Agent Tcl: Hints and snags

Last updated on 1 December 1995

This page lists the known bugs as well as some useful, undocumented features.

Undocumented features

• New flag -noPipeError for the exec command

  exec can be used to create a pipeline. For example,

    exec last rgray | head -1

  pipes the output of the last command through the head command. The problem with this example is that the head command finishes before the last command (last produces hundreds of lines of output but head just echoes the first line and quits immediately). The last command then exits abnormally because it is trying to write to a disconnected pipe. exec fails with the error message

    child killed: write on pipe with no readers

  You can force exec to ignore this pipe error with the -noPipeError flag. The example should be rewritten as

    exec -noPipeError last rgray | head -1

  so that it no longer fails.

Known bugs

• Masks are not transferred when an agent jump or forks. You will have to recreate the masks after a call to agent_jump or agent_fork. This problem is being fixed.

• The restrict command will break if the enclosed script contains an agent_jump or agent_fork command. This problem is being fixed.

• You can not use agent_fork or agent_jump inside a Tk event handler. This problem is being fixed. In the meantime set up the tkwait so that the tkwait will return if the agent wants to jump.

• Tcl gets confused if there are unmatched curly brackets inside a comment. For example, the following Tcl fragment causes an error.

  while {1} {
    # {
  	puts HELLO
  }

  This is a long-standing problem with Tcl itself and will not be fixed as part of the Agent Tcl project. Do not use curly brackets inside a comment and you will be all set.

• gets, puts and read can be interrupted unexpectedly
  Internally Agent Tcl uses interrupts to notify the agent interpreter that a message from another agent has arrived. Unfortunately, if this interrupt occurs in the middle of a gets, puts or read command, the command will fail with the error message

    Interrupted system call

  and will set the errorCode variable to

    POSIX EINTR ...

  This problem is being fixed. A temporary workaround is to reissue the command if the errorCode variable starts with "POSIX EINTR". For example:

      # loop until there is no error
  
    while {[catch {gets $fd} line]} {
  
        # an error has occured so see if errorCode starts with POSIX EINTR
        # loop around and issue the command again if errorCode starts with POSIX EINTR
        # otherwise some other error has occurred and we must handle it
  
      if {![string match "POSIX EINTR*" $errorCode]} {
  
          # handle other errors here
  
      }
    }
  
      # now the next line in the file is in the variable "line"

  This code fragment can be placed inside a procedure so that you do not have to type it over and over. Note that you do not need to worry about interrupts unless the agent is currently registered with a server, i.e., you only need to check for POSIX EINTR inside an agent_begin/agent_end pair.

• Lost upvar references
  This bug has been fixed.

• Lost upvar references to the env array
  An upvar reference to an element of the env array is lost when an agent migrates and is not present in a forked child. Therefore you must recreate the upvar reference after a call to agent_jump or agent_fork. For example, the procedure

    proc run_around machines {
  
      upvar #0 env(DISPLAY) display
  
      set list ""
  
      foreach m $machines {
        agent_jump $m
        append list "The current display is $display.\n"
      }
    }

  will fail with the error message

    can't read "display": no such variable

  The following procedure, however, will work correctly.

    proc run_around machines {
  
      set list ""
  
      foreach m $machines {
        agent_jump $m
        upvar #0 env(DISPLAY) display
        append list "The current display is $display.\n"
      }
    }

  This problem is being fixed.