Technical Difficulties from on Top of the Mountain
2017-04-17
  Technological Stratum
If you're going to hitch your wagon to an evolutionary dead end, you better be good at fixing a flat.

geologic strata In the beginning (for the purposes of our story) was Unix and C. And it was good. You had your shell /bin/sh, the wonderful /bin/[ for all your filesystem commands, and anything else you just wrote a program for.

But as soon as there were more than a few users on a machine sharing a common login script, things got a little complicated. Add in the fact that Unix was designed to be portable, and you had scripts that were running in slightly different environments on different machines, and the need for some logic in these scripts became obvious.

For no understandable reason, the systax for these basic logic statements was bizarre,

    if [ "$uid" -ge 100 ]
    then
      echo "Mere mortal"
    fi

As anyone who's moved from rails to node will tell you, having one syntax beats switching your mental model back and forth constantly. And it was no different back then. Programmers have been universally lazy since the dawn of time--that's why we have so many cool toys. So finally when one particularly impatient programmer had had enough, he decided that there should be a shell that used the same sort of syntax for control that the C programming language used. Thus was born the C shell, or csh for short.

Unfortunately that programmer went on to write many other interesting things, like VI, NFS and Java; and the shell was closely associated with Berkeley BSD, so it did not spread out as fast as a competitor written by David Korn at Bell Labs which became part of the System V distribution widely licensed or copied by such workstation vendors as HP, DEC, IBM and SGI.

Finally as the personal computer advanced into the Unix domain, Brian Fox at the GNU Project wrote a open source shell called Bash which again used a compatible syntax with bourne shell.

Not that csh didn't get any love, Ken Greer had been working on a TENEX like file completition library, and finally patched it into a version of csh which he called tcsh. It gained additional following when it was made the default shell for Mac OS X originally, though Apple later gave into popular pressure and changed the default to bash.

By this time however, machines were faster (megahertz clock speeds), bigger (megabyte memories, gigabyte hard drives), and could support more cycles catering to the user with actual interpreted languages which grew in leaps and bounds. In short order you had your choice of perl, python, tcl, php, ruby and more. And there was no longer any good reason to try and do anything serious in a shell script. A few people kept trying, but nobody took them seriously.

But there are those of us who had to get something done back before the dust had settled, and we picked the best tool we could, and we got as good at is as we good. And once you have your login script which customizes things you want it, there's really not a good reason to change it.

c shell field guide

So I still use csh, well tcsh so I can tab complete things once in a while, but my .cshrc was written many years ago, and except for changing the names of a few directories, its stayed about the same. But I know I'm out on the fringe, and someday the last of us will logoff, and tcsh will be nothing more that a curiosity that computer re-enactment history buffs play with on the weekends.

And when something breaks, your're mostly on your own.

Like it did a few weeks back at work.

I fired up my terminals for the week, like I do every Monday, but instead of showing my a welcoming prompt, the cursor just sat there like the machine was dead. Being the impatient programmer that I was, I hit the interrupt key (^C for mortals), and was rewarded with a prompt. Hmmm. Everything else looked normal, so I interrupted all my terminals and went on about my business.

There was more though. I went to kill a sub-tab in urxvt, and it just stayed there. Wasn't taking input anymore, but it wasn't cleaning up and exiting. Another interrupt character, and its gone. But its getting more annoying. The pattern kept repeating itself: create a new sub-tab, interrupt to get it ready; close the tab, interrupt to get it to go away.

Next week I come in, Monday morning and the machines rebooted, so I have to interrupt all my windows to wake them up. But I have other fish to fry, so I just get around it and go about my business.

Third week begins. I login--same situation. Ok, my extreme laziness now takes effect. Time to figure out what's going on.

First I hunt down all the scripts that the shell executes when it logs in and logs out. There's a different list for login shells, and non login shells are behaving just fine for me. Check out the scripts, run each of them manually, that's not the problem.

We're going to have to poke into the guts of the shell and try to catch it while its stuck. Back in the early days, the tool of choice was dbx, even for programs without debugging symbols or source. You could still observe basic call chains and if you were clever enough, even follow along in the assembly as the program went from one state to another. There was this one time at a client's site, a third party piece of software wouldn't start because it was trying to read an inialization file, and we had the file, we just didn't know where the program wanted it, so I stuck the program in dbx and put a breakpoint on fopen() and low and behold it was looking in /var/spool/x and so we moved the file there and it worked, and the customer thought I was pretty clever--but I digresss. Thankfully there are some better, more pointed tools at our disposal now.

I fire up a new instance of csh, because I already have two dozen instances of tcsh running in all my windows and I'm too lazy to isolate one to test with. Then I find its process ID and feed it to strace(1), which gives me this nugget of information:

    fcntl(6, F_SETLKW, {type=F_WRLCK, whence=SEEK_SET, start=0, len=0}

which is gibberish for stuck locking a file. Still don't know what file is being locked, but its a file that's being accessed at both login and logout.

A bit of creative searching lands me here,

https://rhn.redhat.com/errata/RHBA-2012-0687.html

Addressed in this update is the following issue:

When using multiple shells simultaneously, the command history is saved from all shells in one ".history" file ... This update implements file locking mechanism that uses shared readers and an exclusive writer to prevent the ".history" file from being corrupted.
Hmmm. Ok, I have a .history file, its zero bytes.

    rm -f .history

Problem solved.

 
Life in the middle of nowhere, remote programming to try and support it, startups, children, and some tinkering when I get a chance.

ARCHIVES
January 2004 / February 2004 / March 2004 / April 2004 / May 2004 / June 2004 / July 2004 / August 2004 / September 2004 / October 2004 / November 2004 / December 2004 / January 2005 / February 2005 / March 2005 / April 2005 / May 2005 / June 2005 / July 2005 / August 2005 / September 2005 / October 2005 / November 2005 / December 2005 / January 2006 / February 2006 / March 2006 / April 2006 / May 2006 / June 2006 / July 2006 / August 2006 / September 2006 / October 2006 / November 2006 / December 2006 / January 2007 / February 2007 / March 2007 / April 2007 / June 2007 / July 2007 / August 2007 / September 2007 / October 2007 / November 2007 / December 2007 / January 2008 / May 2008 / June 2008 / August 2008 / February 2009 / August 2009 / February 2010 / February 2011 / March 2011 / October 2011 / March 2012 / July 2013 / August 2013 / September 2013 / October 2013 / November 2013 / December 2013 / December 2014 / February 2015 / March 2015 / July 2016 / September 2016 / December 2016 / April 2017 / June 2017 /


Blogroll
Paul Graham's Essays
You may not want to write in Lisp, but his advise on software, life and business is always worth listening to.
How to save the world
Dave Pollard working on changing the world .. one partially baked idea at a time.
SnowDeal
Eric Snowdeal IV - born 15 weeks too soon, now living a normal baby life.
Land and Hold Short
The life of a pilot.

The best of?
Jan '04
The second best villain of all times.

Feb '04
Oops I dropped by satellite.
New Jets create excitement in the air.
The audience is not listening.

Mar '04
Neat chemicals you don't want to mess with.
The Lack of Practise Effect

Apr '04
Scramjets take to the air
Doing dangerous things in the fire.
The Real Way to get a job

May '04
Checking out cool tools (with the kids)
A master geek (Ink Tank flashback)
How to play with your kids

Powered by Blogger