home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.unix.admin:4646 comp.windows.x:15524 comp.windows.x.apps:829
- Newsgroups: comp.unix.admin,comp.windows.x,comp.windows.x.apps
- Path: sparky!uunet!iWarp.intel.com|ichips!tmcconne
- From: tmcconne@sedona.intel.com (Tom McConnell)
- Subject: SUMMARY: POLL: What is in X/shell files? (REPOST)
- Message-ID: <1992Aug23.002433.18628@ichips.intel.com>
- Originator: tmcconne@sedona
- Keywords: SUMMARY POLL
- Sender: tmcconne@sedona (Tom McConnell)
- Reply-To: tmcconne@sedona.intel.com
- Organization: Intel Corporation
- Date: Sun, 23 Aug 1992 00:24:33 GMT
- Lines: 184
-
-
- [This is a repost. The first attempt, several days ago, seems to have lost.]
-
- The short version:
-
- 1. Put the environment information in .cshrc.
- 2. Put the shell variables and aliases in .cshrc.
- 3. Put terminal setup and console setup in .login.
- 4. Put terminal setup in X resources (ie., .Xdefaults) for xdm.
-
- The long version:
-
- These are the results of the poll of the use of environment files. The
- original questions are these:
-
- 1. What do you use (what is in) .login for?
-
- 2. What do you use (what is in) .cshrc for?
-
- 3. What do you use (what is in) .xinitrc for?
-
- 4. What do you use (what is in) .xsession for?
-
- 5. Do you normally source a separate file for environment variables or
- aliases?
-
- I appreciate the people who participated in this, and I would like to
- thank you for your time.
-
- Results
- -------
-
- The results were, at best, mixed. My original reason for conducting the
- poll in the first place was to determine, in a multi-architecture,
- multi-environment X world, where to put things like path and stty.
-
- There is not agreement, for instance, on where to put environment
- variables. Twelve (12) of the responses used .cshrc, and fourteen (14)
- used .login. Nearly everyone used .login for stty (terminal setup), and
- nearly everyone used .cshrc for "set" and "alias."
-
- The disagreement on where to set path seems to be a result of xdm and
- rsh, neither of which automatically invokes .login. In order to ensure
- that a path is available, the path was often set in .cshrc. For the
- .xsession file under xdm, a common response was to source .login, or just
- to duplicate the information from .login in the .xsession file.
-
- From these responses, I would say that the "best" approach is to use
- what works (brilliant, isn't it :). Here are some general guidelines
- that seem to be in common use:
-
- 1. Most people tend to have a small set of files, with one or two
- additional files that are sourced for environment or alias information.
-
- 2. If rsh is going to be used, the general solution seems to be to
- put the environment information in .cshrc.
-
- 3. If xdm is going to be used, either put all of the environment
- information in .xsession, source a "environment variables" file, or
- put environment information in .cshrc.
-
- From numbers 2 and 3, it seems the most prudent approach is to put
- the environment information in .cshrc. .cshrc also has the shell
- variables and the aliases. This leaves terminal setup and
- console-specific setup for .login. When using xdm, terminal information
- can be set up in the X resources (thanks to Bruno Berstel for this
- example):
-
- * The .Xresources files
-
- ...
- # We know it's Delete because of the xmodmap in .x{initrc,session}
- XTerm*ttyModes: erase ^? kill ^U intr ^C
- ...
- # You should try that : when you select a chunk of text in a window
- # with the shift key down, it's automagically pasted on the command
- # line without any other button press.
- XTerm*VT100.translations: \
- Meta<KeyPress> : string(0x1b) insert() \n\
- <KeyPress> : insert()
- ...
-
- Again, thank you for your time. If anyone would like a copy of the
- responses, I would be glad to send them to you. I would prefer to send a
- uuencoded, compressed tar file of the responses, but any form is fine
- with me. Please send a request via email to:
-
- tmcconne@sedona.intel.com
-
- Cheers,
-
- Tom McConnell
-
- Tom McConnell | Internet: tmcconne@sedona.intel.com
- Intel, Corp. C3-21 | Phone: (602)-554-8229
- 5000 W. Chandler Blvd. | The opinions expressed are my own. No one in
- Chandler, AZ 85226 | their right mind would claim them.
-
-
- Comments
- --------
-
- I only tally the responses that are explicit. For instance, if a response
- says:
-
- Everything is in .login
-
- I only count a response as "everything." I do _not_ add to the number of
- .login responses.
-
- Most of the responses had some commentary about why certain things
- were "good" or "bad." For instance, Bruno Berstel provided a very
- nice background for environment issues in general.
-
- A number of the responses noted that they do not use C shell, but
- some other shell. zsh and the Bourne shell were common. As a result,
- the number of responses for .login and .cshrc are somewhat subjective,
- based on comments in the response.
-
- Several responses indicated that they used global, site, or the
- default file. Global alias files and /usr/lib/X11/xdm/Xsession are
- examples. I have seen differences in Xsession at different sites, so
- I have made no attempt to determine what "should" be in Xsession.
-
- Results
- -------
- Total number of responses: 28
-
- Environment variables in .cshrc: 12
- Environment variables in .login: 14
-
- Everything in .cshrc: 4
- Everything in .login: 1
-
- Terminal setup (TERM, stty, tset) in .cshrc: 3
- Terminal setup (TERM, stty, tset) in .login: 19
-
- 4 responses with no .login file.
- 1 response with no .cshrc file.
-
- Response where .xinitrc is the same as .xsession: 10
- of these> set path in .login: 3
- set path in .cshrc: 3
-
- .xsession sets up environment directly: 8
- .xsession sources a file: 2
- .xsession uses "xterm -ls": 1
-
- Participants
- ------------
- Tom McConnell <tmcconne@sedona.intel.com>
- Steven Ball <steve@cs.anu.edu.au>
- Mark Watson <mark@wehi.edu.au>
- Jim Fulton <lupine!verbosa!jim@uunet.UU.NET>
- Ian Fitchet <I.D.Fitchet@cs.bham.ac.uk>
- Andy Stevens <stevens@madvlsi.columbia.edu>
- Thomas Richter <richter@ki1.chemie.fu-berlin.de>
- Paul Davey <pd@x.co.uk>
- Donn Cave <donn@cac.washington.edu>
- Bruno Berstel <berstel@lacim.uqam.ca>
- Ami Kleinman <ami@shum.huji.ac.il>
- Amos Shapira <amoss@CS.HUJI.AC.IL>
- Tom Tatlow <tatlow@athena.lmo.dec.com>
- Mark Anderson <mark@cs.ucsd.edu>
- John G Dobnick <jgd@uwm.edu>
- Matt Wette <mwette@jpl-gnc-gw.Jpl.Nasa.Gov>
- Richard Pieri <ratinox@meceng.coe.northeastern.edu>
- Daniel S. Lewart <d-lewart@uiuc.edu>
- Niclas Wiberg <nicwi@isy.liu.se>
- Nick Smith <nick@garland.bt.co.uk>
- Steve Rumsby <steve@maths.warwick.ac.uk>
- Dan Connolly <connolly@pixel.convex.com>
- Curt Schroeder <curts@tmpcu.mdc.com>
- Andrew Stoker <mac742@lancaster.ac.uk>
- Bill Trost <trost@reed.edu>
- Jon Marthins <jon@dikas.no>
- Daniel MacKay <daniel@nstn.ns.ca>
- M. Eyckmans <eyckmans@imec.be>
-
- --
- Tom McConnell | Internet: tmcconne@sedona.intel.com
- Intel, Corp. C3-21 | Phone: (602)-554-8229
- 5000 W. Chandler Blvd. | The opinions expressed are my own. No one in
- Chandler, AZ 85226 | their right mind would claim them.
-