Re: Names of Fields in Control Files
On Sun, Sep 26, 2010 at 01:35:00AM -0700, Russ Allbery wrote: Raphael Hertzog hert...@debian.org writes: On Sat, 25 Sep 2010, Jonathan Yu wrote: 22:02:40 rra jawnsy: I don't think we say that explicitly, but RFC 5322 requires it and I can't imagine ever not enforcing that. Although you should check with the dpkg maintainers to be sure. Could we/should we make the Debian Policy more restrictive, and specify explicitly that field names must only be ASCII-encoded? [...] Your comments and feedback on this would be much appreciated. I think this discussion is theoretical and useless. I hope nobody will suggest a field name containing non-ascii characters... I suspect there might be a communication problem that made this come across harsher than it was intended. But I'll mention that one of the things that's sometimes frustrating about trying to nail down the specification and standards around Debian's package format is that aspects of standardization that would be considered completely routine in, say, IETF work are considered theoretical and useless. If we were standardizing things in any other context, one of the very first things we'd do is write an ABNF grammar for Debian control fields, which would immediately and unambiguously state the allowed characters for each component. Ideally, there should be a proper dpkg interface documentation, and the policy document would only need to specify the subset that is mandated by policy. The current situation when the policy team is in charge of maintaining the dpkg interface documentation is awkward, especially since policy does not maintain the dpkg interface. Cheers, -- Bill. ballo...@debian.org Imagine a large red swirl here. -- To UNSUBSCRIBE, email to debian-dpkg-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20100929150556.gb10...@yellowpig
Re: Names of Fields in Control Files
On Sat, 25 Sep 2010, Jonathan Yu wrote: 22:02:40 rra jawnsy: I don't think we say that explicitly, but RFC 5322 requires it and I can't imagine ever not enforcing that. Although you should check with the dpkg maintainers to be sure. Could we/should we make the Debian Policy more restrictive, and specify explicitly that field names must only be ASCII-encoded? [...] Your comments and feedback on this would be much appreciated. I think this discussion is theoretical and useless. I hope nobody will suggest a field name containing non-ascii characters... I don't think dpkg does any special decoding when reading data. dpkg deals with bytes and not characters AFAIK. I'm certainly OK with policy requiring field names to be ASCII. Cheers, -- Raphaël Hertzog ◈ Debian Developer ◈ [Flattr=20693] Follow my Debian News ▶ http://RaphaelHertzog.com (English) ▶ http://RaphaelHertzog.fr (Français) -- To UNSUBSCRIBE, email to debian-dpkg-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20100926075312.gc22...@rivendell.home.ouaza.com
Re: Names of Fields in Control Files
Raphael Hertzog hert...@debian.org writes: On Sat, 25 Sep 2010, Jonathan Yu wrote: 22:02:40 rra jawnsy: I don't think we say that explicitly, but RFC 5322 requires it and I can't imagine ever not enforcing that. Although you should check with the dpkg maintainers to be sure. Could we/should we make the Debian Policy more restrictive, and specify explicitly that field names must only be ASCII-encoded? [...] Your comments and feedback on this would be much appreciated. I think this discussion is theoretical and useless. I hope nobody will suggest a field name containing non-ascii characters... I suspect there might be a communication problem that made this come across harsher than it was intended. But I'll mention that one of the things that's sometimes frustrating about trying to nail down the specification and standards around Debian's package format is that aspects of standardization that would be considered completely routine in, say, IETF work are considered theoretical and useless. If we were standardizing things in any other context, one of the very first things we'd do is write an ABNF grammar for Debian control fields, which would immediately and unambiguously state the allowed characters for each component. I'm certainly OK with policy requiring field names to be ASCII. I think that's probably the right thing to do. -- Russ Allbery (r...@debian.org) http://www.eyrie.org/~eagle/ -- To UNSUBSCRIBE, email to debian-dpkg-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87aan452dn@windlord.stanford.edu
Re: Names of Fields in Control Files
Le Sun, Sep 26, 2010 at 01:35:00AM -0700, Russ Allbery a écrit : Raphael Hertzog hert...@debian.org writes: I'm certainly OK with policy requiring field names to be ASCII. I think that's probably the right thing to do. Dear all, how about simply paraphrasing the RFC 822/5832, which our the source of inspiration ? In that case, the requirement for field names will be to be printable ASCII characters, except colons. I propose the following change in the context the patch that I am preparing for clarifying the Policy's chapter about control files, in bug #593909. diff --git a/policy.sgml b/policy.sgml index be0a505..5c72355 100644 --- a/policy.sgml +++ b/policy.sgml @@ -2493,9 +2493,11 @@ endif p Each paragraph consists of a series of data fields; each field consists of the field name, followed by a colon and - then the data/value associated with that field. It ends at - the end of the line or at the end of the last - continuation line (see below). Horizontal whitespace + then the data/value associated with that field. The field + name is composed of printable US-ASCII characters (i.e., + characters that have values between 33 and 126, inclusive), + except colon. The field ends at the end of the line or at + the end of the last continuation line (see below). Horizontal whitespace (spaces and tabs) may occur immediately before or after the value and is ignored there; it is conventional to put a single space after the colon. For example, a field might Cheers, -- Charles Plessy Tsurumi, Kanagawa, Japan -- To UNSUBSCRIBE, email to debian-dpkg-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20100926150735.gb31...@merveille.plessy.net
