Contains the value for the “stop_when_done” property for newly created action points. This property defines additional elements to stop when a barrier point is satisfied or a thread encounters this action point. You can also set this value using the
When barrier hit, stop value in the
Action Points Page of the
File > Preferences dialog box. The values are:
Contains the default value used when a barrier point is satisfied. You can also set this value using the -stop_when_done command-line option or the
When barrier done, stop value in the
Action Points Page of the
File > Preferences dialog box. The values are:
Contains the control group for the process with the TotalView ID dpid. Setting this variable moves process
dpid into a different control group. For example, the following command moves process 3 into the same group as process 1:
^A: Moves the cursor to the beginning of the line
^B: Moves the cursor one character backward
^D: Deletes the character to the right of cursor
^E: Moves the cursor to the end of the line
^F: Moves the cursor one character forward
^K: Deletes all text to the end of line
^N: Retrieves the next entered command (only works after
^P)
^P: Retrieves the previously entered command
^U: Deletes all text from the cursor to the beginning of the line
Rubout or
Backspace: Deletes the character to the left of the cursor
This example shows two pairs, each delimited by a colon (“:”). Each element within a pair is delimited by any character except a colon. The first character entered is the delimiter. This example uses a “
+” as a delimiter. (Traditionally, forward slashes are used as delimiters but are not used here, as a forward slash is also used to separate components of a pathname. For example,
/home/my_dir contains forward slashes.)
This expression applies a mapping so that a directory named /nfs/compiled/u2/project/src1 in the expanded search path becomes
/nfs/host/u2/project/src1.
Elements that follow are either pids (for process groups) or
pid.tid pairs (for thread groups).
The gid is a simple number for most groups. In contrast, a lockstep group’s ID number is of the form
pid.tid. Thus,
GROUP(2.3) contains the lockstep group for thread 3 in process 2. Note, however, that the CLI does not display lockstep groups when you use
dset with no arguments because they are hidden variables.
Contains a list of information associated with a dpid. This is a read-only value and cannot be set.
Defines the CLI prompt. Any information within brackets ([ ]) is assumed to be a Tcl command, so therefore evaluated before the prompt string is created.
Contains the group ID of the share group for process pid. The share group is determined by the control group for the process and the executable associated with this process. You cannot directly modify this group.
Indicates the scope for newly created action points. In the CLI, this is the dbarrier,
dbreak, and
dwatch commands. If this boolean value is
true, newly created action point are shared across the group; if
false, a newly created action point is active only in the process in which it is set.
As an alternative to setting this variable, you can select the Plant in share group check box in the
Action Points Page in the
File > Preferences dialog box. To override this value in the GUI, use the
Plant in share group checkbox in the
Action Point > Properties dialog box.
Contains a list of all threads in the process pid, in the form
{pid.1 pid.2 ...}. This is a read-only variable and cannot be set.