Hi, First, we don't use zic; we use our own .NET tzinfo parser. I would expect zic doesn't have the problem we've encountered. The most recent change of the Antarctica file (2014-03-21, commit 6285081) introduced three leading spaces before the three leading tab characters at line 267 (second line of the Antarctica/Troll zone definition). Because of the data before the first tab character, our system has an off-by-one error parsing the rest of the line, and doesn't recognize it as a zone definition line. What is the rule regarding continuation lines on zone definitions? We split lines into arrays by tab, which fails in this case where there are leading spaces before the first tab. Since this works in all other cases, is this an error in the data? Or should we follow a different semantic rule? Thanks. David Braverman Chicago
David Braverman wrote:
What is the rule regarding continuation lines on zone definitions?
It's given in the zic man page, which says "Fields are separated from one another by one or more white space characters." Tabs and spaces are both white space characters, and it's immaterial which is used. The other white space characters (other than newline itself) are '\f', '\r', and '\v'.
So, how would zic interpret "<space><space><space>\t\t\t", as on the Antarctica/Troll zone? David Braverman -----Original Message----- From: Paul Eggert [mailto:eggert@cs.ucla.edu] Sent: Sunday 4 May 2014 13:59 To: David Braverman; 'tz@iana.org' Subject: Re: [tz] Troll rule parsing issue David Braverman wrote:
What is the rule regarding continuation lines on zone definitions?
It's given in the zic man page, which says "Fields are separated from one another by one or more white space characters." Tabs and spaces are both white space characters, and it's immaterial which is used. The other white space characters (other than newline itself) are '\f', '\r', and '\v'.
participants (2)
-
David Braverman -
Paul Eggert