Обсуждение: Re: [Fwd: 58A0-9E5C-CE89 : CONSULT from pgsql-slavestothewww (post)]
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 'k, this is a GLOBAL setting, not a local to the list one ... I can remove the approve rule, but here is the complete list of them ... I've never modified these before, either adding or removing, these are teh default ones that get installed ... as you will note with most of htem, they are anchored to the right side, so it isn't just *any* approve: Let me know ... configset GLOBAL admin_body <<ENDAAB /^accept$/i /^reject$/i /\bcancel\b/i /\badd me\b/i /\bdelete me\b/i /\bremove\s+me\b/i /\bchange\b.*\baddress\b/ /\bsignoff\b/i /\bsubscribe\b/i /^sub\b/i /\bunsubscribe\b/i /^unsub\b/i /^\s*help\s*$/i /^\s*info\s*$/i /^\s*info\s+\S+\s*$/i /^\s*lists\s*$/i /^\s*which\s*$/i /^\s*which\s+\S+\s*$/i /^\s*index\s*$/i /^\s*index\s+\S+\s*$/i /^\s*who\s*$/i /^\s*who\s+\S+\s*$/i /^\s*get\s+\S+\s*$/i /^\s*get\s+\S+\s+\S+\s*$/i /^\s*approve\b/i /^\s*passwd\b/i /^\s*newinfo\b/i /^\s*config\b/i /^\s*newconfig\b/i /^\s*writeconfig\b/i /^\s*mkdigest\b/i ENDAAB - --On Saturday, April 21, 2007 17:40:50 +0100 Dave Page <dpage@postgresql.org> wrote: > Here's one... > > -------- Original Message -------- > Subject: 58A0-9E5C-CE89 : CONSULT from pgsql-slavestothewww (post) > Date: Sat, 21 Apr 2007 13:18:36 -0300 > From: pgsql-slavestothewww-owner@postgresql.org > Reply-To: majordomo@postgresql.org > To: pgsql-slavestothewww-owner@postgresql.org > > __ > The following request: > > "(post to pgsql-slavestothewww)" > > was sent to > by World Wide Web Owner <www@wwwmaster.postgresql.org>. > > The request requires your confirmation for the following reason(s): > > GLOBAL ADMIN BODY: /^\s*approve\b/i matched "Approve" at line number 4. > > To accept or reject this request, please do one of the following: > > 1. If you have web browsing capability, visit > > <http://mail.postgresql.org/mj/mj_confirm/domain=postgresql.org?t=58A0-9E5C-C > E89> and follow the instructions there. > > 2. Reply to majordomo@postgresql.org > with one of the following two commands in the body of the message: > > accept > reject > > (The number 58A0-9E5C-CE89 must be in the Subject header) > > 3. Reply to majordomo@postgresql.org > with one of the following two commands in the body of the message: > > accept 58A0-9E5C-CE89 > reject 58A0-9E5C-CE89 > > 4. If you know the administrative password for the pgsql-slavestothewww > list, > all pending requests can be managed by visiting > > <http://mail.postgresql.org/mj/mj_wwwadm/domain=postgresql.org/pgsql-slavesto > thewww?func=showtokens> > > If you do not respond within 7 days, this token will expire, > and the request will not be completed. > > - ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email . scrappy@hub.org MSN . scrappy@hub.org Yahoo . yscrappy Skype: hub.org ICQ . 7615664 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (FreeBSD) iD8DBQFGKldW4QvfyHIvDvMRAnT8AJwPDAFa191tyG9Iwiziay2KeLETtACdFywo 4FUXL8zCMdvSKCGXLdaLK9c= =1fS4 -----END PGP SIGNATURE-----
Marc G. Fournier wrote: > 'k, this is a GLOBAL setting, not a local to the list one ... I can remove the > approve rule, but here is the complete list of them ... I've never modified > these before, either adding or removing, these are teh default ones that get > installed ... as you will note with most of htem, they are anchored to the > right side, so it isn't just *any* approve: In my opinion they are too ambiguous to be used as expressions for rejecting a mail to the list. Maybe mj2 was recently upgraded to a new version that contained an "improved" list of regexes to disallow? > /\bsubscribe\b/i So anyone who sends a mail with the word "subscribe" in it will have it rejected? That seems rather unexpected. Note that it isn't anchored, except by blanks so it seems just using the word in a sentence causes it to fire, unless I'm missing something. > /^\s*approve\b/i > /^\s*passwd\b/i > /^\s*newinfo\b/i > /^\s*config\b/i > /^\s*newconfig\b/i > /^\s*writeconfig\b/i > /^\s*mkdigest\b/i I think these should make an effort to find the arguments just beyond the command and the end of line. Really they should not be there at all. What purpose do they serve, beyond protecting lame list admins who manage to send commands to the list itself instead of the request address? -- Alvaro Herrera http://www.CommandPrompt.com/ The PostgreSQL Company - Command Prompt, Inc.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 None of these messages are rejected ... they are only put over to the moderators queue for later processing ... that is what Dave is having a problem with, that he has to go in and approve them instead of them just going straight through ... As to recent upgrade .. not in the past few weeks, no ... - --On Saturday, April 21, 2007 21:12:27 -0400 Alvaro Herrera <alvherre@commandprompt.com> wrote: > Marc G. Fournier wrote: > >> 'k, this is a GLOBAL setting, not a local to the list one ... I can remove >> the approve rule, but here is the complete list of them ... I've never >> modified these before, either adding or removing, these are teh default >> ones that get installed ... as you will note with most of htem, they are >> anchored to the right side, so it isn't just *any* approve: > > In my opinion they are too ambiguous to be used as expressions for > rejecting a mail to the list. > > Maybe mj2 was recently upgraded to a new version that contained > an "improved" list of regexes to disallow? > >> /\bsubscribe\b/i > > So anyone who sends a mail with the word "subscribe" in it will have it > rejected? That seems rather unexpected. Note that it isn't anchored, > except by blanks so it seems just using the word in a sentence causes it > to fire, unless I'm missing something. > > >> /^\s*approve\b/i >> /^\s*passwd\b/i >> /^\s*newinfo\b/i >> /^\s*config\b/i >> /^\s*newconfig\b/i >> /^\s*writeconfig\b/i >> /^\s*mkdigest\b/i > > I think these should make an effort to find the arguments just beyond > the command and the end of line. Really they should not be there at > all. What purpose do they serve, beyond protecting lame list admins who > manage to send commands to the list itself instead of the request > address? > > -- > Alvaro Herrera http://www.CommandPrompt.com/ > The PostgreSQL Company - Command Prompt, Inc. - ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email . scrappy@hub.org MSN . scrappy@hub.org Yahoo . yscrappy Skype: hub.org ICQ . 7615664 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (FreeBSD) iD8DBQFGKs5e4QvfyHIvDvMRAnucAKCcMhpMYAarWQqtRVyE1Ja9NVYnIgCeJkJI 5BkSMk+qs7SWW+2MlEHBiBk= =BaIv -----END PGP SIGNATURE-----
Marc G. Fournier wrote: > > None of these messages are rejected ... they are only put over to the > moderators queue for later processing ... that is what Dave is having a problem > with, that he has to go in and approve them instead of them just going straight > through ... And given that *all* of the moderation messages from the website include the word 'approve'... I think you can see the irony :-) I'm still confused about why this just started happening though. /D
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 - --On Sunday, April 22, 2007 20:21:20 +0100 Dave Page <dpage@postgresql.org> wrote: > Marc G. Fournier wrote: >> >> None of these messages are rejected ... they are only put over to the >> moderators queue for later processing ... that is what Dave is having a >> problem with, that he has to go in and approve them instead of them just >> going straight through ... > > And given that *all* of the moderation messages from the website include > the word 'approve'... I think you can see the irony :-) > > I'm still confused about why this just started happening though. I don't know ... my only thought is that at some point a configdefault was set on the access_rules for that last ... the only way that would matter is if we *had* put in a special rule, like we did for committers for @pgfoundry.org emails ... I've just added: post allow $host =~ /wwwmaster\.postgresql\.org/ to the access_rules for pgsql-slavestothewww ... see if that fixes it, I'm not sure if access_rules is parsed before or after the GLOBAL TABOO stuff, but if so, that will fix it without having to reduce that other list ... - ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email . scrappy@hub.org MSN . scrappy@hub.org Yahoo . yscrappy Skype: hub.org ICQ . 7615664 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (FreeBSD) iD8DBQFGLDEg4QvfyHIvDvMRAhjiAJ4vEmN4Ly1lXzlTtpLeYVYqGQbr6QCfcyS8 EPfjXeKrKKhUSxJkOh9shKk= =MNbw -----END PGP SIGNATURE-----
Marc G. Fournier wrote: >> I'm still confused about why this just started happening though. > > I don't know ... my only thought is that at some point a configdefault was set > on the access_rules for that last ... the only way that would matter is if we > *had* put in a special rule, like we did for committers for @pgfoundry.org > emails ... > > I've just added: > > post > allow > $host =~ /wwwmaster\.postgresql\.org/ > > to the access_rules for pgsql-slavestothewww ... see if that fixes it, I'm not > sure if access_rules is parsed before or after the GLOBAL TABOO stuff, but if > so, that will fix it without having to reduce that other list ... Hmm, I had a few this morning, but I think they were all slightly before this email. I'll let you know if I get any more - thanks. /D