[comp.mail.elm] I think I might have found a bug

iacovou@cs.umn.edu (Danny Iacovou) (08/31/90)

	I recently upgraded ELM2.3 from PL 4 to PL 6.  I found 2
   interesting "bugs?". 

        The first one is (well ok so this one has a part a and b)
   that when I did a 'configure -d' two items changed in my config.
   set up.
        a) my domain name changed from .cs.umn.edu to .UUCP
           (it's a good thing I noticed this one right off)
        b) the shell changed from the 'C' shell to the 'Korn' shell
           (is Korn shell THAT much better?)

   I had no problem with that "bug?".  What I really care about is
   that now when a user types the magic letters elm for the first
   time and answers yes to the question about making a .elm dir.,
   and even though a .elm dir. is made, no elmrc file is put in it.
   
   Anyone else get this problem? 


-- 

neophytos iacovou                                
university of minnesota                      email:  iacovou@cs.umn.edu 
computer science department

syd@DSI.COM (Syd Weinstein) (08/31/90)

iacovou@cs.umn.edu (Danny Iacovou) writes:
>        The first one is (well ok so this one has a part a and b)
>   that when I did a 'configure -d' two items changed in my config.
>   set up.
>        a) my domain name changed from .cs.umn.edu to .UUCP
>           (it's a good thing I noticed this one right off)
Check you config.sh file, the domain on a -d comes from there.  If it
choose .UUCP, then it was .UUCP in there or '' in there.

>        b) the shell changed from the 'C' shell to the 'Korn' shell
>           (is Korn shell THAT much better?)
This comes from the SHELL environment variable while running Elm, 
or from the config.sh file.

>   I had no problem with that "bug?"
neither is a bug.

>   What I really care about is
>   that now when a user types the magic letters elm for the first
>   time and answers yes to the question about making a .elm dir.,
>   and even though a .elm dir. is made, no elmrc file is put in it.
That is correct and expected behavoir, an elmrc file is not needed
unless the defaults are being overridden.  One will be created if the
options menu is chosen and the options saved.
-- 
=====================================================================
Sydney S. Weinstein, CDP, CCP                   Elm Coordinator
Datacomp Systems, Inc.                          Voice: (215) 947-9900
syd@DSI.COM or dsinc!syd                        FAX:   (215) 938-0235

pmm@mips.COM (Paul M. Moriarty) (09/01/90)

In article <1990Aug31.002402.8495@DSI.COM> syd@DSI.COM writes:
>iacovou@cs.umn.edu (Danny Iacovou) writes:
>
>>   What I really care about is
>>   that now when a user types the magic letters elm for the first
>>   time and answers yes to the question about making a .elm dir.,
>>   and even though a .elm dir. is made, no elmrc file is put in it.
>That is correct and expected behavoir, an elmrc file is not needed
>unless the defaults are being overridden.  One will be created if the
>options menu is chosen and the options saved.

It may not be a bug, but it certainly is a nuisance.  It would be a lot
nicer if I could just direct a user to his/her elmrc and tell them to
modify to their heart's content rather than first telling them how to
create a default elmrc.  

Why should the program say that the .elm directory is necessary
immediately upon invoking Elm for the first time when it isn't planning
on putting anything in the dir until you either create an alias or save
off your elmrc from the options window?  Seems to me that it shouldn't
prompt you for creating the directory until it is actually needed.

I'd strongly suggest that the default elmrc be written if .elm is being
created or if .elm exists and elmrc doesn't.


-- 
Paul M. Moriarty               pmm@mips.com          {ames,decwrl}!mips!pmm 
Sr Systems Administrator
MIPS Computer Systems, Inc                           +1 408 524 8335

fitz@wang.com (Tom Fitzgerald) (09/06/90)

pmm@mips.COM (Paul M. Moriarty) writes:
> It would be a lot
> nicer if I could just direct a user to his/her elmrc and tell them to
> modify to their heart's content rather than first telling them how to
> create a default elmrc.  

People who want to muck with their elmrc's better have a good understanding
of how it works, or they can badly burn themselves.  I'd say let them learn
how the defaults work before letting them modify anything.  It's easy
enough to create a default elmrc.

> I'd strongly suggest that the default elmrc be written if .elm is being
> created or if .elm exists and elmrc doesn't.

I'd just as soon keep things the way they are.  If elm 2.4 or 3.0 or
whatever renames some option fields, or changes an option default, I
don't want to have to modify the elmrc's of every user on this system.
Any user who _has_ built a new elmrc should change it on his own, but
presumably those people know what they're doing.

---
Tom Fitzgerald   Wang Labs        fitz@wang.com
1-508-967-5278   Lowell MA, USA   ...!uunet!wang!fitz

scs@lokkur.dexter.mi.us (Steve Simmons) (09/08/90)

pmm@mips.COM (Paul M. Moriarty) writes:
> It would be a lot
> nicer if I could just direct a user to his/her elmrc and tell them to
> modify to their heart's content rather than first telling them how to
> create a default elmrc.  

fitz@wang.com (Tom Fitzgerald) writes:
>People who want to muck with their elmrc's better have a good understanding
>of how it works, or they can badly burn themselves.  I'd say let them learn
>how the defaults work before letting them modify anything.  It's easy
>enough to create a default elmrc.

pmm@mips.COM (Paul M. Moriarty) writes:
> I'd strongly suggest that the default elmrc be written if .elm is being
> created or if .elm exists and elmrc doesn't.

fitz@wang.com (Tom Fitzgerald) writes:
>I'd just as soon keep things the way they are.  If elm 2.4 or 3.0 or
>whatever renames some option fields, or changes an option default, I
>don't want to have to modify the elmrc's of every user on this system.
>Any user who _has_ built a new elmrc should change it on his own, but
>presumably those people know what they're doing.

I disagree with most of the stuff written so far about the elmrc, and
these two sum up the suggestions pretty well.  My two cents:

Admins should be able to configure elm on a site-wide basis with a
master elmrc file.  Users should be able to override the site stuff
with their own elmrcs.

The stuff about users being able to handle their own elmrcs isn't so.
If the naive user changes one simple thing (such as changing his editor
from vi to jove), all of his settings are saved.  This is a problem two
ways.  First, it causes exactly the problem fitz describes.   Second,
the options displayed in the options are incomplete, so the naive user
can't change them.  We ought to (a) get *all* the options in there, and
(b) when writing an elmrc file, save only those options which are different
from the system default file.