diff options
Diffstat (limited to 'doc/clock.n')
| -rw-r--r-- | doc/clock.n | 37 |
1 files changed, 28 insertions, 9 deletions
diff --git a/doc/clock.n b/doc/clock.n index e609e3c..0823c0e 100644 --- a/doc/clock.n +++ b/doc/clock.n @@ -10,7 +10,7 @@ '\" See the file "license.terms" for information on usage and redistribution '\" of this file, and for a DISCLAIMER OF ALL WARRANTIES. '\" -'\" RCS: @(#) $Id: clock.n,v 1.11 2002/04/22 22:41:46 hobbs Exp $ +'\" RCS: @(#) $Id: clock.n,v 1.12 2003/04/12 19:08:54 kennykb Exp $ '\" .so man.macros .TH clock n 8.4 Tcl "Tcl Built-In Commands" @@ -31,16 +31,35 @@ out by the command. The legal \fIoptions\fR (which may be abbreviated) are: .TP .VS 8.3 -\fBclock clicks\fR ?\fB\-milliseconds\fR? -Return a high-resolution time value as a system-dependent integer -value. The unit of the value is system-dependent but should be the -highest resolution clock available on the system such as a CPU cycle -counter. If \fB\-milliseconds\fR is specified, then the value is -guaranteed to be of millisecond granularity. -This value should only be used for the relative measurement -of elapsed time. +\fBclock clicks\fR ?\fI\-option\fR? +If no \fIoption\fR argument is supplied, returns a high-resolution +time value as a system-dependent integer value. The unit of the value +is system-dependent but should be the highest resolution clock available +on the system such as a CPU cycle counter. .VE 8.3 .TP +.VS 8.5 +If the \fIoption\fR argument is \fB\-milliseconds\fR, then the value is +guaranteed to be an approximate count of milliseconds returned as +a wide integer; the rule should +always hold that \fBclock clicks -milliseconds\fR divided by 1000 is the +same as \fBclock seconds\fR. +.TP +It the \fIoption\fR argument is \fB-microseconds\fR, then the value is +guaranteed to be an approximate count of microseconds returned as a wide +integer; the rule should hold that \fBclock clicks -microseconds\fR +divided by 1000 is the same as \fBclock clicks -milliseconds\fR. +.TP +On some hardware, the counts of milliseconds and microseconds may diverge +from the system clock for short periods; the reason is that they can +be derived from different sources, and a complex procedure is required +to calibrate them. Moreover, Tcl makes an effort never to have the clock +leap forward nor appear to run backward, preferring instead to slow or +speed up the clock frequency slightly until it's back in synchronization. +For this reason, most Tcl programmers need never worry about such +phenomena as leap seconds. +.VE 8.5 +.TP \fBclock format \fIclockValue\fR ?\fB\-format \fIstring\fR? ?\fB\-gmt \fIboolean\fR? Converts an integer time value, typically returned by \fBclock seconds\fR, \fBclock scan\fR, or the \fBatime\fR, \fBmtime\fR, |