[Top] [All Lists]

Notes for the meeting in Paris

2005-08-10 11:36:03

Hi everyone,
I've produced Paris meeting notes from the Jabber logs. Can people please verify them for correctness/completeness.


Thanks to Pete Resnick and Tony Finch for taking Jabber notes.

The WG had an 1:30 minutes long meeting on Tuesday, August 2nd. Several WG members have participated
through Jabber (Cyrus, Ned, Matthew, Ken)

WGLC for the base spec has ended. There were some minor editorial comments on the mailing list, in particular suggestion to change "Syntax:" to "Usage:" everywhere in the document. Philip has also pointed out that he has to change "header" to "header field" to use terminology consistent with RFC 2822.

Vacation draft is ready for WGLC. Jeffrey Hutzelman have done a security review and didn't find any issues.

Kjetil has talked about updated Variables draft. Recent changes include:
 allow sub-namespaces to be all-numeric (e.g. ${foo.1.2})
 allow to SET variables in namespaces that allow for modifications
${0} is now referencing the entire match, not the number of matching groups
 new modifiers :quotewildcard and :quoteregex.
One remaining issue is whether regex related text (i.e. :quoteregex) should be moved to the Regex draft. Alexey was trying to convince everybody that dependency on a document-which-might-take-arbitrary-long-time is a bad thing, in particular because Regex hasn't addressed I18N issues yet. General consensus from people in the room to move :quoteregex to Regex draft, Ned agrees in Jabber.

Q: Security review of Variables has pointed out issue with silent truncation of variables. A: There is a new text in Security Considerations section. Sorry, this should have been listed in the
"recent changes".
Philip has repeated the argument that Ned has maid on the mailing list: if there is a language facility to test if a truncation has occurred, nobody will use it anyway. Jeffrey has replied that he is Ok with not having a test for truncation, as long as the document discusses the issue. The room agrees.

Philip then talked about body draft. A new revision was published as a result of WGLC. Recent changes include:
Charset wording aligned with base spec
Clarified the implicit content-type rules.
New issue: what does it mean to match against non-textual content? Matching is defined for Unicode characters,
so what would it mean to convert a sound or image to Unicode?
Alexey has pointed out that IMAP SEARCH only required to work on text/* parts, suggested to do the same. Philip has mentioned that we could assume that bytes are in iso-8859-1. Many people in the room disagree. David Cridland has suggested that comparators could control which mime types could be searched with them. William Leibzon has pointed out that it would be desirable to be able to search a MIME prolog and epilogue,
several people seemed to agree.
Pete Resnick has asked Ned about default charset for prolog/epilog. Ned has replied that it is US-ASCII,
as defined in RFC 2822.
Cyrus (in Jabber) has suggested to use a separate MIME extension(s) to work with MIME bodyparts properly. Alexey has suggested that maybe we should drop :content, this will certainly make the document simpler.
Cyrus has suggested to leave :content as a simple/quick short time solution.
Philip has asked the attendees if anybody but Sendmail have implemented Body. David said that Exim might.

Philip then talked about Edit header. Philip thought that it should be ready for IETF wide LC, chairs agreed. Scott Hollenbeck reinforced that "writeup" meant that chairs were going to shepherd the document as per new
IESG rules. Chaired confirmed that they were aware of that.

Alexey has talked about Spamtest document, using the slide sent by Cyrus. Recent changes included: removing "-1" from :percent test and a new text explaining how to test that Spam check wasn't performed on the message. New capability "spamtestplus" was added as the result. The document should be ready for WGLC.

Alexey has talked about IMAPflags document. The latest revision got published in spring. It added back internal variable (so that the extension can be used without "Variables", e.g. in Cyrus Sieve), syntax is backward compatible with the old IMAPFlags (revision -03) as implemented by CMU. Cyrus has confirmed that the document should be ready
for WGLC, Ned agreed.

Alexey has talked about Refuse/Reject draft. He thinks that the document is ready for WGLC. Matthew (in jabber) has reminded him that there is one open issue: what should happen if "refuse" is called more than once for a message.
This needs to be discussed on the mailing list (can be done during WGLC).
Philip has said that he didn't like the text on joe jobs in the Introduction/Abstract sections. Alexey has asked for specific suggestions, which Philip has promised to provide on the mailing list. Cyrus should issue a WGLC
for the document.

Chris has presented Ned's Date extension. It adds date test. Can search for a date at the end of any header, so this works with Date, Received and even X400-recieved. Ned has added a comment in Jabber saying that the date stuff was structured the way it was to avoid requiring variables for everything. It is possible to test year, month, day, julian (day number on julian calendar), hour, minute, second, iso8601, or timezone. There is also a test for the current date/time. Ned has asked the WG if any test is missing or should be removed. Cyrus suggested to add "day-of-week". Somebody else has suggested "day-of-year".
Also seconds should be allowed to have value 60.
Tony Finch has suggested to use modified Julian day, which starts counting from the mid-19thc (1863 i think) and days start at midnight, not midday as in the Julian day. Ned agreed with the change.
Cyrus has brought the question about :zone and daylight savings.
Ned/Pete said that zones in the document are numeric, so this shouldn't be an issue. Ned also pointed out that if timezone is not explicitly specified, than a local timezone is assumed and
things should work as expected.
Alexey mentioned that the document is not in WG Charter, but it is Ok to discuss it on the mailing list.
Ned replied that he didn't really care.

Tony Hansen has presented Sieve loops (MIME parts tests). "for.every.part" construct allows to iterate
through the MIME body parts, there is a new "break" construct as well.
Inside the iterator loop the current message is the body part.
Alexey has asked if one can use "fileinto" on a body part. It is not currently clear, many other
Sieve extensions can also be meaningless.
The document also defines a new MIME test, which allows to examine MIME headers/parameters for the current part. If the test is used outside of a loop, the test return true if any MIME part has a particular
Two other extensions are also described in the draft: body part replacement action (it allows to replace a body part with a text/plain string) and enclose action (encapsulate the entire message in a message/rfc822)
Alexey suggested that replace and enclose can be combined.

Q: Is this worth doing?
Alexey: Seems complex, but worth doing. This is in our Charter.

Suggestion to make this Experimental. Alexey replied that we don't need to decide on Experimental versa Standard Track until we get to WGLC. We might get some implementations before the WGLC. Ned commented in Jabber that this might be very difficult to implement for Sun. Jeffrey has asked Ned if making iterator separate from MIME test would help. Ned said that it would.

Alexey recaps issues raised during WGLC for Subaddress. There were some minor editorial issues, also the document is not quite clear that both prefix and suffix forms are allowed, also that multiple characters can be used as a delimiter. Ken needs to submit a new revision before this
can be sent off to IESG.

Review of other drafts has followed. WG needs an editor for SMTP AUTH test (this can't be a part of the base document, as we don't want to have a normative reference to SMTP AUTH document).
Ned agreed to take this one.

Alexey has asked the WG about the state of Notify draft. Barry will start a thread about it, so that the WG can figure out which notification mechanisms should be present and which can be dropped.

Barry has asked about the state of Relational draft. Alexey thinks that it is ready for WGLC.

Action items:
Philip to publish a new revision of base spec to address editorial comments raised during WGLC. Jeffrey should post a short message about security review of the Vacation document.
Ned to post an update to vacation, if he has any changes pending.
Ned to repost his old message about addressing I18N in Regex.
Philip to post a new revision of Body to address searching of non-textual bodyparts.
Alexey/Cyrus to send Edit header to IESG (do writeup, etc.)
Alexey to issue WGLC for Spamtest.
Cyrus to issue WGLC for IMAPFlags.
Cyrus to issue WGLC for Refuse/Reject.
Alexey/Cyrus to issue WGLC for Relational.
Ken to address comments raised during WGLC for Subaddress draft.
Tony/Cyrus to publish combined MIME test draft.
Ned to create a new draft for testing SMTP AUTH.
Barry to revive the Notify draft.
Alexey/Cyrus to propose updated WG milestones on the mailing list.

<Prev in Thread] Current Thread [Next in Thread>