On Thu, Aug 18, 2016 at 07:17:00PM +0200, Johannes Schauer wrote: > Hi, > > Quoting Santiago Vila (2016-08-18 15:01:53) > > However, the way I read this: > > > > https://en.wikipedia.org/wiki/ISO_8601#Combined_date_and_time_representations > > > > there are still two allowed ways to specify the dates and the times: > > the "basic format" and the "extended format". > > > > Both are standard and both are equally correct, but I wish there would > > be a way to use the basic format instead of the extended format, at least > > for the filenames of the build logs, if only because 20160818T113412Z > > is a lot easier to parse than 2016-08-18T11:34:12Z for further > > processing. > > I'm confused about what the problem is that you want to report. > > The current format of the datetime stamp in build log filenames is: > > 2016-08-18T17:09:36Z > > To parse that you write in most programming languages something like: > > strptime("2016-08-18T17:09:36Z", "%FT%TZ") > > To parse a string like: > > 20160818T113412Z > > You would write: > > strptime("20160818T113412Z", "%Y%m%dT%H%M%SZ") > > I find the former format string much easier and shorter than the one to parse > the latter. > > So how is the latter easier to parse?
Try doing that in standard AWK without those fancy time-handling functions :-) For the basic format it would be something like this: date=substr(stamp,1,8) time=substr(stamp,10,6) I leave the extended format as an exercise for the reader. It is in this sense that I said "easier to parse". Also: The extended format makes the filename a little bit longer and a little bit more difficult to handle, as the ":" has to be escaped. In either case: Nobody asked for a way to specify the filename format in a flexible way? For example: If I wanted an integer suffix meaning the number of seconds since the epoch, how I would do that? Thanks.
