I always assumed it was just the same literal T as in the ISO spec:
On Oct 16, 2016 14:02, "Rob Landley" <r...@landley.net> wrote:
> specifies -d as:
> -d date_time
> Use the specified date_time instead of the current time. The option
> argument shall be a string of the form:
> T is the time designator, and can be replaced by a single <space>.
> The problem is, it doesn't say what a "time designator" is. The two
> other hits for "designator" in this page are:
> If the T time designator is replaced by a <space> for the -d
> date_time option-argument, the <space> must be quoted to prevent
> the shell from splitting the argument.
> The -d date_time format is an ISO 8601:2000 standard complete
> representation of date and time extended format with an optional
> decimal point or <comma> followed by a string of digits following
> the seconds portion to specify fractions of a second. It is not
> necessary to recognize "[+/-]hh:mm" and "[+/-]hh" to specify
> timezones other than local time and UTC. The T time designator in
> the ISO 8601:2000 standard extended format may be replaced by
> The touch man page is useless (and points me at the info page, which I'm
> not reading).
> Googling for "posix time designator" brought up wikipedia pages on unix
> time (numeric seconds since 1970), an "epoch coverter" web page where
> you type in a time and it converts it, and then
> is the first possibly relevant hit (isn't it FUN how google mangles
> links when you right click copy link?) but it does not include the word
> "designator". So that's nice.
> The _examples_ on the posix page all use "T" for the "designator", so
> that's what I made my code do when I wrote it. But somebody complained
> (who I promised to stop listening to after they complained about my
> 'wall of text" instead of just taking the "simple" memcpy->memmove
> patch), and although I'm not replying directly to them a bug report's a
> bug report...
> I checked what busybox is doing, and they're funneling stuff through a
> large complicated library function that hardwires in a dozen different
> formats checking them in a lage if/else staircase, and among those are
> two that accept this format with "-" or " " as "designators". (But not
> "T" which is what the posix examples use...)
> Does anybody know what this field is supposed to _mean_? And/or have an
> actual use case in the wild using touch -d?
> Toybox mailing list
Toybox mailing list