Re: [PATCH 3/3] gitk: Allow displaying time zones from author and commit timestamps

> Anders Kaseorg <andersk@MIT.EDU> writes:
> > On Sun, 29 May 2011, Paul Mackerras wrote:
> 
> > > > @@ -11018,7 +11018,18 @@ proc prefsok {} {
> > > >  proc formatdate {d} {
> > > >      global datetimeformat
> > > >      if {$d ne {}} {
> > > > -	set d [clock format [lindex $d 0] -format $datetimeformat]
> > > > +	if {[string match {*%[zZ]*} $datetimeformat]} {
> > > > +	    if {[catch {set d [clock format [lindex $d 0] -timezone [lindex $d 1] -format $datetimeformat]}]} {
> > > > +		# Tcl < 8.5 does not support -timezone.
> > > > +		global env
> > > > +		set zone [lindex $d 1]
> > > > +		set env(TZ) "IDK[string range $zone 0 2]:[string range $zone 3 end]"
> > > 
> > > What is this about?  Where is the IDK prefix for timezones defined or
> > > described?
> > 
> > Yeah, sorry, that deserved a bit more explanation.  This is a kludge to 
> > get Tcl 8.4 to format dates in the right timezone.  IDK is an arbitrary 
> > made up 3-letter code (“I Don’t Know”), since a UTC offset can’t generally 
> > be converted into a zone name.  The format of TZ is described at:
> > http://www.gnu.org/software/libc/manual/html_node/TZ-Variable.html
> > http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap08.html
> > 
> > Actually, reading that again, I just realized that I need to invert the 
> > sign.  Also that it can be longer than 3 letters; I’ll use “Unknown” 
> > instead.
> 
> Why not use UTC+N timezone (note: please check of for +HHMM it is
> UTC+HH or UTC-HH) for timezone with given numeric offset from
> Coordinated Universal Time?
> 
> BTW. UTC because http://www.nist.gov/pml/div688/utcnist.cfm#cut