sysdefs.h: ifdefs for locations of mail directory, rmail and vi using M_XENIX define o modified Makefile's for appropriate compiling flags above o Chmod 777 /usr/spool/mail o Chmod 600 /usr/spool/mail/* o src/encode.c - added a dummy crypt() function so ld would't complain. Encryption should be ifdef-able. Originally I tried to modify Configure.sh to set up everything for Xenix and concluded that to maintain compatability with everything else would require a bit of work. The time might be better spent getting Larry Wall's Metaconfig and writing a new Configure script from scratch. Would appreciate any comments, tips, patches, etc. I'm sure someone else out in netland already has this running on SCO Xenix 286/386. Elm is a great program (thanks, Dave!) and I would like to contribute some time and energy to improve it particularly in the ease of installation. -Bill ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Name: William D. Burton US Mail: American Academy of Arts and Sciences 136 Irving St., Cambridge, MA 02138-1996 Audible: 1-617-576-5023 UUCP: ...!husc6!amcad!billb ARPANET: billb%amcad.uucp@husc6.harvard.edu ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
donegan@stanton.TCC.COM (Steven P. Donegan) (04/12/88)
In article <150@amcad.UUCP>, billb@amcad.UUCP (Bill Burton) writes: > In article <10861@codas.att.com> writes: >I am running XENIX 386 and would like to use the Elm >mail system. I have the latest source code for Elm but >cannot get it to run on my system. What can I do to >make it work with XENIX 386. I have Rel 2.2.2 with >latest 386 development system. I have elm 1.7 working quite well on my SCO 2.2.1 (286) system. The only problems I have noted so far are: 1) Encrypted mail does not work. Either with the hacked up non crypt code or with a 'real' crypt.o from an earlier development system (before irrational security paranoia set in). 2) If a user is in elm in some functions and gets disconnected the session remains running - SIGHUP appears to have been defused. If you want a copy of my sources let me know, or if enough people want this code to justify the bandwidth usage I'll be happy to post. -- Steven P. Donegan Sr. Telecommunications Analyst Western Digital Corp. donegan@stanton.TCC.COM
karl@ddsw1.UUCP (Karl Denninger) (04/13/88)
In article <150@amcad.UUCP> billb@amcad.UUCP (Bill Burton) writes: >In article <10861@codas.att.com> writes: >>I am running XENIX 386 and would like to use the Elm >>mail system. I have the latest source code for Elm but >>cannot get it to run on my system. > >Would appreciate any comments, tips, patches, etc. I'm sure someone >else out in netland already has this running on SCO Xenix 286/386. > Well, we have ELM 1.7b working here on SCO V/386.... although at the moment I don't have a "clean" copy so making diffs would be difficult... The changes required were not extensive; we don't have any problems with core dumps or other strange happenings... We are running ELM along with smail 2.5 (with the patches posted to the net a while).... works great!! I will see if I can backtrack the changes and get some idea of what was modified. --- Karl Denninger | Data: +1 312 566-8912 Macro Computer Solutions, Inc. | Voice: +1 312 566-8910 ...ihnp4!ddsw1!karl | "Quality solutions for work or play"
mike@ists (Mike Clarkson) (04/14/88)
In article <13@stanton.TCC.COM>, donegan@stanton.TCC.COM (Steven P. Donegan) writes: > > I have elm 1.7 working quite well on my SCO 2.2.1 (286) system. The only > problems I have noted so far are: > > 2) If a user is in elm in some functions and gets disconnected the session > remains running - SIGHUP appears to have been defused. Yes I've seen this too. Why is this happening? If the line drops while you are composing a message in the editor (in our case micro-emacs 3.9) then the elm and the emacs are left running on the login line. So the next poor sucker to try logging in has 2 processes competing for the same port: login and emacs under elm. If you type anything in, login may get it or emacs may get it, and if you type ^X^C real fast to login, emacs will exit! Not good for my dialup users. How is this possible? What on earth does Elm do to allow itself to avoid SIGHUP ? Puzzled and displeased... -- Mike Clarkson mike@ists.UUCP Institute for Space and Terrestrial Science York University, North York, Ontario, CANADA M3J 1P3 (416) 736-5611
chip@ateng.UUCP (Chip Salzenberg) (04/14/88)
I've been using Elm 1.7 beta for some months now under SCO Xenix/286 (versions 2.1.x and 2.2.x). WARNING FOR XENIX USERS RUNNING ELM: As distributed, Elm creates mail lock files in the _wrong place_: Locking style Lockfile name ------------- ----------------------------- V7 (used by Elm) /usr/spool/mail/username.lock Xenix /tmp/username.mlk You must change Elm's locking code (and fastmail and filter, if you use them) to reflect this difference. Otherwise you run the risk of corrupting your mailbox when new mail comes in. In article <150@amcad.UUCP> billb@amcad.UUCP (Bill Burton) writes: >I have been recently attempting to get Elm 1.7 working on SCO Xenix 286 >and it appears that I have succeded for the most part. The one problem >is that occaisionally I get a "Segment Violation" and sometimes a core >dump. This happens if I use the "builtin" pager, or when I try and save >the configuration options. Try looking at stack backtrace after the core dump (the adb "$c" command). That should give you a hint. >Have compiled with the following flags (if I remember correctly): > CC: -O -Ml2 -Mt24 -LARGE > LD: -Ml2 -F 4000 I used "128" instead of "24", otherwise the same. >Here is a brief list of some of the changes I made: >o hdrs/sysdefs.h: ifdefs for locations of mail directory, rmail and vi > using M_XENIX define >o modified Makefile's for appropriate compiling flags above >o Chmod 777 /usr/spool/mail >o Chmod 600 /usr/spool/mail/* >o src/encode.c - added a dummy crypt() function so ld would't > complain. Encryption should be ifdef-able. Chmod 777 is a BAD IDEA. Do you want other people removing _your_ mail file? Instead make /usr/spool/mail 775 and make Elm setgid bin. Then Elm can write to that directory but hostile users can't. (And Elm does check for your real permissions, so other users can't just use Elm to read your mail.) >Elm is a great program (thanks, Dave!) Hear, hear. [Eight-line signature deleted; shame on you, Bill!] -- Chip Salzenberg "chip@ateng.UU.NET" or "codas!ateng!chip" A T Engineering My employer may or may not agree with me. "I must create a system or be enslaved by another man's." -- Blake
donegan@stanton.TCC.COM (Steven P. Donegan) (04/16/88)
In article <181@ists>, mike@ists (Mike Clarkson) writes: > In article <13@stanton.TCC.COM>, donegan@stanton.TCC.COM (Steven P. Donegan) writes: > > > > I have elm 1.7 working quite well on my SCO 2.2.1 (286) system. The only > > problems I have noted so far are: > > > > 2) If a user is in elm in some functions and gets disconnected the session > > remains running - SIGHUP appears to have been defused. In one of it's source modules elm 1.7 traps signal SIGHUP, SCO Professional and SCO FOXPLUS also have this 'feature'. I expect it has to do with some misguided sense of file integrity. When I have tested my 'fix' to the source I will be happy to post. The problem with elm 1.7 was not reported by me, but by my local xenix guru, news feed and friend Jack Vogel of Turnkey Computer Consultants. -- Steven P. Donegan Sr. Telecommunications Analyst Western Digital Corp. donegan@stanton.TCC.COM
jack@turnkey.TCC.COM (Jack F. Vogel) (04/18/88)
In article <18@stanton.TCC.COM> donegan@stanton.TCC.COM (Steven P. Donegan) writes: >In article <181@ists>, mike@ists (Mike Clarkson) writes: >> In article <13@stanton.TCC.COM>, donegan@stanton.TCC.COM (Steven P. Donegan) writes: [deleted text...] >> > 2) If a user is in elm in some functions and gets disconnected the session >> > remains running - SIGHUP appears to have been defused. > >In one of it's source modules elm 1.7 traps signal SIGHUP, SCO Professional >and SCO FOXPLUS also have this 'feature'. I expect it has to do with some >misguided sense of file integrity. For all of those interested the problem here is in the elm 1.7 source file called syscall.c. It only happens in certain circumstances because the SIGHUP signal is only ignored while in a function called sys_call(). The comments in the call indicate that the signal is ignored for sendmail purposes so I should expect that simply commenting out the line : signal(SIGHUP,SIG_IGN); should take care of the problem. However, I have not had time to look over the code in any detail, so cannot vouch for all the ramifications of doing this, nor have I been able to test it yet. Good luck all, -- Jack F. Vogel Turnkey Computer Consultants, Costa Mesa, CA UUCP: ...{nosc|uunet}!turnkey!jack Internet: jack@turnkey.TCC.COM
woof@hpfcda.FC.HP.COM (Steve Wolf) (04/19/88)
> The comments > in the call indicate that [SIGHUP] is ignored for sendmail purposes I added this fix. The problem was that if a user composed a mail message, sent it, and then logged out, the 'sendmail' child process would catch the SIGHUP and quit. The mail message would be lost. If you don't use 'sendmail', it is probably safe to take out the signal() line. Steve Wolf Hewlett-Packard UUCP: {ihnp4|hplabs}!hpfcla!woof Fort Collins, CO ARPA: woof%hpfcda@hplabs.HP.COM